UNIVERSITY OF CALIFORNIA, SAN DIEGO

Prominence: from Sensation to Language

A dissertation submitted in partial satisfaction of the
requirements for the degree Doctor of Philosophy in
Cognitive Science and Linguistics
by
Tracy Cameron Mansfield
Committee in charge:
Professor Ronald W. Langacker, Chair
Professor Chris Barker
Professor Jeff L. Elman
Professor Kathleen A. Hubbard
Professor Carol A. Padden
Professor Walter J. Savitch
1997
Copyright
Tracy Cameron Mansfield, 1997
All rights reserved.
iii
Signature Page
The dissertation of Tracy Cameron Mansfield is approved,
and it is acceptable in quality and form for publication on
microfilm:
Chair
University of California, San Diego
1997
iv
Dedication
To Deana and Caitria...
One Dragon.
v
Epigraph
CADE: It will be proved to thy face that thou hast men about thee that
usually talk of a noun and a verb, and such abominable words as no
Christian ear can endure to hear.
William Shakespeare, King Henry VI
Daddy! Watch this... Pnk ... Prple ... Pnk ... Prple ... They
make different shapes!
Caitria Mansfield at 3 years of age
vi
Table of Contents
Signature Page .................................................................................................... iii
Dedication........................................................................................................... iv
Epigraph................................................................................................................v
Table of Contents................................................................................................ vi
List of Figures and Symbols .................................................................................x
List of Tables ...................................................................................................... xi
Foreword............................................................................................................ xii
Acknowledgments ............................................................................................ xiv
Vita.................................................................................................................. xviii
Abstract ............................................................................................................. xix
CHAPTER 1: The Difference that Prominence Makes .............................................1
1 Prospecting the Claim: Collecting Data...........................................................10
2 Refining the Ore: Excluded Data.....................................................................15
3 Tools: Theoretical and Terminological Background.......................................20
3.1 Discourse I: Sensation to Communication..............................................20
3.2 Discourse II: Expectation and Revelation ..............................................30
3.3 Discourse Summary................................................................................37
4 Polishing the Stones: Display of Analytic Results ..........................................38
4.1 Elaboration..............................................................................................38
4.2 Revelation ...............................................................................................43
4.3 Linked Instances of Abnormal Prominence............................................46
4.4 Summary: The Functions of Abnormal Prominence ..............................48
5 Conclusion .......................................................................................................49
vii
CHAPTER 2: The Difference between Routine and Volitional Prominence ..........50
1 What prominence IS .........................................................................................54
1.1 Prominence IS Anatomical Configuration...............................................55
1.2 Prominence IS Physiological Production ................................................56
1.3 Prominence IS Physiological Perception.................................................72
1.4 Prominence IS Baseline Excursion..........................................................76
1.5 Prominence IS Effortful Motion..............................................................78
2 Where prominence GOES ..................................................................................80
2.1 Prominence GOES on Points of Punctuation ............................................80
2.2 Prominence GOES on Notes of Music ......................................................83
2.3 Prominence GOES on Peaks of Tunes ......................................................91
2.4 Prominence GOES on Units of Syntax....................................................109
2.5 Prominence GOES on Beats of Rhythms ................................................121
2.6 Prominence GOES on the Beat unless it is Making Sense......................130
3 What prominence DOES ..................................................................................131
3.1 Prominence DOES Semantics .................................................................131
3.2 Prominence DOES Context .....................................................................136
3.3 Prominence DOES Discourse..................................................................137
3.4 Prominence DOES Theme and Rheme ...................................................145
3.5 Prominence DOES Outward Focus .........................................................154
3.6 Prominence DOES Inward Focus............................................................166
3.7 Prominence DOES Focus, both Outward and Inward.............................169
3.8 Prominence DOES Proportion.................................................................172
4 Prominence MAKES Sense...............................................................................176
CHAPTER 3: Elaboration......................................................................................178
1 Power .............................................................................................................181
2 Precision.........................................................................................................192
2.1 Intolerance ............................................................................................192
2.2 Position .................................................................................................202
3 Conclusion .....................................................................................................209
viii
CHAPTER 4: Revelation .......................................................................................211
1 Change and Counterpart ................................................................................219
2 Setting ............................................................................................................220
3 Substitution ....................................................................................................222
3.1 Regular Substitution .............................................................................223
3.2 Recapitulation .......................................................................................226
3.3 Disambiguation of Reference and Pronouns ........................................229
3.4 Reference and Parallel Structures .........................................................233
3.5 Summary of Substitution ......................................................................241
4 Addition .........................................................................................................241
5 Derivation ......................................................................................................243
5.1 Conventional Pairs................................................................................244
5.2 Conventional Sets .................................................................................250
5.3 Basic Conventional Domains................................................................257
5.4 Complex Conventional Domains..........................................................263
5.5 Contextual Pairs....................................................................................264
5.6 Contextual Sets .....................................................................................266
5.7 Basic Contextual Domains....................................................................267
5.8 Complex Contextual Domains..............................................................267
5.9 Derivation Summary.............................................................................270
6 Conclusion .....................................................................................................271
CHAPTER 5: Linked Instances of Volitional Prominence....................................272
1 Dissociation ...................................................................................................273
2 Coordination ..................................................................................................277
3 Timing............................................................................................................278
4 Conclusion .....................................................................................................280
ix
CHAPTER 6: Conclusions and Future Research...................................................281
1 Live Data........................................................................................................282
2 Increased Complexity ....................................................................................282
3 Increased Selectivity......................................................................................284
4 More Modalities of Communication..............................................................284
5 Developmental Studies ..................................................................................287
6 The Big Picture ..............................................................................................288
6.1 Reasonable Speculation ........................................................................288
6.2 Actual Results .......................................................................................290
Afterword..........................................................................................................294
Appendix I: The Anatomy of Prominence....................................................295
Appendix II: Distribution of Data ..................................................................298
Appendix III: Other Corpus: Sources for Abnormal Prominence ...................301
Appendix IV: Brown Corpus: Sources for Abnormal Prominence ................303
Author Index.....................................................................................................312
Subject Index ....................................................................................................316
References.........................................................................................................332
x
List of Figures and Symbols
1-1: Wireframe Necker cube .....................................................................................1
1-2: Necker Cube Elaborations .................................................................................3
1-3: Necker Cube Resolved.......................................................................................4
1-4: Transduction ....................................................................................................20
1-5: Perception ........................................................................................................23
1-6: Cognition .........................................................................................................23
1-7: Articulation ......................................................................................................25
1-8: Discourse .........................................................................................................26
1-9: The Archive (Atemporal Model) .....................................................................32
2-1: Sample of Steeles Transcription.....................................................................85
2-2: Palmers Six Tone-Patterns .............................................................................95
2-4: The Tone Pattern Generator in Action...........................................................124
2-5: Baseline Units ................................................................................................127
2-6: Focus, Presupposition, and Assertion............................................................155
3.8: Direct Iconic Proportion Symbol (^).............................................................175
I-1: Anatomy of the Vocalization Tract ...............................................................295
xi
List of Tables
2-1: Butlers Association of Punctuation with Accent............................................82
2-2: Firbas Appearance/Existence Scale..............................................................142
2-3: Firbas Attributive Scale................................................................................143
2-4: Prominence Components Associated with Focus..........................................174
II-1: Prominence Frequency (Functions in Order of Analysis) .............................298
II-2: Prominence Frequency (Functions Ranked by Total Instances) ...................298
II-3: Expanded Functional and Grammatical Categories.......................................299
II-4: Expanded Categories with Subtotals .............................................................300
xii
Foreword
ROPER So now youd give the Devil benefit of law!
MORE Yes. What would you do? Cut a great road through the law to
get after the Devil?
ROPER Id cut down every law in England to do that!
MORE (Roused and excited) Oh? (Advances on ROPER) And when the
last law was down, and the Devil turned round on youwhere
would you hide, Roper, the laws all being flat? (He leaves him) This
countrys planted thick with laws from coast to coastmans laws,
not Godsand if you cut them downand youre just the man to
do itdyou really think that you could stand upright in the winds
that would blow then? (Quietly) Yes, Id give the Devil benefit of
law, for my own safetys sake.
Robert Bolt, A Man for All Seasons
It all started so very quietly.
Niccolo di Niccoli, a 15th-century Vatican clerk, designed what we think of
now as cursive handwriting in order that whole words could be written without having
to lift the pen off of the paper, and it was copied by European clerks who referred to it
as the Italian hand. The italic typefont itself was created by Aldus Manutius in 1501
as a small typeset representative of the Italian hand for use in pocket books, and since
it was developed as a primary typeset, it was not originally used to set off anything,
but rather it composed the entire text.
Fifty years later, there was still neither use nor mention of italics in John Harts
manuscript, The Opening of the Unreasonable Writing of our Inglish Toung; however,
just twenty years after that (1569), he did describe the use of italics for prominence in
xiii
his foundational volume on English linguistics, An Orthographie, Conteyning the Due
Order and Reason, Howe to Write or Paint Thimage of Mannes Voice, Most Like to
Life or Nature.
Some time between then and now, italic fell into disuse as a primary textual
font. I suspect that this happened after the 17th century when the Dutch started to
emphasize vertical rather than angled types, but before the introduction of the linotype
machine in the 1880s, along with the adoption of a standardized handwriting
pedagogy, such as Spencerian script in the mid-1800s.
Contemporarily, italics are used to mark excursions of variable length away
from the main stream of a written text. Energetic gestures, both spoken and signed, are
also articulated as departures away from the primary flow of conversation, and these
two types of asides have come to be associated with one another. Some of these breaks
are violent emotional explosions, while others are rigidly controlled tightenings of
meaning which require great articulatory finesse, but of course most asides fall short
of these extremes. Such gestural behavior actually spans language and species
boundaries, and provides insight into the earliest forms of communication.
In this dissertation, I explain just what the devil is going on, while being
mindful of the trees, for my own safetys sake.
xiv
Acknowledgments
First and foremost, I thank Deana for giving me precisely what I needed to
complete this project, namely the time to work while she fielded distractions, and the
motivation to keep working when all I wanted to do was make up for lost sleep. Her
love and her confidence in me is borne out by the fact that this dissertation is finished.
I also want to thank Caitria, my daughter, who during the time that all of this
work was going on was only one-through-four years old. She was very generous with
her time, advice, food, smiles, and love, all of which helped to sustain me when I just
felt like chucking the whole mess in the trash.
My parents, Lorna and Dewey, have always come through for me, no matter
what kind of help I needed. An education is merely one of the many luxuries that they
have afforded me, as well as the freedom to spend more time with my family, and the
understanding that I am a fortunate man indeed.
Jeff and I are brothers: always best friends; rarely bitter enemies; and usually
partners in business, war, crime, or justice. When I needed to take a break from life
and go to grad school, he helped me to get there and stay there. When I needed to take
a break from grad school to come back to life, he found me a wife. I know that sounds
grossly primitive, but its essentially what happened. Without Jeff, I wouldnt have
gotten into trouble nearly as often as I did, but then again, without Jeff, I also wouldnt
have gotten anywhere nearly as interesting as I have.
xv
The King of Imagination provides me with an inspiration of which I am sure he
is wholly unaware. The way Bill lives and who he is always show me how thin the
reality threshold is, and it brings me to believe not just that anything is possible, but
entirely probable.
The greatest gift that I could wish for my friends is that they might have the
opportunity to come to know each other as I have known them. Each of them is part of
who I am, part of the person who was able to write this book. They might each take a
little satisfaction in knowing that it is their speech patterns that have helped to form
my intuitions about the structure of the language that I use, and in that way their own
speech is given special attention. Then again, they might not give a hoot... some of my
friends are just like that, bless them.
Thank you Margaret Langdon, for admitting me into the linguistics program on
the simple condition that I promised to actually show up.
And speaking of promises, Ron Langacker has been friend enough to allow
himself to be persuaded by my most outrageous demands, perhaps despite his better
judgment, all because he felt that he could see the glimmer of promise in some of my
work. Just as importantly, he has also been advisor enough to ensure that I made good
on that promise. He has stuck his neck out for me so often that had Lamarck been
correct in his theories, Ron would have become giraffid long before now. I did all this
work just to prove to him that I am not, in fact, lazy.
xvi
Without Chris Barker, my most confident response to formal semantics
questions would have been a resounding, Duh? He wins points for blithely aligning
himself with Caitrias perspective, cheerfully crouching down when talking with her,
much as he must have done when dealing with earlier drafts of this dissertation.
Kathleen Hubbard and I used to hang over the back fence in the midst of doing
our chores to swap stories about our various aches, pains, setbacks, and triumphs.
While the fence was figurative, the empathy was literal, and that helped me to get
through this project just as much as her pointing out that I needed to whip my anatomy
into line, not to mention my physiology.
Jeff Elman, Carol Padden, and Walt Savitch are the external members of my
committee, and my contact with them over the years has been limited by my
reluctance to let them read any draft that I thought might tempt them to quit. Despite
this neglect, they have supported my efforts by agreeing time and again to serve on
successive committees. Year after year, they have maintained their confidence in my
ability to complete this project, which is finished in part precisely because I didnt
want to disappoint them. I hope that they can now feel that their faith was well placed.
Karen van Hoek, Suzanne Kemmer, and Shanti Hofshi all gave their invaluable
commentary on earlier drafts of this material, which started out as The the Paper.
Randall Rose was exceedingly generous with his time and insightful commentary
when the project had reached ten times its original size.
xvii
Karen Anderson, Sharon McGill, and Nancy Washington have all helped me to
thread my way through a maze of petitions, program requirements, language exams,
fee deadlines, and other tasks designed to test the strength of my desire to complete the
program. The fact that people do graduate is testament to their skill and patience.
Linda Murphy culled the library, and gave the graduates salvage rights. Unable
to buy books to lift my spirits, this treasure helped me to feel better. These castoffs
brought light to a number of dark corners within linguistics, and I found many early
drafts written by linguists before they were famous. This perspective gave me hope.
When I first got to UCSD, Rock Hunter was toiling ceaselessly to wring a
series of reluctant miracles out of the guts of a computer system adapted from an IBM
Selectric II, a Model F crank starter, and a grumpy transponder from the Ark that was
still sending out fitful distress signals to God. Without Rocks devoted efforts, I simply
wouldnt have had the resources available to work on this or any other project.
I would also like to say that I consider myself very fortunate to be able to count
a number of people among the staff and TAs at the Muir Writing Program as my good
friends. What I know about teaching I owe to them, and some of my favorite times at
UCSD were spent with these people.
Thank you, Spirit, for your inspiration.
Finally, thank you everyone else who deserves to be listed here, but whom my
erratic memory has refused to pester me about in a timely fashion... sine vobis non.
xviii
Vita
August 8, 1962 Born, Castro Valley, California
1984 University of California at Davis
B.A., Linguistics; Minor: Psychology
Internship: raptor rehabilitation center
1987-1992; 1995-1996 Teaching Assistant, Muir Writing Program
1987-1992 Graduate Student Association Representative
1991 University of California at San Diego
M.A., Linguistics; C.Phil., Linguistics
1997 University of California at San Diego
Ph.D., Cognitive Science and Linguistics
xix
Abstract
Abstract of the Dissertation
Prominence: from Sensation to Language:
by
Tracy Cameron Mansfield
Doctor of Philosophy in Cognitive Science and Linguistics
University of California, San Diego, 1997
Professor Ronald W. Langacker, Chair
Two written corpora were sifted to uncover 1,452 contemporary American
English words in declarative sentences that were isolated for prominence with italics.
This research shows that conversants rely upon a direct iconic proportion between
phonological and semantic intensity, resolving ambiguous form-meaning associations
by exceeding the dimensional limitations of sequential segmentation. A speaker/signer
aligns a model of reality adopted by a looker/listener with its own, either identifying a
word-internal meaning disparity through ELABORATION, or revealing a word which can
xx
repair a word-external mismatch with REVELATION. These functions have their basis in
the cognitive abilities of POWER and PRECISION, whose primitive progenitors provided
the threshold across which sensation was adapted for communication.
Elaboration signifies either an increase in semantic POWER (huge is very
huge) or a discharge of EMOTION. Meanings portrayed with essentially spatial imagery
are articulated with greater PRECISION, often as INTOLERANCE (all is absolutely all).
Also, parcelling semantic variations as a types instances allows a POSITION in that type
to elaborate a words meaning. PROPER instances are paragonal (where green is a
real green-green), and odd ones are PERIPHERAL (where green is a weird green).
In revelation, contextual mismatches are identified by shifting the intensity or
location of primary prominence. The prominent word is exchanged for a counterpart
located in a parallel setting either in discourse or in common knowledge. SUBSTITUTION
exchanges an explicit, single-word discourse counterpart. DERIVATION exchanges
implicit or multi-word counterparts through an appeal to conventional or contextual
associations. Greater prominence is used to repair broader mismatches in construal,
correcting errors which are rooted more deeply inside common knowledge or memory.
With LINKED close instances, mutual prominence is interpreted as significant.
Mutually prominent words are taxonomically DISSOCIATED as labels, or COORDINATED
according to their pre-existing causal or consequential links. Both dissociation and
coordination can be augmented by the TIMING provided by close structural parallels.
1
CHAPTER 1
The Difference that Prominence Makes
There can be no difference that doesnt make a difference elsewhere
William James, Pragmatism (Lecture II)
Lets start out with a familiar Necker cube:
Figure 1-1: Wireframe Necker cube
This sensory illusion is intriguing precisely because it displays a classic form-meaning
ambiguity. Its lower-dimensional construal might be that of contiguous polygons (or
triangles .), but its higher-dimensional construal is that of a cube, where one of its
two central vertices (A or B) is repelled into the background away from the viewer,
while the other is one drawn into the foreground toward the viewer. The figure has one
FORM, but perceptive qualifications such as scanning and dimensionality conspire to
broaden the potential range of its MEANING, where any of these form-meaning pairs is a
SYMBOL.
A
B
2
Language is full of similar form-meaning illusions, such as the classic ...it is
my privilege to extend to you a laurel, and hearty handshake (Laurel and Hardy?),
the generic Joe ran into John and he fell down (which fell?), and the prosaic object,
digest, contract, permit (nouns, verbs, mixed?). Each of these examples contains a
core phonological form, as in the stressless versions of /k.n trkt/ or Use the example
you wrote. Such a kernel can be associated with a variety of meanings depending
upon its embellishment, as in the nominal /k.n trkt/ versus the verbal /k.n trkt/
(a WORD-INTERNAL MEANING DIFFERENCE), or Use the exmple you wrote versus Use
the example you wrte (a WORD-EXTERNAL or CONTEXTUAL MEANING DIFFERENCE).
Changes made outside of an utterances core, such as with pauses and pitches,
disperse or refract this cores association among different members of a set of meaning
variations; however, in the absence of external cues specifying a form-meaning pair, a
default meaning can be assumed. I say can because some of these defaults are much
stronger than others. With no clarifying external cues, the core form of the phrase. I
have my /s k.t/ in the drawer probably refers to the tie and not to the predicament,
while a laurel is usually the leafy crown and not the actor, yet stressless /k.n trkt/
might just as easily be a noun as a verb. Language users rely upon the phonological
underspecification allowed by these defaults, such as when readers inflect writing.
External cues, such as prominence, can be used to advance or evade such implications.
Normal prominence signals congruence between the message it embellishes and these
defaults, which implies a proper understanding between language users, but abnormal
prominence warns of potential or actual meaning differences and misunderstandings.
3
Prominence is only one of the cues which can resolve a linguistic illusion, but
it is favored over other changes which tend to promote rather than clarify formal
ambiguities. In fact, some modifications integrate so well with an illusions form that
they do not affect its potency at all, as in the application of shading or reduplication to
this Necker cube:
Figure 1-2: Necker Cube Elaborations
The first change (from 1-2.0 to 1-2.1) shades the sides surrounding the right central
vertex (B), while those around the left central vertex (A) are left unshaded, and then the
second change (from 1-2.1 to 1-2.2) reduplicates the shaded cube. Unlike prominence,
shading and reduplication conform to the dimensional limitations established by the
illusion, which in this case would be something like the extended surfaces of the three
pairs of parallel planes which intersect to create the original cube (1-2.0). Assimilated
changes cannot help to resolve the original formal ambiguity: the central vertices of
the shaded cube can still be easily construed as moving in or out as surrounded by their
shaded (B) or unshaded (A) sides, as can all of the central vertices of the reduplicated
cubes (making nine or ten cubes, depending upon which go in, and which come out).
A
B
1 2 0
4
Like prominence, however, the following cylinders potentially exceed the
illusions limited dimensionality, because they exist outside of its determining planes:
Figure 1-3: Necker Cube Resolved
They meet that potential when they exhibit incongruous behavior (occlusion) denied to
the strictly planar components, where that behavior specifies a relation between those
planes that would otherwise have been definitionally unrelatable. This allows for at
least partial resolution of the illusion in 1-3.1, which does not directly relate A with B,
and for complete resolution in 1-3.2, which does.
More specifically, A can still be construed as projecting outward in 1-3.1, but
only under two sets of circumstances. First of all, the effects of the small cylinder can
be ignored because the line it occludes is isolated from the illusions pivotal elements,
namely A and B, which allows the viewer to pay more attention to those vertices than
the cylinder. Second of all, the occlusion can be absorbed (so A still projects forward)
when the line is interpreted as bent under the cylinder, in congruity with the original
illusions dimensional limitations. The larger cylinders resolution is much stronger,
because it is neither isolated from the pivotal elements, nor is its occlusion slight
enough to allow it to be assimilated. The relation between A and B is direct.
A
B
B
A
B
0 1 2
5
As sequential segmental symbolism, spoken language is also of necessarily
limited dimensionality, just like the cube, and so a core sequence like brown eyes and
hair would be prone to ambiguous interpretation were it not for extradimensional cues
like the pitches and pauses in: 1) [(brwn) (eyes and hair)]; or 2) [(brwn eyes) and
(hir)]. Elaborations such as big dark brown eyes and hair do not resolve the illusion
because they accommodate the cores dimensionality; however, external cues found in
context or prosody can be used to exceed that dimensionality, and resolve the
ambiguity.
With this resolving function in mind, lets take a look at the following four
passages, all of which were plucked straight out of the Fiction: Romance section of
the Brown corpus (Francis and Kucera, 1961; with line numbers from the corpus
superscripted for identification):
[1-1] The ward was a small one, four beds, kept reserved for female
alcoholics.... there was something about her Alice But this
woman's name was Rose Bancroft!

She looked about sixty, though I recalled that the chart gave her
age as forty-four. An ugly scar disfigured the somewhat familiar
puffy face, already marred by the tell-tale network of broken
red veins that heavy drinkers carry. Her coarse hair was two-
colored bleached blonde and its real, dirty gray. Oh, could it
be? No, no it was an unfortunate resemblance, that was all it
was.

[Rose speaks] Im not going to be one of your guinea pigs. Let

your pupils learn on someone else, doctor. Just let me die in
peace.
I stared at her, almost speechless. Her little speech was totally
out of character with the sort of person I thought she was. Even
her voice had taken on a more cultivated tone.
P19 1400
6
[1-2] Leaving Cathy with them [for two weeks], Myra had gone out
to the Coast for a supposedly brief visit; but she had stayed all
winter, and Cathy had stayed all winter too with them.
That was an awfully long two weeks.
P15 0270
[1-3] If he ever did such a thing again she'd die of shame.
P03 1530
[1-4] I know you feel badly, but that sounds like such a queer thing
for you to say.
Does it? he asked. Yes, perhaps. I'm supposed to joke about
things, aren't I? But sometimes life can be rather a
disappointing business.
P14 1250
The most obvious question to ask seems fairly harmless, namely: What are the italics
for? Our objective is to find out how the isolated italicized words differ in form and
meaning from their counterparts which have had their italics removed.
Just take a first pass at explaining these meaning differences. In [1-1], so-called
external contrast doesnt provide a consistent explanation: stared is not a counter
to some other process like glanced; instead, while a normal stare is a lingering
rather than a wandering gaze, the focus of stared is intense enough to penetrate a
patients alcoholic veneer. In [1-2], long is six times longer than long, and two
weeks last an entire Winter. If the ever in [1-3] were just plain ever, then he
could still occasionally get away with doing such a thing without her dieing of shame
(or perhaps even noticing). This is because ever normally allows for some tolerance
which the additional prominence on ever squeezes into nothing. Finally, in [1-4],
you does not simply contrast with another pronoun in context, or mean you of all
people, but rather it encompasses only those qualities found in the normal you (such
as that the person in question normally jokes about things) while leaving out queerer
behaviors which have suddenly cropped up in the current you (his seriousness).
7
This rough treatment shows that none of these examples are merely externally
contrastive, and that they are unified by express, positive similarities in their behavior,
and not just by a simple lack of normal prominence patterns. To begin with, both the
penetration of stared and the endurance of long are greater in POWER than normal;
in addition, both ever and you define more tightly controlled, albeit figuratively
spatial areas, where the normal degree of tolerance allowed in the portrayal of their
imaged or imagined boundaries is now drawn with stronger PRECISION. Greater power
and precision are both more elaborate portrayals of increased INTENSITY. To the degree
that these Brown corpus examples are valid representations of audible gestures, this
first approximation suggests that an utterances phonological intensity is in DIRECT
ICONIC PROPORTION to the semantic intensity construed in its meaning, where intensity
manifests itself as power or precision. Right now, this proportion is as gross-grained as
it can get (binary), but finer gradations will be supported later. This is an intriguing
beginning, and would be strong enough in itself to motivate a more formal analysis.
But now take a look at this next set of examples, taken from the same section
of the Brown corpus (Fiction: Romance), in which external contrast does play a
significant part in making sense of the changed meaning:
[1-5] the sound coming through the walls like something on the
other side of the curtain, so you knew they heard you when they
were quiet, and while you lay wondering what they had heard,
you listened.
P09 0220
[1-6] So they stayed quiet and hung not on what he said but on how
he said it, not listening exactly, but rather, feeling.
P09 1140
[1-7] Holy mackerel, thats the most unique dog I ever saw.
P16 0950
[1-8] Roy smiled he did have a nice smile.
P03 1670
8
These are not instances of word-internal meaning changes, and they do not display a
straightforward application of increased intensity; instead, they make reference to
COUNTERPARTS in the discourse, and they are marked by changes in the normal
placement of prominence. This shift incidentally results in an increase in intensity,
which can then be augmented even further to reflect a greater strength of conviction.
These counterparts vary liberally in their EXPLICITNESS and ENCAPSULATION. For
example, in [1-5], you and they are single words actually appearing in the context,
and so they are entirely explicit and thoroughly encapsulated counterparts for one
another. When an explicit counterpart is more broadly DISTRIBUTED, you get a sentence
like [1-6], in which the counterpart to feeling is a more rambling passage listing
alternatives to feeling such as not what he said but how he said it, not listening
exactly. An IMPLICIT version of such a widely distributed counterpart can take the
form of a CONVENTIONAL SET like the personal pronouns, as in [1-7], where other
people (an implied he, they, or some other pronoun) have probably seen other dogs
which have struck them as peculiar, but the girl represented by I has not. Finally,
implicit counterparts can still be tightly encapsulated, in that the implicit (did) not
against which did contrasts in [1-8] does not appear in so many words (and so is
implicit), but negation itself is still a tightly encapsulated notion.
In [1-1] through [1-4], the abnormally prominent words each displayed a direct
iconic proportion between the increased power or precision of their phonological form
and their meaning. In [1-5] through [1-8], this phonological increase is proportional to
the disparity rupturing the conversants reality models, as reflected in the strength of
9
conviction needed to realign them. This is iconic because greater phonological effort is
expended to more deeply penetrate either discourse, memory, or belief structures. The
actual level of effort expended is intense (greater-than-primary) only when it needs to
distinguish a word which is already located at the default site of normal, primary
sentence stress; otherwise, a primary level of prominence makes a distinct enough
impression.
To this point, power and precision have been described only as providing the
bases for purely linguistic functions, showing that contrast and emphasis identify
behaviors which are elemental only in linguistic alchemy; however, this analysis will
go on to suggest a scenario in which their primitive progenitors are the two global,
fundamental cognitive abilities which are still used to evaluate either the force of a
single sensation (absolute magnitude), or to compare the difference in force between
two or more sensations (relative magnitude). Empirical support for this intuition will
show that prominence has roots in cognitive behavior primitive enough to represent
the first step across the threshold from incoming sensation to outgoing
communication, and then on into language in general, as well as across specific
languages and language modalities.
Now, this strikes me as a very pretty description, but so far it rests on a rough
analysis of nine instances in eight whole examples, and that analysis relies in turn on a
network of vocabulary items whose definitions have yet to be presented, much less
supported. This introductory chapter will go on to describe the fashioning of this
handful of sparkly stones into glittering jewelry.
10
The enticing lustre which glows just beneath the surface of these rough gems is
alluring enough to encourage not only the collection of additional data (1), but the
refinement of the two resulting gathered corpora into successive data sets (2). Some
terms will not be fully defined until chapter 2, but the tools needed to sort, classify, and
polish these sets will be laid out in this chapter (3), as will the theoretical and
terminological settings within which these tools work. After that, an overview of the
results of this analysis will be ready for display (4), and the organization of the rest of
this analysis will then be set out in the conclusion (5).
1 Prospecting the Claim: Collecting Data
This analysis targets any contemporary American English morpheme or word
in a declarative sentence that has been isolated for abnormal prominence with italics or
underlining, where abnormal or volitional prominence is defined in chapter 2. To
begin with, I sifted a new corpus of four million words in 38 books to expose 938 such
instances, where a prominent word was selected only if the material surrounding it was
not so marked. I used this OTHER CORPUS to generate some preliminary results, then
culled the conventionally recognized BROWN CORPUS (Francis and Kucera, 1961) to
form a fresh, independently-sourced data set. The Brown corpus gathers 500
American English texts samples of approximately 2,000 words each from sources
printed in 1961, where source categories range from Miscellaneous: Government &
House Organs to Belles-Lettres (cf. Francis and Kucera, 1979). Once sifted, I used
the new set of Brown corpus examples to validate and develop the framework.
11
Several methods have been developed to portray these form-meaning patterns,
but no standard exists which is familiar enough to native speakers that I could simply
generate a comprehensive set of representative examples at will, because readers of
this analysis would not be able to discern without error the same sound pattern for a
specific example as I had intended. Furthermore, while many linguistic studies are not
faced with the problem of having to convey intonation patterns in written form, this
analysis was not afforded that luxury, and even after having disallowed questions,
exclamations, and commands, the data involved presented wide variations on the basic
declarative theme, variations which could lead different readers to interpret the same
written example in different ways unless they were given other hints to guide them,
such as an extensive body of context.
One of the major consequences of dealing even indirectly with intonation was
that I could not rely upon examples which simply came off the top of my head. A
sentence like The cat is on the mat does not present enough context that I could
count on a native speaker being able to reproduce either the baseline or the abnormal
prominence pattern reliably enough that I could then go on to compare the semantics
of the two patterns in detail. A change of perceived prominence on the part of the
reader would render a change in the semantics from their perspective, which would
ruin the chance that the example would convey its intended meaning. In fact, showing
that such a small shift in form can cause a significant change in meaning is the point of
performing this analysis in the first place. It was important, then, to make sure that the
data presented here would be read with the intended prominence pattern.
12
This consideration made it necessary to cling to the most basic principles of
data gathering, starting with the precept that the strongest examples would be those
which could be recovered by another analyst, such as the reader; therefore, written
recorded instances isolated with italics or underlining seemed to be an obvious source
of stable examples. Because I wanted to minimize the misinterpretation of the
prominence patterns in the examples, I did not actually analyze any TRANSIENT data in
this analysis, since it had no recorded source and was only stored in memory; however,
I did interject generated examples sparingly into the text as brief illustrations. The
analysis proper relies solely upon WRITTEN RECORDED data, using the preliminary
results to establish the identities of a specific set of readily reproducible prominence
behaviors.
Recorded or not, searching through a vast body of data for instances of specific
linguistic behavior requires a lot of work, but it becomes easier if you establish a
background against which the salient subset stands out. The more dimensional
differences there are between the figure and the ground, and the greater the contrast
within any of these dimensions, the better. For example, a test for color blindness
would be easy if spots of a particular color were not only raised up off the page (an
additional dimension), but were raised well up off the page (high contrast within that
dimension). That is just what the search for instances of abnormal prominence is like.
The phonological features distinguishing an instance of prominence as abnormal are
iconic, so those instances then stand out from their surroundings naturally, making
them easier to find.
13
For example, any spoken word which is supposed to draw greater attention to
itself than would normally be accorded to others in the primary conversation stream is
going to be rendered in a tone of voice other than that which is used for the main flow,
as in John F. Kennedys inaugural address (January 20, 1961): Ask not what your
country can do for you; ask what you can do for your country. In addition, Kennedy
accompanied each of the prominent words not, you, and you with its own
emphatic pointing gesture, and he punctuated the words following the second you
with a trail of quick jabs, making a trail of weaker gestures which kept time with his
increasing rate of speech. Words which are audibly or visibly prominent in this manner
simply stick out from their surroundings, and the type of change in such a words
phonological form varies with the type of change in its meaning.
Likewise, a written word representing a sound which is not equal to those
around it will be printed with a font which is also not equal to those around it, and
there is a positive correlation between the way in which the sound and the font are not
equal. They are not equal to their environments in equivalent fashions, subject to the
restrictions imposed by their own modalities. For example, a loud noise is bigger than
its surround, and a capitalized font is bigger than its surround, and so a capitalized
(bigger) font is used to represent a louder (bigger) sound. It is easy to identify the word
which represents the loudest sound in this set: bang; Bang; and BANG. This has
a natural enough appeal that operators of on-line services discourage the use of all
capitals in writing electronic messages, because it comes across as aggressive
shouting, LIKE THIS. In short, written prominent words also stick out.
14
To return to the color blindness test, a problem would arise if spots were raised
not only due to closeness to the targeted quality (a particular color), but because they
merely had more of any quality greater than a baseline set of characteristics. This
would mean that brighter, larger, or more sharply resolved spots might all be raised,
simply because they were saturated with those respective qualities. This would not
only obscure the color pattern, but examples outside of the intended data set would be
prone to being tagged for analysis. Sticking out simply on the basis of more hides
targets with more of a particular quality.
Similarly, all-capital and bold fonts would be excellent markers of the data
analyzed here, were it not for an ambiguity which makes them equally good indicators
of just plain loud noise. Such words are used to represent pure shouting often enough,
and they appear rarely enough (often simply identifying technical terms), that it is not
efficient to gather them. Just as I would not collect instances of shouted material from
audio or video data, I have ignored those very few instances rendered in all capitals or
bold fonts in the written recorded material.
Italics, on the other hand, appear relatively frequently in a printed text among
regular words (average = 0.0003, standard deviation = 0.0004). I dont know when it
became standard practice to use italics to mark an aside, but that it is a contemporary
standard is not in doubt. Examples of italicized words in context were collected for
analysis, as were examples of underlined words in the one text which substituted
underlining for italics (Source [19], p. 302). None of the sources resorted to any other
alternative to italics, such as all capitals.
15
Each instance in the corpora representing an analog of spoken prominence was
collected into the database, even when it was couched in an objectionable message.
Each was gathered with enough context to ensure that a native speaker had an
excellent chance of using the same prominence pattern as that intended by the analyst.
While the main analysis of this data begins in chapter 3, the following section
describes the types of examples which were not included for analysis.
Finally, to avoid font conflicts, I will adhere to the following standards from
this point forward: 1) words which are treated specifically as objects of analysis will
be underlined in both the examples and the main text; 2) small capitals will introduce
technical terms; and 3) italics will only be used in the examples and the main text for
abnormally prominent words which are specifically not being analyzed. In examples
with more than one abnormally prominent word, the one which is specifically under
analysis will be underlined, and the rest will remain in italics. Multiple underlining
will be used later on in one small section to mark variable degrees of prominence.
2 Refining the Ore: Excluded Data
The data sets were pruned to reject only those instances which did not convey
analogs of spoken prominence, and which therefore were not relevant to this analysis.
The guidelines for this exclusion were developed while gathering data from the Other
corpus, and so it provides many of the examples illustrating this section. The numeric
labels superscripted on the Other corpus examples identify 1) a source text and 2) page
number, separated by a dot, where those sources are listed in Appendix III, p. 301.
16
To begin with, instances were passed over when the italics were used to flag a
nonconventional word or phrase, namely those expressions which were not commonly
familiar to the users of the language in which the embodying text was written:
[1-9] The [Persian] men crying love poems in an orchard on any
summers night are as often as not the lutihaw, mustachioed
toughs who spend most of their lives in and out of the local
prisons, brothels, and teahouses.
G05 0720
This included words which were foreign to any language either as nonsense (like
Winnie-the-Poohs tiddely
9.49
poems), or as having been assigned a significantly
different meaning than normal (such as where leaning
24.30
was used to describe a
type of forceful telepathic intrusion). Articulations such as grunts, groans, moans,
sighs, and other sound effects were also ruled out due to their lack of conventional
status (whoosh
13.318
, chomp
8.262
, Cchhwip Pttooey
8.344
, umph
7.405
).
Conventionality is a matter of degree, and so there are some borderline cases,
such as shtick
10.156
, Wunderkind
10.26
, apartheid
F48 1690
(circa 1961), and:
[1-10] According, then, to what I take to be the prevailing view, these
rioters were merely a handful of irresponsible, Stalinist-
corrupted provocateurs.
F42 1560
When not marked, these words are sometimes treated as conventional English, and it
is difficult to determine if their italicized versions are: 1) prominent and conventional;
2) prominent and nonconventional; or 3) non-prominent and nonconventional. This
ambiguity prompted their rejection from the data set, knowing that any such instance
could easily be integrated later once its status as conventional had been established.
17
I ruled out any instance that only acted as a label for a person (Oscar P.G.
Frank
P03 0610
), magazine (Nation
G70 1170
), vehicle, brand, caption, trademark, book,
brochure, legal case, movie, play, television show, recording or other entity. This
included the metalevel use of labels, such as when words or letters of the alphabet
addressed their own use (the one-time shibboleth of socialism
G21 1530
, the
word friendship
G49 1230
, The word death
11.60
, He is the only dog I ever knew
who could pronounce the consonant F.
2.24
).
This same guideline was extended to cover the deliberate correction of words
which shared no more than an accidental phonological similarity. In such cases, the
phonological difference between two words is emphasized, and the contextual
difference in meaning is corrected only as a side-effect, like this:
[1-11] REINER: Youve lived so long, did you ever have an accident
in all this time?
2000: An accent? Always.
REINER: An accident.
2000: Oh, an accident. Yes, in the year sixty-one, I was run over
by seven men fleeing a lion. They ran me over.
19.15
In [1-11], the only semantic tie shared by accent and accident is a pretty schematic
havability, which did not motivate the error, but merely allowed the mistake to have
been made in the first place. In other words, while it does not seem all that likely that
accident would be heard as excellent, it is not so unlikely that accidents could
be heard as excellence. If this tie were semantic, then pluralizing should not affect
their likelihood of being mistaken. Such cases seem similar to the data under analysis
here, but any contextual meaning similarities between the words are adventitious.
18
While many italicized Latin and Greek scientific terms, such as Coleus, were
ruled out, it was not because they were categorized as foreign terms, because some of
their non-italicized forms are now plain English (coleus, vitamin, gluteus maximus,
macrobiotic, acetaminophen); however, such terms were ruled out when they were
used specifically as labels (100 grams of tomato juice can contain two Drosophila
maggots
11.28
). Such scientific terms are usually, but not always, capitalized when
they are intended as standard plant and animal labels, but this distinction is obscured
when the word is capitalized at the beginning of a sentence, and so all of those
sentence-initial cases were ruled out.
Italics can pull part of a text out of the primary flow of conversation, marking
that section as dwelling in a reality aside from that occupied by the main body. In this
way, words from another source are often quoted in midstream, such as when a sign is
being read, or when thoughts are interjected (a dark look that said later
24.51
).
The anthropomorphized thoughts of nonhumans often take this form, where telepathic
invaders resort to italic assaults (You shall die,
24.111
), and animals look messages
at us (we are helpless puppies in your presence.
14.xviii
). I left out italicized
multiword sequences due to their infiltration by such examples, but good candidates
for exceptions (had they been encountered) would have been italicized verb-particle
constructions and multiword nouns.
Italics which only indicate that a person is raising their voice to be heard in
general (and not for a meaning change) were not included in the database. Such
phrases tended to be discluded anyway because they were usually not declarative:
19
[1-12] Tea, Nadine screeched.
P18 0770
[1-13] Well, said her mother, one of the pigs is a runt. Its very
small and weak, and it will never amount to anything. So your
father has decided to do away with it.
Do away with it? shrieked Fern. You mean kill it? Just
because its smaller than the others?
29.1
Fern does an awful lot of shrieking, yelling, and crying out in defense of her pig, and
so whether these instances were viewed as exclamations or interrogatives or both, they
had to be discluded. (In the tradition of William Goldman, I would briefly like to
mention that the pig does not get hurt at this or at any other time.)
Italics are commonly used to set off different mechanical properties within a
text, such as dedications (To Judith Schnerk
10.dedication
), sections, and headings for lists
and outlines:
[1-14] Air: Whipped into the white stuff to make it frothy.
Sugar and Corn Syrup: The filling is about 42% sugar by dry
weight.
11.16
Instances such as these do not reliably reflect gestural prominence in a natural way,
and so they were not included.
It should be clear, then, that instances of data were only ruled out when they
did not reflect analogs of spoken, isolated, abnormal prominence in an American
English declarative sentence, and that no potential instance of data was discarded
simply because it controverted any part of the theory being developed in this analysis.
The disallowed data simply tended either to be spoken with normal prominence levels
and placement, or they came across as if the words were quoted.
20
3 Tools: Theoretical and Terminological Background
Discourse will be characterized from two perspectives, the first of which is that
of an isolated individual reacting to its environment, and the second of which is that of
a group of individuals interacting with one another in a shared environment which
changes over time. The first set of stages brings us from sensation to communication
(3.1), and the second moves us on from communication and into language (3.2).
3.1 Discourse I: Sensation to Communication
This analysis assumes popular divisions between sensation and perception, as
well as a model of sensory transduction as an ultimately electrochemical set of events.
My analysis doesnt rely upon the accuracy of these accounts, but their familiarity
eases the transition from sensation to communication. This analysis does rest upon
previous research (Langacker, 1987: 100ff; Bolinger, 1986: ch. 6f; to begin with), in
which power and precision are portrayed as fundamental cognitive abilities evaluating
perceptual events. Discourse will be portrayed with a looped signal path, but only that
part of that path needed to illustrate TRANSDUCTION is shown here (S
o
S
i
brain):
Figure 1-4: Transduction
S
o
S
i
RECIPIENT
{
B
R
A
I
N
SIGNAL
PATH
STIMULUS
GENERATOR
21
For the moment, assume that there is a stimulus generator (or reflector) banging air
molecules together or radiating light wavicles as output signals (S
o
) over some range
of directions in its environment. Some of that energy traces a signal path intersecting a
recipients sensory organs, which then transduce that portion of the energy patterns,
including that portions power, into incoming electrochemical signals (S
i
).
The signals power can be PROXIMAL or DISTAL. Forgive my metaphor, but if
you pitch a dart at me, I can directly evaluate its impact as a proximal event simply by
waiting to feel it. This measure is precise even if I have never had any experience with
darts, because my evaluation is made after the signals reception, and is based solely
upon the energy of the impact (the articulatory energy put into the signals energy)
rather than any previous experience of my own. The same goes for any sensory signal.
I can also indirectly evaluate the strength of that same signal as a distal event
by observing its generator (pitcher, speaker, signer, or whatever). Did the gesture
appear efficient or wasteful? Did the articulation seem to be energetic? Was the motion
terse and concentrated, or sustained and broadcast? How has it felt to me when I have
made the same sort of gesture? The accuracy of this indirect measure must rely upon
my previous experience (learned or innate) with similar events, where this experience
can be as specific as my having pitched one dart at another person, or as general as my
often having tossed aside inconsequential objects. The greater the experiential overlap
between generator and receiver, the greater the likelihood that an indirect evaluation
will be accurate, and that the intended meaning will be conveyed without the signal
having to be directly powerful, which is a relatively wasteful encoding.
22
A directly powerful signal should have its origin in an efficient transfer of
articulatory energy from the gesture which generated it, because the alternative would
be enormously wasteful, namely where the power of the signal only represents a
fraction of the articulatory energy. Such a waste is not conducive to the survival of the
generator. Direct power has its advantages in primitive systems of communication,
because the message gets across even when the generator and the receiver have little
or no experience in common: signals which either overwhelm or saturate a sensory
field will be evaluated as powerful (ICONIC). To avoid waste, the transfer of articulatory
energy to signal energy has to be essentially equal. Evaluated as a proximal event, the
strength of that signals form has an equally intense meaning (again iconic), namely
that the sensory field is being overwhelmed or saturated. Shared experience with this
primitive equation promotes the development of communication systems which can
rely upon the additional evaluation of signals as distal events, systems like language.
When it comes to these more sophisticated signal systems, signal strength is not
necessarily PROPORTIONAL to meaning, but perceived articulatory effort is.
This proximal/distal distinction will be important later when experience is
defined more clearly (cf. chapter 2, 1.3). Whether interpreted as a proximal or distal
event, the sensation of a signal is a local consequence of a remote event. You dont
really see lightning way over there, but rather on your retina, and then in your brain.
Even with touch, the pressure itself, though immediate to the surface layer of the skin,
is remote from the nerves. Sensors use (electro)mechanical processes to transduce
these remote energy patterns into local electrochemical signals for further processing.
23
The next step transports the local signal to the brain through PERCEPTION:
Figure 1-5: Perception
Roughly, SENSATION will be a signals reception and the immediate transduction which
aids that reception (reaction of rods, cones, cochlear nerves, and so on). PERCEPTION
will be any transduction and transmission beyond the sensory organ which is needed to
get the electrochemical signal to the brain, plus its reception there as a PERCEPT. There
is a great deal of controversy surrounding the brains boundaries, but that is beyond the
scope of this work, and definitions that precise have no critical bearing here.
Now, what the brain does with these percepts will be a part of COGNITION:
Figure 1-6: Cognition
PERCEIVER
{
B
R
A
I
N
S
i
COGNIZER
{
B
R
A
I
N
24
In short, the actual input is assigned to perception, and the output beyond the brain is
just another type of (electrochemical) signal generation, but what happens to percepts
within the brain is what I am going to circumscribe as cognition. There is no way that I
am going to pretend to be able to describe cognition proper, but I am going to propose
two cognitive abilities that this analysis, and other research, suggests are necessary for
semantic processing. Keep in mind that a signals intensity can be encoded in proximal
or distal terms. Proximal attributes should be processed more easily than distal ones,
because they require no access to experience for evaluation; however, they should
contribute to experience afterward to facilitate processing of distal signals. In that
sense, proximal processing is a primitive precursor of distal processing.
The absolute intensity of an individual signal at a given point in time will be its
POWER, and the relative intensity of one signal compared to another from which it is
distinct (in space, time, or both) will be their PRECISION, which makes primitive
precision a measure of how closely one signal approximates another in power. The
processes that evaluate an iconic signals power (whether proximal or distal) are
essentially evaluating the power of that signals generator; likewise; those cognitive
processes which evaluate the precision between two or more such signals are
evaluating the relative power of those signals generators. This amounts to the
perception of 1) the absolute size or strength of an event, such as a perceived emotion,
color, or threat, and 2) the relative proportions of multiple events, such as perceptual
changes in general, or the relative hugeness or imminence of a threat in specific. At
this point, PROMINENCE is the evaluation of a signals intensity as high.
25
The set of cognitive functions which actually evaluate a signals power will be
treated here as if they were a single entity, which will also be called POWER; likewise,
the set of functions which evaluates relative power will be treated as a unitary entity
called PRECISION (cf. comparison; Langacker, 1977: 100). In their most primitive
forms, power and precision are essential functions across chordate behavior, whether
or not vermiforms (for example) are actually suggested as providing the evolutionary
source for these functions in human cognition.
Now take a look at all of this from a generators perspective, where that
generator will soon come to be discussed as it represents a speaker or signer:
Figure 1-7: Articulation
Having processed the percept, the brain can send signals to the articulators. The action
of articulation in the diagram above begins when the internal electrochemical signal
leaves the brain, and continues as it becomes bodily motion, which essentially releases
another signal into the outside world (S
o
), ready to be picked up by something else
(S
i
).
S
o
S
i
GENERATOR
{
B
R
A
I
N
SIGNAL
PATH
26
Properties of intensity as conceived by the generator can be encoded iconically
into these signals, which can then be reflected in the motion of the articulators when
rendering that signal. This allows for a portrayal of a person as sensing something
huge or intense in their environment, perceiving it, and then articulating a signal in
response which is also large or intense. The reception of this subsequent signal forms
the overlap which closes the coil where it was originally opened above at transduction.
The only reason that this is a coil instead of a plain circular loop is because I am
casually associating each generator with the configuration of a human head, so the
articulators are below the receivers, and because I am choosing two participants.
Having set all of this up, here is what I think of as discourse:
Figure 1-8: Discourse
In summary, some participant begins things by taking on the role of generator, drawing
in perceptions from the outside and adding them to their current set of conceptions,
then evaluating this conceptual set and articulating an audible and/or visible gesture.
S
o
S
o
S
i
S
i
PARTICIPANT
{
B
R
A
I
N
SIGNAL
PATH
PARTICIPANT
(ENVIRONMENTAL
STIMULI)
27
This expression adds to the environmental input for the next participant, and so on.
During this discourse, a signal can be perceived that contains iconically encoded
information that signals prominence. This high intensity can be transferred to the
physical form of the electrochemical signal sent to the articulators, and thus
prominence in a string of words can have come iconically and proportionally from the
largeness of an environmental signal generator. Note that the level of the prominence
can then be encoded iconically and proportionally into the signal which is to be
received by the next listener or looker.
Cognitive abilities such as power and precision are developed in an
environment which presses the specific need to evaluate the intensity of perceptions.
To the degree that communication itself becomes vital for a species in that
environment, developing the additional ability to communicate or externalize
conceptions or perceptions of prominence (as articulation) would present a great
advantage over only being able to process incoming perceptions of prominence. A
species has a distinct advantage if it is composed of individuals which can not only
perceive danger, but which can communicate the imminence of that danger to other
individuals. Given the iconic nature of these signals, and the universality of the
behavior in reaction to these signals (fight/flight), some universality across systems of
communication and language should also be expected.
The value of language universal gestures, whether visible (stabbing) or audible
(explosion sound-effects), is not readily apparent when removed from a primitive
environment, but it is clearly valuable at a communication universal level. Bigness and
28
loudness convey threat and authority, and power (often sheer size) is expressed visibly
or audibly with such ferocity that even insects employ such pure representations of
threat. For example, Ohala (1983: 7) supports the existence a cross-species frequency
code which equates an emitters smallness with high pitch, and its largeness or threat
with low pitch and loudness. My work suggests that these primitive prominence
gestures are the earliest of those used for communication. Given our personal,
continual experience with prominence, it seems unreasonable to suggest otherwise
than that it was beneficial to adapt an internal resource, namely the cognitive
evaluation (or articulation) of prominence to serve an external purpose: the
communication or physical articulation of prominence. The trick is to clarify this
internal-to-external (reception-to-production) adaptation, and to make the way in
which a cognitive ability is a resource seem natural.
The adaptation of external resources is familiar. New civilizations cannibalize
their ancestors, and then leave ruins providing little more than a nice flat place upon
which to rebuild. Physical and cultural resources such as old tires and folktales are
repeatedly adapted to fit places and purposes for which they were not originally
intended: carved stones from the colonnade in the Temple of Artemis are now part of a
functioning aqueduct some distance away; pagan traditions have been co-opted by
Christian holidays; an ancient check for weapons is our handshake; content verbs are a
source of modals; and muddy main streets become paved Main Streets. This analysis
supports a scenario in which the primitive cognitive abilities of power and precision
were adapted in a similar way, externalizing the internal articulation of prominence.
29
This introduction will continue to trace the path of this adaptation, crossing the
threshold from communication to the initial linguistic use of power and precision. This
portrayal is supported by current research tied to cognitive function (and sensation and
perception) promoting the candidacy of iconic or indexical gestures as the earliest
units forming the foundation of subsequent linguistic evolution (Armstrong, Stokoe,
and Wilcox; 1995). Previous research on the language functions of normal levels of
prominence are covered in chapter 2, and the remainder of the dissertation houses the
analysis proper, the results of which reveal the systematic application of power and
precision to create a pair of prominence functions in language.
This analysis of abnormal prominence is a bridge between previous research
on the primitive communication functions of intense prominence and the linguistic
functions of the weaker or normal forms of contemporary prominence; I will use it to
forge the final link in a chain of adaptation, a path starting at the primitive end with the
transduction of sensations of intensity, and terminating at the contemporary end with
the most strongly grammaticized portrayal of weak prominence, namely syntactically
predictable normal sentence stress (as described in chapter 2).
Over the last sixty years, linguistic analyses have uncovered the stages of the
adaptation of prominence in an order running counter to the sequence in which these
functions were originally developed. This is a normal, if not necessary, pattern in the
analysis of any adapted resource. For example, nine major strata of Troy have been
excavated out from underneath Hissarlik in Turkey, and you have to drop down over
15 meters (and 5,000 years) before finally hitting the remains of the Bronze Age
30
fortress at the bottom. It is not uncommon for more mysteries to be uncovered than are
solved until researchers make the origin of a resource accessible, at which time the
questions raised by previous analyses of that resource can be answered by the forward
propagation of the new evidence through the chain of its adaptation. This is simply
another way of looking at the chain of adaptation as being enervated.
The obstacles for prominence are not as strictly layered as this, and so a more
appropriate metaphor would be the alignment of windows of opportunity. All of the
following propositions must be regarded as only mildly offensive for the path to be
clear: 1) mechanism is able to tolerate mentalism; 2) competence can appreciate
performance; 3) cognition and scientific disciplines can both thrive when distributed;
and 4) primitive behavior is shared behavior, both among species and among forms of
communication. These windows have been open simultaneously for less than ten years
(cf. chapter 2), and so the timing of this analysis is no coincidence.
3.2 Discourse II: Expectation and Revelation
The previous section describes discourse primarily as it functions during the
articulation of signals that are immediate to the cognizer both in space and time, where
those signals are usually environmental in origin. This section will incorporate the
cognizers memories by describing discourse over space and time, including beliefs,
portraying language as an activity shared among many cognizers. In general, the
discussion will begin with the portrayal of one cognizer, and work toward the
description of several of them acting in concert.
31
This analysis rests on a portrayal of language use and its users pieced together
from Langackers work on discourse (1991: 3.1.1), Fauconniers study of reference
(1985), Gunters material on context and prominence (1974; cf. chapter 2, 3.2), and
analogs in other approaches. Lakoff (1987) is generally my source for ideas about
idealized cognitive models (ch. 4) and radial categorization (ch. 6). I also spent a lot of
time with Hawkins (1978), and the way that I approach (in)definiteness is due in large
part to his analyses, as well as Langackers applications of them. All of these facilitate
the description of cognizers working in concert.
To begin with, while any given person (P) is experiencing the world or reality
with which we are all supposed to be familiar (R), that person will update their own
individual set of EXPECTATIONS about reality in accordance with those experiences
(R
P
). Those things which remain stable in that persons experience are those ideas
about reality which will rarely get updated, and the longer that something meets their
expectations, the less likely they are to update that portion of their personal view of
reality on the basis of minor disappointments. Something which is expected to never
happen might, in fact, happen; it just might not happen often enough to change that
persons expectations about the likelihood of its occurrence. This phenomenon is
known in psychology as per-SE-verence (Krull, pc).
Such stable expectations are described by that persons IDEALIZED COGNITIVE
MODELS (ICM). It is important to emphasize the fact that stable does not mean highly
likely to occur. For example, there are conventional notions insisting that anyone can
win the lottery, or that anyone can become President of the United States. These ideals
32
are maintained as stable despite the low expectation of seeing them realized, and they
can be supported right alongside a high expectation that they will not be realized. In
fact, that they are not realized is yet another ICM.
A persons set of ICMs is their individual ARCHIVE (small a), and it is the store
of those things about reality which that person holds to be highly resistant to change.
Their view of reality represents the NONPHYSICAL CONTEXT of their conversations, and
their archive is the stable subset of that mental reality model. Since any given person
should expect things from their individual archive to be given, obviously true, or
not open to question, their archive is a location to which a person readily refers
during a conversation:
Figure 1-9: The Archive (Atemporal Model)
The Archive
(archival core):
ICM
#n
ICM
#2
ICM
#1
R
Pn
R
P2
R
P1
Reality (R)
individual
archives:
1 2 n
33
I will give evidence below which suggests that this intuition about archival knowledge
being given is so strong that each of the participants in a conversation (P1, P2 Pn)
will expect their individual set of archival knowledge (R
P1
, R
P2
R
Pn
) to be shared
wholesale by all of the other participants (R
P1
= R
P2
= = R
Pn
). Of course, this is a
false consensus because their individual archives will necessarily intersect much less
closely than any of them expects (R
P1
R
P2
R
Pn
).
It follows that ICMs should be similar from person to person since they are
building expectations about what the same reality will be like. Different people should
find a core of similar things to be stable about reality: gravity keeps you on the ground;
other drivers are idiots; and dogs are great. ICMs dealing more with discourse tell you
that when you want to refer to something, you point to it somehow, or you point in the
direction of its location. These are all intuitions that will hold from person to person;
however, to the extent that their archival knowledge is different, it becomes necessary
to distinguish between any given persons individual archive and the core of archival
knowledge shared by any given set of people. This core is the ARCHIVE (big A), and it
is naturally the material in the intersection of all of the individual personal archives.
A person expects their experience of their PHYSICAL CONTEXT or WORLD to be
shared by others: if someone feels wind, they expect others to feel wind; if they see a
tree, they expect it to be seen; if they taste something sweet, or feel something soft
ad somnium. Like the intuitions mentioned above, these sensations are supposed to be
equivalent from person to person, but there will be degrees of difference arising from
such pragmatic factors as people facing different directions, or from a person not using
34
one or more of their sensory modalities (due to nasal congestion, blindness, and the
like), and so a distinction must be made between any given persons experience of
their world (small w) and the experiences held in common by any set of people
grouped together in the same WORLD (big W).
Just as there are central and peripheral densities in our perception of the world,
the archive has areas distinguished by their accessibility. Chafe (1973) first describes
this viscosity in terms of a familiar depth-continuum, defining surface, shallow, and
deep memory; later on (1994), he modifies this portrayal to align better with his newer
descriptions of active, semi-active, and inactive states of consciousness (ch. 5). In both
works, Chafe shows that there are normal sentence structures and intonations designed
to plumb each of these three depths or activation states.
Chafes earlier work suggests that specific prominence patterns can be used to
reach particular depths of a cognizers memory, and he says in the later work (ch. 6)
that accessing less active information exacts a greater activation cost, and prominence
is an expression of that cost. Not only would activation of ones own inert memories
(as experience for distal events) be costly, but so would activating any equally inert
memories in someone else through the energetic articulation of prominence. The
results of my analysis suggest that this cost is in direct proportion with memory or
inertial depth. Chapter 2 portrays the coin of the realm as quanta of energy, which
intense prominence spends in an iconic expression of energy. One cognizer can even
pay the cost of a proximal event, activating deep or inert material in another cognizers
mind, which brings us to the discussion of cognizers working together as a group.
35
When people converse, the roles of SPEAKER or SIGNER (S) and LISTENER or
LOOKER (L) will usually be passed around from person to person, and a temporary
reality model or DISCOURSE SPACE will be created for the purposes of conversation. The
Archive (the intersection of the nonphysical contexts held by the participants) plus the
World (the intersection of their physical contexts), plus the information from earlier
stages of the discourse, all add together to make the CURRENT DISCOURSE SPACE (CDS).
The CDS is updated frequently, on a faster than per-clause basis, and the most
recently updated portion of the CDS is called the IMMEDIATE DISCOURSE SPACE (IDS).
The construction of the INITIAL STATE of the CDS begins before any words have been
spoken. It is the state of the CDS which exists before any conversant tries to update it
by taking on the role of S and drawing something into it from the surrounding context,
or by referring to any instance which already exists in the initial state of the CDS.
In short, the Archive plus the World equals the initial state of the CDS. In long,
the physical context includes anything of which all of the conversants share a current
sensory perception in their physical surroundings, and the nonphysical context is the
intersection of all of the conversants personal views of reality, so the initial state of
the CDS includes anything in the physical context which provides sensory experience
for all of the conversants (warm sun, rough floor), their perceptions of this stimuli,
plus the core of archival knowledge shared by all of the conversants. The initial state
of the CDS is the set of all entities (types, instances) of which the conversants are
already mutually aware just prior to the beginning of any discourse, and so those
entities do not need to be drawn into the IDS by any form of elaboration.
36
This temporary model is any given persons CDS, and it lives for the length of
the conversation. Since no two people actually use the same brain, each person must
have its own CDS, even though people behave as if they were manipulating only one
CDS between them. As stated earlier, this feeling of being in consensus about shared
information is strong, but always at least slightly false; however, to the extent that the
people are supposedly building a shared body of information between them, I will
refer to the CDS as if it were one entity, rather than as if it were a set of multiple,
overlapping, near-identical copies, with one copy being distributed to each person.
Such a view does not change this analysis, it merely makes the CDS much easier to
discuss.
Given this definition of the CDS, it is surrounded by (and is a subset of) its
CONTEXT, which is the union of its physical and nonphysical contexts. The set of
information known commonly by all of the members of the group (the intersection or
Archive) is a subset of the sum of the information known by all of the members of the
group taken as a whole (the union). S can update the CDS by drawing new material
into the IDS from the context of the discourse, including material drawn recursively
from older stages of the CDS which have fallen farther away from proximal time as
the CDS is updated. This suggests that there is also a temporal component to the
discourses context, parallel in structure and function to the physical and nonphysical
contexts, but exploring this possibility will have to wait for future research. In any
event, some of this new material has the potential to update the Archive in general, and
therefore the archive of any P involved in the discourse.
37
This is all going to have a direct bearing on the characterization of the
prominence function which will come to be called REVELATION, which comes into play
when one P is disturbed by a perceived mismatch among R
P
s. When it is Ps turn at
being S, P will point out the error by making a correction which it thinks will realign
the R
P
s to its satisfaction. That correction will be rendered as an utterance containing
an abnormally prominent word. The greater the degree of the error, or the greater its
inertia, the greater the intensity of the prominence. In other words, a small mismatch
outside the archive might not even warrant correction, but an error within the archive
will usually require abnormal prominence of primary intensity, and a misalignment
within the Archive will probably elicit intensely abnormal prominence. Variations on
this behavior will be discussed in chapter 4, and a number of preliminary examples
will be given in 4 of this chapter.
3.3 Discourse Summary
I would like to describe how the results of this analysis are distributed across
the claims that I have made up to this point. Although the direct iconic proportion
holding between the form and meaning of linguistic prominence is supported by the
results of this analysis, the gradation of that proportion is not as finely measured as it
could be with audio or video data. As this analysis continues, I will demonstrate that
there are two linguistic functions of abnormal prominence, namely ELABORATION and
REVELATION, a division which is also supported directly by the results of this analysis,
and which is reinforced by a straightforward process of simple reasoning.
38
In this section, I also talked about two cognitive functions, namely power and
precision, and both of them have their existence supported by additional research
outside of this analysis. I will show that their interaction supports the division between
elaboration and revelation, and that they can be used to systematically organize a set of
subfunctions for elaboration. The infrastructure of revelation will be defined later
according to components derived from the definition of discourse. Finally, I would
like to observe that this system as a whole hangs together in an entirely comfortable
and consistent fashion.
4 Polishing the Stones: Display of Analytic Results
I now offer you a preview of the results of this analysis. The following
functions will be illustrated with data selected from the Other corpus, because I am
saving the Brown corpus instances for the analysis proper. First, I will guide you
through an exhibition of elaboration (4.1), then move on to revelation (4.2), and
follow up with the functions of linked instances of abnormal prominence (4.3).
4.1 Elaboration
L interprets ELABORATION as singling out meaning variations whose semantic
intensity most closely equals the greater phonological intensity used on a word by S,
where that equation is a direct iconic proportion. In essence, L applies that greater
semantic intensity to the words meaning in an amount equal to the perceived formal
intensity. This meaning change is word-internal, as if you were to change your shirt
with a pair of scissors. As a change in intensity, it naturally lends itself to two
39
variations, namely elaboration for power, or elaboration for precision. The first type of
change simply uses the prominences POWER to represent a similar force added to the
meaning of the word. The second type of change applies additional energy as
PRECISION to control the degree of tolerance normally associated with the meaning of
that word. These two linguistic subfunctions are named so as to explicitly associate
them with the cognitive functions from which they are posited to have been adapted.
There are actually two kinds of power, the first of which directly reflects the
inherent intensity of the size, speed, or other characteristic of the prominent word:
[1-15] REINER: Im a little queasy about this, telling tales about
Presidents and Presidents wives.
2000: Theyre all a little power-crazy, right? And they love to
do it. Lets face it. They love it. They love it. They LOVE it.
REINER: Yes.

2000: Let them have a mistress or two. Let em, cause they
have to do it.
19.89
In both of these cases, the additional intensity of the prominence is associated with an
increased energetic intensity in the prominent processes, where love is stronger than
love (but not as powerful as LOVE), and have is obsessive when compared to
have. Note that these are both changes in meaning internal to the word, and not
external changes, such as having love swapped out for hate, or for any other
process. That would be a revelation instead, which is analyzed in section 4.2.
Power can also be added as an adventitious reflection of the intensity of Ss
emotional discharge, rather than any powerful characteristic inherent in the word
itself. Only five examples exist in the Other corpus, and seven in the Brown corpus:
40
[1-16] Trouble with women like you is you always want to argue.
4.166
[1-17] 2000: when I was a kid, when I got married, it was in caves.
Caves. Caves.
19.80
A confrontation can be heated and intense, but in these examples prominence displays
the intensity of Ss emotions, and not the ferocity of the tiff. Inherent and emotional
power overlap, and so it is better to think of them as polar rather than as discrete.
Energy does not always explode, but rather sometimes it allows for a greater
control over another process. Words are normally used with a range of tolerance in
their meanings, such as where the word none is understood to mean none except for
a couple of insignificant exceptions. Precision removes this tolerance, such that
none means absolutely none, with no exceptions whatsoever. This next example
shows how this INTOLERANCE affects the boundaries of all:
[1-18] Alright, fine. Here.
Chomp.
Could I try one more, please?
You said that was the last one.
Well, I made a mistake. Can I have that one? The one going in
your mouth? Thats the one I really want.
Youre sure this time?
Yes. Thats the one that will satisfy my curiosity about all
potato chips. I swear it this time.
Its not the potato chip he wants. He just wants to know that he
can have another potato chip afterward.
8.263
In [1-18], all is not used in simple contrast with some or none, neither is it set
against some discretely parcelled measure of all, but rather prominence has removed
any tolerance that all normally displays toward exceptions, and so now all means
41
absolutely each and every last one of them. It is clear from the context (I swear it
this time) that in other cases when the dog has begged for chips, all was allowed to
mean almost all (as in all chips eaten right here and now). The greater the reduction
of the tolerance, the greater the phonological intensity of the abnormal prominence.
When a words meaning variations are treated as a set, precision can identify a
POSITION within that set or TYPE as special. Those instances precisely in the center are
good, typical, PROPER members of the type (some of which are PARAGONS), and those
on the border are badly atypical PERIPHERAL members. One of my favorite examples of
abnormal prominence marks a noun as a proper member of its type:
[1-19] The unwritten manifesto of this revolution states that the Negro,
backed by a number of whites in every section of the land, is
finished with being classed as not quite human; that he is no
longer humble and patientand unlettered; and that an
astonishingly large group of Negro scholars and journalists and
artists are expressing their resolution with courage and skills.
They are no longer colored people. They are people.
5.150
The intent of the abnormal prominence is to emphasize the homogeneity implied by
the word people, and thus to make the centrality of this particular instance of the
word readily apparent, but one of the strongest effects is to make the eradication of the
nouns modifier absolutely clear. There is no longer any modifier drawing this instance
of people away from being anything other than a good, absolutely typical member of
its type. People represents a high standard which has been met.
Of the two prominent words in this next example, wrong is being treated as
peripheral by S, and fit is a negative use of prominence for proper position:
42
[1-20] and something about the way she held hands was just
wrong. Our fingers didnt line up right.
8.146

I think that what she did was slide her fingers in too early so
they were all out of sync with mine.

We just didnt fit.

In the case of wrong, S is trying to find a phonological form which matches up with
the meaning he (Riser) wants to convey (hence the pause before the word), but that
meaning is just too strange for words, so S compromises. S would like to find a form
that has conventional ties to the meaning he has in mind, but he finally gives up and
chooses something close. That mismatch, and Ss evident feeling of discomfort over
the hand-holding, add together to get wrong interpreted as peripheral under
abnormal prominence. In contrast to the positive use of proper position with people,
in which a standard was met, didnt fit shows the effects of negation (didnt) on
the use of proper position, where S holds up the proper instance as a standard which is
not being met.
All of these subfunctions of elaboration involve Ls interpretation of greater
phonological prominence as providing the additional energy necessary to create or
select an internal change in a words meaning. This energy is construed as being in
direct iconic proportion to that which is needed to augment the power inherent in a
words meaning or in Ss emotional charge, or to exert tighter control over a words
range of tolerance, or to locate an instance within its type. The experimental evidence
in support of this proportion will be reviewed in detail over the course of chapter 2.
43
4.2 Revelation
REVELATION is a contextual or word-external meaning change, so you just put
on a new shirt rather than tailor the old one. The three types of revelation vary with the
kind of mismatch in R
P
s that S wants to correct. When S can identify the error with a
single word elsewhere in the discourse, S uses an abnormally prominent word called a
CHANGE as a replacement for the error. This error is called the changes COUNTERPART.
In SUBSTITUTION, this change is just one word, which is completely ENCAPSULATED and
EXPLICIT. The portion of the discourse against which this error stands out so glaringly
is called the SETTING. When this counterpart is an actual gap in a setting, S fills that gap
with a change, making an ADDITION. When the counterpart is DISTRIBUTED across more
than one word, or when it is less than fully explicit, then S suggests a change arrived at
by a process of DERIVATION.
In a substitution, there is an explicit, encapsulated (unitary) counterpart in the
discourse setting which gets corrected by the abnormally prominent change:
[1-21] sometimes its not a beautiful baby. And I cant lie. I tell them
right to their face, No, thats a monkey youve got there.
8.331
[1-22] All this requires that the phone be modified. According to
Caristi, it would be difficult, but probably not impossible, to
detect room noises via the ringer circuit of an unmodified phone
on the cradle, too.
11.87
The discourse counterpart for monkey is unitary and explicit, namely baby, and
for unmodified it is modified. In each case, the counterpart is representative of
what S thought that L held to be true, which did not match up with Ss own beliefs.
44
When the context has the prominent word filling in a specific gap in the
setting, then that gap is referred as a NULL COUNTERPART, and the process is one of
addition:
[1-23] The chief task of a librarian is to get people to read.
You feel so? My feeling, Mrs. Kennicott, and I am merely
quoting the librarian of a very large college, is that the first duty
of a conscientious librarian is to preserve the books.
4.93
[1-24] Look, John, everbody takes a crack at me cause I been a
preacher. A preacher aint nothin but a man.
Yeah, buthesa kind of man, else he wouldn be a
preacher.
7.305
S is not using abnormal prominence here to refute an implication of unconscientious,
careless, irresponsible, or any other notion with which conscientious might form
a conventional pair, but rather it is specifically adding a modification which had been
missing altogether, and thus is filling in a perceived gap. Similarly, S interposes kind
where no modification had gone before, making the gap itself obvious. There arent a
lot of examples of addition in the corpora, but they are distinct in their behavior.
Derivation is used when the counterpart is more than one word, or is less than
wholly explicit, but is not null. The independence of derivation as a revelation subtype
is supported by Princes (1981) INFERRED information, and by Needham (1990), who
shows that the mention of an object in the background PRIMES information about its
parts and qualities. This lack of a single, easily identifiable counterpart means having
to ferret one out of the background, often by appealing to associations with the change,
such as conventional or contextual pairings, or conceptual domains.
45
Sometimes an explicit counterpart is not unitary because it appears as a list:
[1-25] But a hundred and fifty years ago nobody judged a river on its
hydroelectric output, or even perhaps on scenic beauty. It was a
wayor an obstacle.
38.244
This is a case of way versus both output and beauty. Sometimes the counterpart
is unitary, but not explicit, such as when an implicit reference is made to the other
member of a familiar conventional pair, as in [words vs. pictures]:
[1-26] Victorias Secret is big trouble. Thats a good-looking
catalogue.
In all fairness, its more than a catalogue. Its a lovely story, a
novella, really, that I keep at my bedside, and every night I read
a few pages and see what those wacky girls are up to this time.
There are no words, of course, but you can put together the
story.
8.313
Such distributed counterparts still characterize Ss perception of Ls beliefs, with S
revealing those beliefs to be wrong, as backed up by a forceful expression of authority.
When it comes to revelation, then, three subfunctions are distinguished by the
explicitness and encapsulation of their counterparts. If the counterpart is unitary and
explicit, the subfunction is substitution. If the counterpart seems to be a gap in the
setting, then the subfunction is addition. If the counterpart is distributed, then the
subfunction is derivation. The changes in the form and the meaning of the marked
word rely upon the prominence being abnormal, just as they did in elaboration, but
because S uses revelation to correct an error in the discourse, the iconic relation is
proportional to the strength that S needs to promote its correction.
46
4.3 Linked Instances of Abnormal Prominence
Sometimes instances of abnormal prominence occur near one another, but this
is simply coincidental (CLOSE INSTANCES OF ABNORMAL PROMINENCE) unless context and
timing conspire to relate them (LINKED INSTANCES OF ABNORMAL PROMINENCE). Context
can dissociate or coordinate prominent words, an effect which can be augmented by
the rhythmic patterns arising from the use of parallel structures.
DISSOCIATION occurs in a context in which prominent instances are pushed
apart, where a multiple substitution establishes what amount to taxonomic labels:
[1-27] Soil consists of mineral particles, organic mattersome living,
like soil bacteria, and some dead and in the process of decay
plus air and water in the pore spaces between the particles. The
shapes, sizes, and relative proportions of the different mineral
particles determine the type of soil; the quantity of organic
matter influences the soils quality. Soils, then, are almost
infinitely variable. Nevertheless, gardeners speak of three
general typesclay, sand, and loam.
12.74
Dissociation emphasizes the distinct differences between the normal meanings of two
or more words, solidifying the borders which are used to divide a domain, which in
this case is the domain of soil. Special word-internal meanings are not being drawn
out by prominence, so this is not a form of elaboration, but rather the normal meanings
are serving a special purpose in a prominent context, which is revelation.
A previously defined taxonomy can be replaced by its prominent reflection, as
in the following:
[1-28] Im the one unable to live without you.
No, Im totally unable to live without you, I swear.
8.304
47
This reference swap is [Im (ref. x) you (ref. y) vs. Im (ref. y) you (ref. x)],
which is a change of reference without a change of grammatical form.
COORDINATION occurs in a context which portrays a multiple substitution as if it
were tying words together, rather than forcing them apart:
[1-29] For years theyve been promising us phones where you can see
who youre talking to. I think theyre putting it off because if
people could see you, you wouldnt be able to lie anymore. You
cant say, Oh, I was just leaving. They see youre in your
pajamas, they know youre not leaving.
8.322
Coordination stresses what the normal meanings of the two words have in common,
which goes on to promote a causal or consequential link unseen in simple contrasts. In
this example, knowing is a consequence of seeing, and the connection between the
two is given a gentle boost by the rhythmic timing of their shared prominence.
Each of the two corpora holds one distinct example of TIMING (plus a number of
mildly close cousins), in which parallel structures organize the prominent points which
they surround into a rhythmic pattern, augmenting the dissociation or the coordination
specified by the context. The following example displays the clearest case of timing in
the Other corpus, specifically a dissociation augmented by timing:
[1-30] I began to think more about what I felt and feel more about
what I thought.
37.72
The sing-song, rhythmic quality is evident, and this pattern helps to emphasize the
dissociation between thinking and feeling. The Brown corpus example holds a
five-instance rhythmic coordination, and it will be analyzed in chapter 5.
48
Notice that these sequences of linked prominent words do not determine any
new functions, but are essentially contextual variations of revelation. In other words,
the exact same sequence of abnormally prominent words could either be dissociated or
coordinated depending upon the initial barriers or links forged by the context (when it
does not leave them as mere coincidence). Either way, the strength of the link between
these instances can be augmented by the rhythmic timing of the prominent beats in any
parallel structures which might surround them.
4.4 Summary: The Functions of Abnormal Prominence
Everything comes down to two linguistic functions for abnormal prominence,
namely elaboration and revelation. Elaboration signals a word-internal meaning
change, where the phonological intensity of the additional prominence is interpreted as
being in direct iconic proportion to an increase in the semantic intensity of the changed
meaning, either in terms of power or precision. Revelation makes a contextual or
word-external meaning change, specifically a correction in a mismatch between reality
models as perceived by S. The type of revelation S uses depends upon the explicitness
and encapsulation of the counterpart to that change. An explicit, thoroughly
encapsulated (unitary) counterpart is exchanged in a substitution. A null counterpart is
filled in an addition. Finding a more broadly distributed counterpart requires more
work, namely derivation. In the case of revelation, the phonological intensity of the
abnormal prominence is interpreted to be in direct iconic proportion to the effort
expended by S in promoting the correction.
49
5 Conclusion
These functions are all directly supported by this analysis of American English
morphemes and words in declarative sentences which have been isolated for abnormal
prominence with italics or underlining. The existence of these functions, the detailed
description of the systems pervading their behavior, and the existence of a direct iconic
proportion between form and meaning, all of these have been derived from a thorough
analysis of these abnormally prominent words. Much of this data has been dismissed
for decades as trivial matters of contrast and emphasis, but just look at how much is
going on in the data sifted from just two corpora, and this isnt even the main analysis.
Chapter 2 synthesizes firm definitions for normal or routine prominence and
abnormal or volitional prominence from a thorough review of previous research on
prominence anatomy, physiology, patterns, and functions. It will provide support for
the portrayal of these two linguistic functions as having been adapted from
communication functions, which had been adapted from cognitive functions designed
to evaluate sensations of intensity in a physical environment. Chapter 3 is the first part
of the analysis proper, and it lays out all of the details concerning elaboration;
likewise, chapter 4 goes on at length about revelation. Chapter 5 is used to display
these functions as applied to linked instances of volitional prominence. Finally,
chapter 6 maps out my plans for performing further research on data recorded in an
audiotaped or videotaped format rather than a written one, as well as the eventual
extension of this project to encompass cross-modal volitional prominence, all of which
will include the analysis of increasingly intricate prominence patterns.
50
CHAPTER 2
The Difference between Routine and Volitional Prominence
You can bear a little more light?
I must bear it, if you let it in. (Laying the palest shadow of a stress
upon the second word.)
Charles Dickens, A Tale of Two Cities
One of the first things that I would like to do is suggest alternatives to loaded
terms like normal and contrastive in the interests of setting aside their implication
that willful control of ones language is in some way a display of degenerate behavior.
Even neutral pairs like typical and paratypic attribute far too little frequency to the
amount of time that conversants spend ensuring that they have really been understood
as well as they would like. It turns out that prominence behavior can most usefully be
draped along a continuum which reflects the deliberate, conscious, VOLITIONAL control
a user exercises over its language at one pole, and at the other describes how deeply a
user can surrender that control to a variably mechanical, cognitive default or ROUTINE.
All of the language behavior along this continuum is human, and so normal, and all of
it occurs frequently enough to be typical at least in quantity, and close enough in
proportion that to call any of it atypical would be misleading.
Volitional prominence is the type S uses when it is concerned that it will not be
understood correctly without resorting to effortful measures to ensure that it is making
sense. S draws particular attention to those parts of its utterance which hold the key to
51
aligning Ls model of reality with its own. This should sound familiar from the digest
of this analysis given in the introductory chapter, and I will give two more examples of
volitional prominence here, one each for elaboration and revelation:
[2-1] They have some of the best papers, and current-events
discussions so interesting.
4.108
[2-2] And see, my friend, that you make your house a home. A house
is a mere skeleton of bricks, lath, plaster, and wood; a home is a
residence not merely of the body, but of the heart.
16.209
Just to make sure that the function of volitional prominence is clear in each case, I will
point out that the strong phonological forms of best and so are interpreted by L as
selecting word-internal meaning variations which can be construed with equally great
power. The substitution of home for house in [2-2] is detailed in the subsequent
text, and the additional articulatory force reflects the strength of Ss admonition, lifting
the change up out of its surroundings and into the path of Ls direct attention.
Volitional prominence, then, brings no new articulators to the communicative
act; it just uses the same set of routine articulators with greater power and precision.
Some of the research reviewed in this chapter has measured this effort as an increase in
the expenditure of energy quanta, down to the cost of the motor units recruited for the
fortified motion of the standard articulators. The reader can experience this effortful
articulation directly by acting out the written data, or indirectly by reconstructing that
data empathetically, since they represent distal events. This personal experience allows
for a much more finely graded set of judgments about the proportions of phonological
and semantic prominence involved than would be granted by the italic font alone.
52
Now lets take a look at these same examples without their italics, as if S were
blithely assuming a well-aligned interpretation on Ls part:
[2-3] (a) Thy have some of the bst papers, and current-evnts
discussions s interesting.
4.108
(b) Thy have some of the best ppers, nd current-events
discussions so nteresting.
4.108
[2-4] And se, my friend, that you mke your huse a hme. A huse
is a mere skleton of brcks, lth, plster, and wod; a hme is a
rsidence not merely of the bdy, but of the hart.
16.209
To begin with, note that the prominence placements and levels vary between [2-3](a)
and (b) despite the identity of their syntactic structures, so prominence is not utterly
routine, as in being mechanically predicted according to syntax. In fact, the notion of
this type of prominence being routine is most evident when you watch how S directs
the focus of a conversation. This tracking shows up most plainly in [2-4], because it
was used earlier for a revelation, and this would only have occurred in an environment
which was conducive to prominence identifying a contextual meaning difference. The
standard strength of the focal prominence, and the pattern of subsidiary prominences
dependent upon those points: those are matters of routine. S can rely on this routine
filling in the details even while S is preoccupied with another task. No effort has to
be put toward determining just how not-strong the phonological form of the prominent
word is going to be, because the lack of real intensity is only broadly reflective of the
strength that L attributes to the construal of that words meaning.
When S yields all volitional control over prominence to routine, its placement
and level fall prey to the rhythmic patterning to which all human activity is susceptible:
53
[2-5] And se, my friend, that you mke your huse a hme. A huse
is a mere skleton of brcks, lth, plster, and wod; a hme is a
rsidence not merely of the bdy, but of the hart.
16.209
[2-6] (a) You walked n and sat dwn n the couch.
(b) I just hte when my teth bte my fork.
(c) Constantne, you will boge til dawn!
[2-7] How knd of you to let me come.
Note the familiarity of the sing-song pattern in the list of materials in [2-5], and the
ease with which [2-6](a), (b), and (c) can all be chained to the same rhythmic pattern
(da-da-DAA da-da DA-DA-Da-daa). If you gave [2-7] a Lerner-and-Lowe intonation
(My Fair Lady), then you surrendered to rhythm (daa-DAA Dada daa.daa.daa.daa).
Some of these rhythms (question, command), or perhaps the rules for their automatic
generation, are so common that infants communicate quite capably with them before
and without words. Rhythm, then, is at the opposite pole from volitional prominence.
These definitions are the result of reviewing over 400 years worth of research
into English prominence, where those antecedent analyses addressed: 1) what it is; 2)
where it goes; and 3) what it does. This chapter is divided by this taxonomy. What
prominence IS embodies the anatomy, production, and perception of prominence (1).
Where prominence GOES holds notational schemes for draping intonation contours
over an utterance according to its sense, music, syntax, or metrical structure (2).
What prominence DOES gathers analyses which have identified the function of
prominence over the years as targeting interest, information, discourse changes, theme
and rheme, and finally inward and outward focus (3). The cumulative outcome of this
massive collective effort will be distilled at the end of this chapter (4).
54
1 What prominence IS
Lungs, tongues, lips, throat... these speech organs display readily observable
mechanical behaviors that are reflected in the functions of familiar tools like a bellows
or woodwind. The buzzing of the voicebox can be felt right there at ones fingertips.
Phonologists might consider this access to be a mixed blessing, since it lends itself to
the proliferation of the same sort of intuitive, pervasive myths that turned dolphins into
mermaids. Anyone who can take a deep breath can feel their chest rising as their lungs
fill with air, and they will confidently tell you that this air is being pushed out of their
lungs as they talk, and that the harder they push, the louder they sound. Everyone is an
expert, and anyone can come to the same common-sense conclusion that louder noises
become painful, so this same air must push on the eardrums more strongly, and so
prominence is a more forceful expulsion of air. A number of linguists once came to
similar conclusions, when looks were so clearly revealing that any evidence to the
contrary was questioned sooner than the actual appearances.
Theres no problem with such portrayals, just so long as two things are kept in
mind that balance the value of the intuitively appealing surface descriptions evenly
against the results of deeper scientific observations, namely: 1) that motion observed
at the surface does not represent everything thats going on at all levels; and 2) just
because there is more going on at deeper levels, that doesnt mean that people dont
behave according to what they perceive on the surface, or according to what they feel
is going on inside their bodies. This distinction is crucial here because prominence has
been studied not only in terms of its production (1.2), but its perception specifically
55
in regard to its production (1.3), and it turns out that the perception of prominence
has less to do with the mere reception of an acoustic signal than it does with the
derivation of an articulatory gesture out of that signal (1.4).
Strictly anatomical studies are much less controversial (1.1).
1.1 Prominence IS Anatomical Configuration
Routine and volitional prominence are articulated with the same physiological
structures, so they will be differentiated by their activity, and not their anatomy. To
begin with, the respiratory system is a resource adapted for vocalization, and so none
of its structures is devoted originally to making sound. Lungs, for example, are swim
bladders opened to the outside. This system is then subject to the same constraints
governing the power and speed of skeletomuscular activity in general. The laryngeal
muscles can pull the arytenoid cartilages more tightly, more quickly, or more often
than normal, but only at a greater cost than normal. Sequences of activity (walking,
chewing, breathing) will exhibit rhythmic behavior when not deliberately disturbed.
It is not, then, important to restate the respiratory/vocal anatomy right here;
besides, there are so many good dissections already available that just picking any year
and a letter renders such references as Ladefoged, Lehiste, or Lieberman, all in 1967.
Because it is of ancillary importance to show the wide array of mechanisms available
to prominence for variably energetic activation, material digested from a number of
such sources has been isolated in Appendix I (p. 295).
Researchers are now in agreement as to the respiratory structures involved in
speech; however, they do not agree on how those same articulators function together,
56
so the physiological portions of their studies are reviewed here. Simply parroting the
latest consensus will not suffice, because analyses of respiration and prominence have
naturally been developed in parallel over time, and so changes which have been
wrought in the portrayal of prominence are best understood in comparison to
contemporary analyses of respiration and vocalization.
1.2 Prominence IS Physiological Production
At the turn of the century, two manners of characterizing the respiratory system
flowed together, the first of which appealed to the well-established metaphoric
portrayals using bellows and bagpipes, and the second of which was based on new
anatomical research such as that performed by Mller (1848: ch. II), who blew air
through larynges excised from cadavers (p. 1009 in particular). The ability to precisely
measure the activity of specific muscles in living specimens was available no earlier
than the 1960s, and so the metaphors persisted during the earlier decades, continuing
to be used as the basis of much of the first modern research into prominence.
In these early studies, the linguistic function of prominence is disregarded in
favor of its mythical physiological functions. Such references were refuted as research
methods improved and their results became more widely disseminated. Many linguists
cite the early works of Jones and Bloomfield as influential, and yet in that work at the
time, it was naturally taken for granted that superficial observations were enough to
provide an accurate description of the respiratory mechanism. Stetson goes a step
farther, interpreting his experimental results in support of a bellows metaphor that he
assumed must be accurate, or more likely that he simply never thought to question.
57
Jones (1918) is taken as fundamental by many researchers (Klinghardt, Palmer,
and Schubiger, to name a few), and he describes stress simply as the degree of push
from the diaphragm (p. 24, 101). This seems nave now, but was likely inoffensive
at the time, since it was congruent with a very simple, common view of respiration
dynamics. In fact, researchers continue to portray stress as greater effort, only as
applied to their current model of respiration. Jones did, however, demonstrate a good
deal of foresight in consistently differentiating stress from prominence, as follows:
...prominence depends upon combinations of quality with length, stress
and (in the case of voiced sounds) intonation. (p. 24, 101)
The prominence of sounds may be due to inherent sonority (carrying
power, 101), to length or to stress or to special intonation, or to
combinations of these. (p. 55, 209)
The effect of prominence... is often produced by certain combinations
of [stress and intonation]. (p. 275, 1008)
But this kind of prominence is only important to Jones when it comes to syllables
within words, because the functions which make words prominent in sentences are all
described in terms of physiological stress, and not length or intonation; therefore, there
is no sentence prominence by his definitions.
Bloomfield (1933) is also often referenced in prominence studies, but it is his
association of loudness with stress that later researchers mention, doing nothing other
than to point out that he was wrong. They attribute a disproportionate strength to the
attention that Bloomfield paid to this equation, which amounted to one sentence:
58
Stress that is, intensity or loudness consists in greater amplitude
of sound-waves, and is produced by means of more energetic
movements, such as pumping more breath, bringing the vocal chords
together for voicing, and using the muscles more vigorously for oral
articulations. (p. 110f)
The problem here is not that stress is tied to effort, but rather that loudness is tied to
stress as stress is equated with effort. These are not experimental results which were
once taken to be fundamental, and which later research proved to be wrong, but rather
only the same nave, Jones-ish reliance upon common contemporary premises.
Stetson (1951) represents a crucial development in the support of these sorts of
assumptions, because he worked at a time when the technology allowed him to make
gross measurements of muscular activity on living speakers, getting immediately
below the level of superficial observation. The problem was that this science hadnt
yet been around long enough to wield any real authority over the metaphors that had
already been in place for centuries, and in fact it made no impact on the structure of
Stetsons own favored metaphor.
Stetson comes to three main conclusions, the first of which is that a pulse of air
is made by the intercostal muscles with every syllable. This chest pulse pushes the
subglottal air pressure beyond the mean value maintained by the abdominals (p. 96).
The second is that open syllables involve a collapse of the lung which is countered by
an expansion of the external intercostals. The third applies to stressed syllables, which
have their chest pulse supported by an additional contraction of the abdominal
muscles, primarily the rectus abdominis. Like the results generated in many
pioneering studies, these will all come to be dismantled by their successors.
59
Stress, then, is proposed to be the motion of an additional set of muscles,
where everything in Stetsons Motor Phonetics is muscular motion. Word accent
(prominence) is a matter of stress, specifically duration and intensity, but not pitch:
The logical analysis of sound as involving duration, pitch, and
intensity, is responsible for the belief in pitch as an accentual factor; if
duration and intensity affect accent, then why not pitch? But stress is
not a matter of the properties of sound, it is a matter of the coordination
and culmination of a movement. A movement must involve a stress, a
pulse, and a movement must involve time, but a movement does not
involve pitch.... In speech all stress involves increased force of the
syllable movement, i.e. of the chest pulse. (p. 95f)
Because changes in pitch do appear, Stetson had to find a consistent way to dismiss
them, and so he blamed them on changes in laryngeal musculature which are merely
incidental to tensions in the other musculatures of speech (p. 95). Naturally, this
view will also come to be dismantled.
Stetson made kymographic measurements of body wall movements, took
subglottal pressure readings through a tracheal puncture, and derived lung pressure
readings via a balloon in the stomach. While changes in such readings do suggest that
there are muscles moving, they cannot specify which muscles those are. Stetson read
such indications of general muscle movement only in congruence with the activity of
those specific muscles which mapped well onto the simple bellows metaphor he
supported:
When the chest is slightly inflated for speaking, the air is not under
pressure; like a hand bellows for blowing a fire, the volume is
increased, but the nozzle is open and there is no flow of air....
60
If one makes quick strokes of the hands while holding the inflated
air bellows, the nozzle emits little pulses of air; as the quick strokes are
repeated the air pulses reduce the volume of the air, and the arm
muscles must bring the boards of the bellows closer and closer to
accommodate for the loss of air. (p. 1)
As others note, this does not reflect respiratory factors now taken for granted, such as
the elastic recoil of the lungs (Lieberman, 1968). While Stetson gets criticized for not
having the benefit of the results of the later research which his own made possible, he
is also being held accountable for interpreting his data to fit his theory.
So, Stetsons work gets mentioned by other researchers for two reasons,
namely to treat it as the first study which aggressively measured physical activity
during speech without killing the patient (which involved procedures as invasive as
intramuscular probes and tracheotomy), and to chide Stetson for having assumed that
his experimental results would necessarily support his assumptions about the way that
respiration must be working. Similarly, the metaphors underlying my own analysis
(such as force/finesse, spatial imagery, adaptation, and so on) are just as susceptible to
being modified in congruence with technological innovations, so dont rely too
strongly on their specifics. For example, it might turn out that there is some highly
influential factor in the environment (and so in human cognition) which makes a triad
with force and finesse, in which case the subtypes of elaboration would have to be
adjusted accordingly, just as Stetsons bellows metaphor has been altered.
These early studies leave syllable prominence made up of diverse factors like
quality, loudness, stress, and intensity, but not pitch according to Stetson, who details
the mechanisms upon which Bloomfields vigorous effort would work, all of which
61
is interpreted in support of the prevailing mechanical intuition, rather than modifying
it. It took the following studies to really give in-depth observation the upper hand over
intuitive assumption, and for prominence to be associated with measurable changes in
the full set of fundamental frequency, duration, intensity, and vowel quality, rather
than just some exclusive (and prevalent) subset like duration, pitch and intensity alone.
To begin with, Lifshitz (1933) performed the first of several studies which tied
loudness to at least duration and amplitude, rather than just assuming that it was
equated with stress. House and Fairbanks (1953) is similarly straightforward: changes
in fundamental frequency affect vowel quality. This study is cited by researchers in
support of experiments which demonstrate that pitch is an indicator for prominence,
where those researchers go on to point out that changes in vowel quality must then
also be available to provide clues to prominence in the speech signal.
Fry (1955) is often cited in support of claims that changes in duration and
intensity affect the perception of stress, and is referred to less often to show that
duration actually has a greater influence than intensity. There are also cases where it is
cited in reference to pitch as an influence on stress, but Fry never addresses pitch.
Something else to keep in mind is that these other researchers apply Frys conclusions
more broadly than Fry did. The stress tested here is that which makes part of a word
prominent enough that it can be identified as either the noun or the verb form of an
otherwise identical pair, where the five words tested were object, subject, digest,
contract, and permit. The duration effects were consistent only when just the vowels
were measured for change, and the intensity shifts were less consistent, though
62
significant. While these specific syllable-stress results might well be applicable to
other words, or to sentence stress in general, Fry does not make these claims, and in
fact he makes it quite plain that the scope of the conclusions application is narrow.
Liebermans work (1960) shows that word stress judgments are affected by
formant and fundamental frequency, pitch peak, and duration (psychoacoustic: 1965;
acoustic: 1960, 1967). He brings together the results of several studies, suggesting that
The so-called phonemic pitch levels described by Pike, Wells, Trager and Smith,
and others... may merely reflect the phonetic effect of the discontinuities caused by the
coupling of the subglottal system to the larynx (p. 35). One particular study (1963)
suggests that speakers avoid harmonics of the 300 Hz subglottal resonance because it
will ruin pitch clarity. (My review of such phonemic pitch level studies begins with
Ripman, 1922, p. 99).
Lieberman is effectively saying that no matter what set of articulations and
energy expenditures defines the complex of activities that generate prominence, the
speaker might avoid those processes which would contribute to the disturbance of the
clarity of pitch. The resulting gaps which appear to delimit phonemic pitch clusters or
levels are just the avoided areas in which the pitch becomes unclear due to interference
from the harmonics of the 300 Hz subglottal resonance (the actual rate of which has an
unspecified variance among speakers). If speakers were to consistently display this
avoidant behavior, then the proposed proportion between phonological and semantic
intensities could not be purely continuous for all modes of articulation, because pitch
would manifest a stepwise discontinuity. The increased energy expenditure might still
63
be fluid, but at least a portion of the articulation of prominence would have discrete
targets, which might well influence other aspects of the articulation. Of course, this is
all moot to the extent that Bolinger (1951; p. 107 in this analysis) finds pitch to be
continuous, and without any significant discontinuity.
In another study (1968), Lieberman supports Ladefogeds (1967) correlation
between subglottal pressure peaks and stressed syllables, but only when such a
syllable is the only stressed one in an utterance. Lieberman says, peaks in subglottal
air pressure are generally, but not always, one of the underlying manifestations of
emphasis (p. 1160). Since the words are still perceived as stressed despite the lack of
an accompanying peak of subglottal pressure, Lieberman suggests that word stress
must be able to be marked by something other than this sort of direct effort (perhaps a
marking which facilitates distal reconstruction), and mentions that they all still have
longer duration.
These results are less persuasive when the test itself is analyzed. For example,
when subjects were asked to produce one emphasized word in a sentence such as, Joe
ate a big bowl of borscht last night, or, Joe ate a big bowl of borscht, the word
isolated by prominence was accompanied by a peak in the subglottal air pressure.
When the subject was asked to emphasize more noncontiguous words in a similar
sentence, namely Joe ate a big bowl of black borscht, this accompaniment was not
consistent. The problem is that rhythmically, this last example sounds odd. What I
would like is to see this study performed on a more normal multiple-emphasis
utterance such as, Never kick a man in the foyer.
64
One important aspect of this study is that Lieberman was able to test some
indirect means of measuring subglottal air pressure, namely the tracheal catheter, as
well as esophageal and tracheal balloons, demonstrating that they could return results
which were not significantly different than those gathered by his more direct method,
namely tracheal needle puncture, just so long as a mathematical correction was made
to adjust for the elastic recoil of the lungs.
In summary to this point, the stress vocabulary broadens beyond intensity
and loudness, and even routine levels of prominence are said to be generated by
muscular systems which simultaneously affect an inexact set of stress, fundamental
frequency, duration, and vocal quality. As the bellows metaphor is discarded, a more
sophisticated model is built for the generation of peaks in subglottal pressure which
always accompany the isolated instances of routine syllable prominence. No specific
explanation is given as to how those peaks would necessarily manifest themselves in
terms of all of the other demonstrated changes in fundamental frequency and so forth,
neither is any consistent pattern of change defined; furthermore, as research crosses
into Lieberman and Ladefoged, all of the previous studies on syllable stress are
subsumed as if they automatically applied to words when patterned into sentences.
Most specifically, the strength of prominence is shown to be a matter of increased
activity within a set of articulators, rather than the additional function of one or more
ancillary articulators. The following studies show that the range of articulator activity
over which prominence manifests itself can be measured in terms of the expenditure of
discrete quanta of energy, down to the recruitment of individual motor units.
65
Malone (1926) analyzes accent and intonation into discrete units of absolute
quantity, placing his system in opposition both to van Ginnekens definition of accent
(1907: 287) and Jones (1909a) use of intonation contours (cf. 2.3, p. 91), where
Jones is held to be representative of current pitch and intonation research. Two quotes
are compared to begin the argument. The first passage is pieced together and presented
by Malone as if it were a direct quote from material in van Ginneken, where mental
energy or mental effort is intended to stand in opposition to physical effort:
I define accent as the mental energy which one phoneme has more of
than any of the others, and which manifests itself explicitly in resorting
more strongly to one of five qualities, namely: intensity; height;
quantity; timbre; and articulation. (p. 371; my loose French translation)
Malone then presents a modified form of a quote from his own earlier work:
The enunciation of any speech unit necessarily involves the
expenditure of a certain minimum of energy. Such a minimum may
therefore be looked upon as inherent in the unit. Any additional
increment of energy would then constitute the accent of the unit. .......
(p. 371; cf. Malone 1923: 6)
There is a minimum amount of energy inherent in the production of any speech unit,
and anything above and beyond that is the energy required by the accent.
Van Ginnekens weakest syllables would have no lower limit on their energy,
being successively relatively weaker than those in their environment, whereas Malone
sets up the energy used by the weakest syllable as a base quanta, with increases in
strength involving multiples of that base amount. Accent can then have an absolute
strength, not just a relative strength. Malone divides this energy into: 1) the energy of
66
production, which is dynamic energy giving dynamic accent or stress; and 2) the
energy of maintenance, which is static energy, giving tonic accent or pitch.
Having said this, Malone does not seem to do much with it. He reanalyzes a
passage originally annotated by Jones (1914b: 190-195), using a system which is
essentially like an earlier version of Palmer (1922), and he makes some casual notes
about the differences between American and British intonation. He records changes in
dynamic and static energy, but there is no accounting in terms of the absolute energy
increments he posited.
Ladefoged collects ten years worth of stress research into one book (1967),
research he performed both alone and with a number of other people (Draper,
Ladefoged, and Whitteridge 1957, 1958, 1960; Ladefoged, Draper, and Whitteridge,
1958; Ladefoged, 1960, 1962, 1963; Ladefoged and McKinney, 1963). Ladefoged
dismantles Stetson (1951) in a businesslike fashion, but is just as open when it comes
to implementing changes in his own previous work, readily revising portions which
had been criticized by Kunze (1964; cf. Ladefoged, 1964, for his reply).
To begin with, Ladefoged provides a detailed description of the respiratory
systems activity during speech intensities ranging from soft to shouting, based on
electromyographic readings from an iron maidens worth of intramuscular probes, as
well as air pressure measurements with tracheal and esophageal balloons. These
readings identify the movement of specific muscles, leading straight to one of the first
results, which is that It is quite clear that there is no simple correlation between
intercostal activity and syllables; and... there is no evidence in Stetson to the contrary
67
(p. 20). Although this dispels the myth of there being one chest pulse for every normal
syllable, there is an intercostal burst which accompanies a variation in the degree of
stress (as well as one which increases expiration for voiceless sounds).
When it comes to Stetsons rule for open syllables, Ladefoged says, We could
not find any evidence for Stetsons statement that English syllables with a certain kind
of phonetic structure are always checked by the action of the external intercostals (p.
25). Ladefoged says that the external intercostals are only active during quiet talking
after a deep breath, and not during normal conversation.
Finally, the rectus abdominis does not reinforce the internal intercostals during
the production of a stressed syllable, with the possible exception of very emphatic
stressing, when lung pressure may be unusually high (p. 25). During normal
conversation, however, the abdominals are only active at end of a very long utterance.
Ladefoged also runs a number of tests to determine correlations with increased
subglottal pressure, and finds it to be proportional to volume velocity, vocal cord
frequency, fundamental frequency, and peak sound pressure. In addition, loudness is
proportional to both peak subglottal pressure and peak sound pressure. A little extra
subglottal pressure goes a long way:
These results reinforce the view stated earlier that stress is best
described in physiological rather than acoustic terms. Because of the
interaction of vowel quality and intensity, and the trading relationship
between intensity, frequency, and duration (Lieberman, 1960), there is
no single, simple acoustic event that always occurs in all stressed
syllables in spoken English. But it is apparent that every stress is
accompanied by an extra increase of subglottal pressure. (p. 46)
68
Stress is a matter of effort measured in terms of subglottal pressure, and when stress is
applied to a syllable (prominence), that syllable is altered by some unpredictable
subset of those factors which are proportional to that increase in pressure.
hman (1967) runs with the idea of quantum prominence. To begin with, no
matter how chaotic a complex wave might look, if it is periodic, it can be analyzed as
the sum of a set of very well-behaved wave patterns. hman projects this idea onto the
notion that prominence patterns are rhythmic complex waves, developing a computer
model which shows that the salient features of these intonation patterns in simple
utterances of a number of [Swedish] dialects can be simulated by means of a single
positive step as input to the sentence intonation filter and an appropriately timed
negative pulse as input to the word intonation filter (p. 47). Just one positive pulse at
the beginning of the first syllable of the stressed word, and then negative pulses at the
word stresses, and the prominence pattern for the intonation surfaces.
Unfortunately, there is an insidious danger in the appeal of this study, and it lies
in the same notion which spawned it, namely that any periodic wave can be broken
down into simpler waves. As long as prominence is modeled as an approximation of a
periodic wave, the experiment has no alternative but to work. Even at that, it would be
nontrivial for the experiment to demonstrate that models of prominence patterns could
be generated with a single positive pulse and a couple of negative pulses, were it not
for the fact that the patterns in question are so short that it isnt hard for human speech
to exhaust the mathematically possible results within those constraints, such as high-
low-high, high-high-low, and so forth.
69
The physiological realization of this energy pulse is posited to work as follows:
These pulses would reflect an instantaneous addition of a quantum of
physiological energy to the speech production system as a whole. This
energy is distributed spatially, over the articulatory, phonatory, and
pulmonary channels, as well as temporally, over the time segment of
the utterance immediately following the onset of the pulse. (p. 33)
A pulmonary pulse contracts the intercostal muscles ballistically, and the articulatory
intensity translates as faster and more vigorous movements of the tongue, lips, velum,
and jaw, as well as the articulatory components of the larynx (p. 34). This all sounds
very interesting, but again it is just speculation.
hman analyzes Swedish, and while its prominence will vary from that of
English, it is not different physiologically. This is supported by Gay (1978, below). In
addition, part of Lehistes (1970) book is a comparison of stress research performed on
a number of languages in addition to English, and Lehiste concludes that:
Prominence may indeed have several physiological and acoustic
components. Intensity and fundamental frequency are probably both
factors in the production and identification of a stressed syllable.
Languages may differ in the relative importance of one or the other
feature, and in the relative independence of the two features. (p. 146)
Lehiste also mildly implies that a language using one feature for lexical discrimination
will use the remaining features for prominence. A tone language might use stress or
duration to mark prominence rather than pitch. To the degree that prominence is a
matter of energy expenditure, it is so for all physically articulated languages, although
they vary in the proportion with which this energy is distributed across the set of
available mechanisms, as described by Lehiste.
70
Kent and Netsell (1971) support a direct proportion between stress and effort
that was actually measured rather than programmed. They had speakers utter carrier
phrases of the type My _ again and I _ again, with different amounts of stress on
the inserted word. Their speaking motion was recorded with cinefluorography, which
required the application of radiopaque patches and paste on the subjects articulators.
They found that increases in stress are associated with increases in the muscular
activity of the peripheral speech apparatus, influencing the articulation of the jaw,
lips, and tongue (p. 43). Articulation was enhanced in the direction of the presumed
target, with results such as increased lip protrusion for rounded vowels, or tongue
position shifting anteriorly and superiorly for a high-front vowel.
Sussman and MacNeilage (1978) reduce the difference between intense and
routine effort simply to saying p rather than p. They show that greater effort
entails 1) the recruitment of additional motor units, which means that any given unit
does not have to fire as often as it did over the course of an utterance, 2) an increase in
the motor unit discharge rate, with a shorter interval between the successful activation
of motor units, and 3) a sharply reduced variability in recruitment intervals. At the
articulator level, this amounts to the mandible lowering earlier for an open vowel,
greater displacement of the jaw, with the whole movement being characterized as
more carefully orchestrated and having greater precision, with the target more
forcefully attained. Greater effort leads to greater power and precision.
Gay (1978) shows that when stressed vowels are spoken, increased speaking
rate equals decreased duration. When unstressed vowels are spoken slowly, even if
71
they are of the same duration as the quickly spoken stressed vowels, there is a
reduction in fundamental frequency, and some loss of overall amplitude and vowel
color. Gay concludes from these results that, two separate and independent
physiological mechanisms control changes in speaking rate and lexical stress, one that
horizontally compresses the string and the other that modulates overall articulatory
effort. A change in duration is a deliberate strategy of the first, while only a
consequence of the second (p. 229). Gay has managed to differentiate a deliberate
control over a change in spoken duration from the increased duration that is a
consequence of stress.
As the 1950s became the 60s, then, speculation about prominence as energy
gave rise to actual studies in which that energy was measured. Ladefoged (1957-1967)
used electromyographic readings from intramuscular probes as well as air pressure
measurements with tracheal and esophageal balloons to describe intensities of
respiratory activity ranging from soft to shouting, identifying the movement of specific
muscles. This set the stage for the association of vocal activity with specific measures
of expended energy.
hman (1967) provided a mathematical model for prominence as a variable
energy expenditure, which were described as energy quanta or pulses to reflect the
notion that prominence contours are the result of an interpolation between stressed
points. Beginning with Kent and Netsell (1971), there was an actual measurement of
extra effort specifically associated with stress, followed by Sussman and MacNeilage
(1978) showing how this energy is expended in terms of the recruitment of actual
72
motor units, and then finally Gay (1978), who has energy spent on a prominence-
specific change in duration that is isolated from the mechanism which controls
speaking duration. The difference between routine and volitional prominence should
then be measurable as lesser and greater energy, even if that measurement is
stochastic. All in all, this is one case in which intuition is borne out by science.
So far, volitional prominence is characterized by the expenditure of some
amount of energy which triggers some mechanism (or set of articulatory, phonatory,
and pulmonary mechanisms) to bring about an increase in subglottal pressure, which
has proportional effects on some subset of volume velocity, vocal cord frequency,
fundamental frequency, peak sound pressure, loudness, and somewhere along the line,
duration. The greater expenditure of energy in the articulators can lead to both more
powerful and more precise activation. Naturally, it would be nice to be able to be more
specific at precisely those points where I have resorted to some amount, some
mechanism, some subset, and somewhere, but such precision is waiting on even
further technological advances.
1.3 Prominence IS Physiological Perception
It would also be nice to know if this production research has anything at all to
do with what we actually perceive as prominence, hence this section, which supports
the ability to evaluate signals as distal events.
Lehiste and Peterson (1959) is a study of primary stress on a few noun/verb
pairs (like Fry, 1955), showing that stress is perceived in terms of many influences on
the acoustic signal, which are listed as speech power, fundamental voice frequency,
73
vowel quality, and duration (p. 435). Objective measurement is left for future
research, but vowels which are subjectively judged to be equally loud are evaluated by
the researchers as equal in physiological stress or effort, not actual loudness. They
anticipate the Motor Theory of Speech Perception in saying, It is our belief that the
interpretation of the speech signal by a listener is based on a very complicated set of
auditory parameters by means of which he makes an interpretation of the speech
production (p. 429). These findings have a parallel in the work of Josephs, Giesler,
and Silvera on the judgment of quantity (1994), which shows that a task is interpreted
as if it required greater effort if nothing other than its perceived size is increased.
Ladefoged (1967) performs more tests, showing that when vowels such as /i/
and /u/ are produced with the same subglottal pressure as /a/ and /./, the /i/ and /u/
(suffering greater impedance) will have a peak sound pressure about 5 dB lower than
the /a/ and /./, but will be judged to be just as loud. The same effort put into an /i/ and
an /a/ will give you a louder /a/, measured objectively; however, when subjects are
asked to judge the loudness of /i/ and /a/ produced with the same effort, the two will be
rated as equally loud. Ladefoged says, It is, of course, probably true that what
listeners were doing when they were judging the loudness of the words in this test was
assessing the amount of effort they themselves would have to make in order to produce
corresponding sounds (p. 40). This result is also consistent with the Motor Theory of
Speech Perception, which is built around the notion that the hearer perceives a gesture
and not an acoustic signal, suggesting that the hearer draws upon their experience in
having produced the sounds they are trying to understand. In the 1963 version of this
74
theory (Liberman, Cooper, Harris, and MacNeilage), vocal activity is a matter of
independent articulator movements, where those movements are identified with
subphonemic features. Those features are implemented by the selection of invariant
motor commands sent to the muscles prior to the encoding of an utterance. The hearer
is assumed to appeal to their own speech-motor system to perceive these commands.
When Folkins and Abbs (1975, 1976) showed that articulation was not a matter
of independent movements, but coordinated complexes, Liberman and Mattingly
(1985) revised the Motor Theory, and individual movements became choreographed
articulator gestures. The hearer was no longer trying to perceive the invariant motor
commands assigned to particular subphonemic features, but rather trying to ferret out
the equally abstract control structures for the articulatory gestures.
Mattingly and Liberman (1988) give a general auditory mode for ecologically
arbitrary events and a specialized phonetic mode for ecologically important events
like biologically important sounds (p. 775), saying that, The phonetic mode... is
heteromorphic in the sense that it is specialized to yield perceived objects whose
dimensionalities are radically different from those of the proximal stimuli (p. 779f).
Essentially, ecologically arbitrary sounds are evaluated proximally, and ecologically
important ones are distal events. This makes sense in that important events are more
likely to have experience associated with them, and regular communicative value.
Formants heard homomorphically sound like noises, and formants heard
heteromorphically sound like speech. It would be in the latter case that a speaker
would make a kinaesthetic appeal to their speech-motor system.
75
Fowler and Rosenblum (1991) dislike the Motor Theory in its specifics. They
provide support for the perception of speech as gesture, but point out that no actual
evidence is given in any version of the Motor Theory to suggest that the hearer makes
a kinaesthetic appeal to a specialized cortical system. There is another way to explain
why listeners recover phonetic gesture. It is that phonetic gestures are among the
distal events that occur when speech is perceived and that perception universally
involves recovery of distal events from information in proximal stimulation (p. 43).
In order for this point to be clear, I need to distinguish their use of perception from
the definition that it is given in my analysis. Although there are certainly specialized
sensors involved in the recovery of signal information (eyes, ears, nodes of Ranvier),
there might well be a great deal of equivalence (universality) in the electrochemical
transport of the transduced signal to the brain without regard to which type of sensor
originated the signal, and on top of this, at least some of the receptive areas of the
brain are also specialized. The alternative that Fowler and Rosenblum are trying to
support is as follows: once that signal is in the brain, so to speak, and is actually being
evaluated, they suggest that the only processes needed to reconstruct the distal event
are universal, and do not necessarily require an appeal to any specialized modules like
a speech-motor system. As far as congruence with my analysis is concerned, I would
expect primitive power and precision to be universal, but in order for the articulatory
force of a signal to be usefully iconic, that is to say, for a language user to be able to
feel the gesture, as such, elaboration and revelation rely upon cognitive abilities which
access personal experience with speech as it appeals to linguistic power and precision.
76
In summary, Ladefoged has anticipated much of the Motor Theory, much of
which is still speculative, such as whether people actually access their motor cortex in
the evaluation of speech. Although Fowler and Rosenblum propose a diffuse,
universal cognitive recovery of information from distal events as an alternative to a
direct appeal to specialized cortical systems, the following study on Swedish suggests
that there really is some sort of target or unvoiced speech-specific entity that is an
ideal, and that this ideal is accessed during speech.
1.4 Prominence IS Baseline Excursion
Whereas the studies up to this point have all been treating prominence as if it
changed the otherwise normal, ideal form of a sound, Lindblom (1963) studies how
the unstressed syllables in a word change, effectively portraying the sounds in stressed
syllables as normal. In other words, it is not stress that deforms a sound, but the lack of
stress. Although a vowel phoneme can be realized in a more or less reduced fashion,
the talkers intention that underlies the pronunciation of the vowel is always the
same, independent of contextual circumstances. A vowel target appears to represent
some physiological invariance (p. 1778). In specific, a speaker of Swedish (another
language that uses heavy stress) was found to consistently undershoot the formant
frequencies of eight Swedish vowels in unstressed contexts, such that they tended to
sound centralized. The shorter duration gave the speaker just that much less time to
approximate the ideal, and it also resulted in a closer proximity to the surrounding
consonants, causing greater coarticulatory interference.
77
Lindbloms study shows that a speaker aims for an ideal form of a sound, even
though its realization will always fall short unless the sound is stressed. The speaker
does not deliberately overshoot the target in order to have the actual articulation land
closer to where that ideal would be, but rather the speaker shoots directly at the target
no matter what, even though it is going to fall short. The listener is expected to
understand this, and to extrapolate the intended ideal target from the consistently
undershot realization. Bolingers analysis of reduced vowels (1986: 347-360) supports
the translation of these results onto English, where any given reduced vowel is the
undershot version of a full vowel target.
Lieberman (1967, ch. 3) builds a case in favor of interpreting an infants
attention-getting cries as innate precursors to the intonation contour of the unmarked
breath-group found in most languages. Both the cries and the breath-group typically
display a rising-falling fundamental frequency contour with a duration of one or two
seconds, and a gradual decrease in the baseline frequency. Lieberman presents studies
which show that deviations from this crying pattern can signal neurological and other
abnormalities, suggesting that there is such a thing as a routine, innate base intonation.
Lieberman also includes references to research showing that infants learn intonation
patterns sooner than language.
While I agree with Liebermans conclusion that there is an innate pattern there,
I do not accept some of the projections that he makes based on that conclusion. He
characterizes the breath-group as being designed to segment speech into sentences,
which seems backwards. Why would breathing need to adapt to speech? Vocalization
78
in general should have had to adapt itself to respiratory patterns, and so it would be
better to posit an innate breath-group designed to optimize the exchange of gases and
so forth, and assume that the unmarked breath-group is the one which least disturbs
respiration while allowing for vocalization.
1.5 Prominence IS Effortful Motion
These studies cough up a great deal of information, but the results are not
predictive so much as they are simply statistical. For example, prominence is not
created by an increase in subglottal pressure, but it is virtually always accompanied
by just such an increase. In those cases of prominence where this pressure is absent,
the slack is picked up by features such as formant and fundamental frequency,
duration, and vowel quality. Each of those features has its own set of studies which
proves it to be more important than any of the others in the perception of prominence.
This substitution of features applies not just to the utterances of individual speakers,
but to the prominence strategies chosen by languages as well. In addition, given two
equally powerful acoustic signals, the hearer will attribute greater prominence to the
sound which requires the greatest gestural effort to make, which means that the hearer
must know what it feels like to make such gestures.
To continue, technological advances in research methods allowed observations
to recede below the surface, where they were made not only at an increasing depth, but
at a greater minuteness of scale. The results of such readings did not undercut the
traditional intuitions about respiration and vocalization (e.g. prominence is extra
physical effort), but rather they pulled the support out from under the common sense
79
projections about those processes which had up until then accommodated surface
observations. The traditional portrayals simply adapted themselves to the newly
revealed evidence, giving the bellows a recoil spring. Prominence remains physical
effort, and the only thing that has changed is the array of mechanisms to which that
effort is applied, plus the intuition that this expenditure is now made in discrete quanta.
Prominence still plays into all of the familiar metaphors which involve the expenditure
of energy, such as those which equate a raised voice with anger and heat, or a more
precise articulation with an effort of will or a narrowed focus of attention.
The intuitions which have changed the most are primitive ones which apply to
sensation in general. The first is that all sensations are local, even though they all seem
to be remote. That this is a counterintuitive system is no surprise: sensing a threat on
your retina is nice, but projecting the threats relative location on an environmental
image is better, and doing so intuitively is best. This external-to-internal equation
would be facilitated by any iconic relationships holding between physical and mental
structures. The second is that not all of the information about an event is evaluated
proximally for all signals, but rather cognition involves the reconstruction of distal
events through at least some kinaesthetic appeal to specialized cortical systems in at
least some cases, and probably to more universal cognitive processes in other, if not
all, cases. Whether proximal or distal, the effect of perceiving an increase in the
energy of a gestures form is to attribute a similar increase in energy to its meaning.
These intuitions about effort and iconicity are also of primary importance to
studies which explain where prominence goes (2) or what it does (3).
80
2 Where prominence GOES
When prominence is treated as a fluid motion or holistic process, it becomes an
intonation contour, with critical points along curves acting as borders and stresses, like
high-tension towers lifting draped lines above a rolling landscape. Pitch is no longer
levels, but inflection, and its closest cousin is music, the only difference being that
speech is held to be continuous, while song is made up of discrete notes. These shapes
are not applied to speech to cause a meaning, but are drawn to it according to qualities
determined by an utterances sense. The critical points at which intonation contours
are aligned with these sense groups were first transcribed as punctuation (2.1), and
then as musical notation (2.2). As phonemic principles came to be applied to studies
of prominence, intonation fragmented first into tune peaks (2.3), then syntactic units
(2.4), and finally rhythmic beats (2.5). When the prominences meaning carries no
more importance than just where it goes, the user has often relinquished control to
rhythm as a lowest-level default, and prominence is just a beat in a larger pattern.
2.1 Prominence GOES on Points of Punctuation
Here, prominence has a shape as you draw it you can feel its form, and this
physical feeling is iconically associated with the emotional rise and fall of the sense of
the meaning. A period is a short stab, but it can be made more lingering with a tail that
turns it into a semicolon. (Parentheses are enclosures that hold energy and meaning
inside.) And how bold the exclamation point! These are the first attempts to represent
prominence in terms of a physical activity that reflects the emotional feeling or
meaning of the contour it supported, namely its tune.
81
Hart (1569) supports at least three different tunes, which are differentiated as
early as the beginning of a sentence: And for the mark of the interrogative and
admirative, I would think it more reasonable to use them before than after, because
their tunes do differ from our other manner of pronunciation at the beginning of the
sentence (folio 41, side 2, with font and spelling standardized). In other words,
question marks and exclamation points might be better placed where their associated
tunes start to differ from the declarative, namely at the beginning of the sentence (e.g.
Spanish); however, he stops short of trying to implement this spelling reform, seeing
the matter is of no great moment. Whole tunes are associated with an emotional sense
or mood, which gets signified by a punctuation mark of critical shape. Harts
predecessor to this work holds the earliest English version of what became a standard
system of three accents for marking syllable stress, with one iconically-shaped mark
each for higher/sharper, lower/flatter, and circumflexed tunes (1551, ch. 10, folio 164).
In Butler (1632), Tone is tune/pitch, Sound is force/stress (p. 54, ch. 4, 1),
and Accent is the expressing of one syllable of a word above others; with a little
higher tune and longer time, in whatsoever Tone or Sound (p. 54, ch. 4, 1-2). The
dimensions of an utterances Tone, Sound, and Pauses follow punctuation, where
Butlers system is comprised of both Primary and Secondary Points (p. 58, ch. 4,
3.1). The secondary points have no accent, and are the apostrophe, dash, dieresis, and
hyphen. The primary points are either simple (period, colon, semicolon, comma) or
mixed (exclamation point, question mark, parentheses, square brackets), but only
simple primary points are tied to a consistent pattern of accent, as follows:
82
Table 2-1: Butlers Association of Punctuation with Accent
Tone is final, unless drawn away by emphasis. The secondary points are defined in
terms of accent, but not in as tidy a fashion, where their patterns vary with the primary
point that they have expressed or understood, in them (p. 60, ch. 4, 3.2), which can
involve rising Tone. Hultzn (1939: 42) equates this with the two-tune system adopted
by Jones (1918), but only does so by interpreting as rising Butlers description of a
level tone: it continues the tenor or tone of the voice to the last word (p. 59, ch. 4,
3.1). This doesnt look like a two-tune system any more than does the three-tone
system found in Walker (1787: 15f), given Butlers continued tune and Walkers
Monotone (2.2).
So punctuation once signified the sense of an utterance, or the shape of the
mood over which the intonation was to be draped. The actual punctuation mark was
tied to the end of the contour, although Hart suggested placing it at the beginning of
the sentence, where the difference in intonation begins. Hart doesnt say what those
tunes are, but Butler suggests the typical components: rising; falling; and level tones.
Accent (Intonation Pattern)
Tone
(tune/pitch)
Pause Sound (force/stress)
sense sentence
S
i
m
p
l
e

P
r
i
m
a
r
y
period fall (on the
last word)
long
perfect
perfect
colon shorter than
for a period
imperfect
semicolon continued
(to the last
word)
imperfect middle ofn: orn.
comma shortest more imperfect
83
2.2 Prominence GOES on Notes of Music
As the 18th century became the 19th, the treatment of prominence as a musical
contour became increasingly prevalent. This outlook initially opened up in response to
a demand to transcribe spoken intonation for efficient English teaching, and the
perception of that demand only grew through the 20th century in response to the
consequences of the World Wars. One of the first issues in this musical alignment was
the matter of the most appropriate degree of complexity necessary to capture the needs
of language, as seen in the flow from Sheridan (percussion alone), to Burnet (drums),
to Steele (complex vocal orchestration), to Walker (simplified melody, as later in
Sweet), and finally to Odell (musical, but without strict musicality). The adaptation of
the iconic physicality and spatial orientation of the straightforward tone strokes to
musical notes on staves was a natural one.
Sheridan (1762) holds that there are special villains responsible for English
accent being commonly misconstrued as musical in nature, namely those masters of
grammar schools who revived the teaching of ancient literature. He contends that they
decided that in the absence of knowing just what it was that the Greeks had meant by
their accentual marks, those marks must have described tonal inflections. Sheridan
admits that the Greeks must have done something with their language to make it sound
pleasing, as Foreigners listened to their Orators, tho they did not understand their
language, with as much pleasure as we do Italian singers (p. 45), but he concludes
that whatever it was they did do, it was nothing like the artificial inflections recreated
by the teachers of Greek. He also cuts off any inquiry into the musicality of English
84
accent by saying that the only modern use of musical accent is pursued by the Chinese,
of whom we know but little, and a thing about which we can have but very obscure
ideas (p. 46). The reader is supposed to flush all of their old misinformation about
English accent in order to make room for new (mis)information, which is that English
accent is not musical, and is not a matter of inflection. It only distinguishes one
syllable in a word from the rest, either by duration or stress, which is simply called
smarter percussion (p. 41).
Burnet (1773) wrote six volumes on the origin and evolution of language in
general, as well as the state of English in particular, and deep inside the second volume
he buried a characterization of syllable stress as nothing better than the music of a
drum, in which we perceive no difference except that of louder or softer (v. 2, ch. 4, p.
300). Earlier, he had differentiated syllable and sentence stress as follows:
As to accents in English, Mr. Foster... would fain persuade us, that in
English there are accents such as in Greek and Latin. But to me it is
evident that there are none such; by which I mean that we have no
accents upon syllables, which are musical tones, differing in acuteness
or gravity. For though, no doubt, there are changes of voice in our
speaking from acute to grave, and vice versa... these changes are not
upon syllables, but upon words or sentences. (v. 2, ch. 4, p. 298)
That there is truly no other difference, is a matter of fact, that must be
determined by musicians. Now I appeal to them, whether they can
perceive any difference of tone betwixt the accented and unaccented
syllables of any word. (p. 299f)
Burnet ends up recanting when Steele meets and exceeds this challenge, changing his
perspective in later editions of his work.
85
Steele (1775) associates speech with a complex musical representation (p. 51):
Figure 2-1: Sample of Steeles Transcription
The bottom portion of an accent is a straight or circumflexed line iconically describing
the variable path of quarter tones (lines and spaces) over which the voice traces a slide,
and the top portion is a mark of quantity (cf. Thelwall, 1812: xliv). Acute and grave
accents are simply monotonic slides which traverse most of the scale. Steele has gone
to great lengths to show Burnet some tones differing in acuteness and gravity, as well
as all of the other kinds of musical notation applied to speech, saying further: Neither
is it like the intonation of the chorostates, or precentor in our cathedrals, where the
change of tone is made between one sentence and another, or between one word and
another; that is, where the change is made, not upon syllables, but upon words or
sentences (p. 15). It is in the face of this system that Burnet decides to change future
editions of his work.
Although Walker (1787) is much more practical than Steele, the notion that
singing can be reduced to notes so as to delineate them on paper also suggests to
Walker that the same might be done for speaking (p. 1):
, , ,
....... .............................................
3
c
:
{ {
And now, if ever westood in need of mature
, , ,
86
ALL vocal sounds may be divided into two kinds; namely, speaking
sounds and singing sounds. Singing sounds are such as to continue a
given time, upon one precise point of the musical scale, and then leap
as it were from one note to another; but speaking sounds, instead of
dwelling on the note they begin with, slide either upwards or
downwards to the neighboring notes, without any perceptible rest on
any; so that speaking sounds are exactly of the same kind with those
which are produced by a violin when a finger slides up and down the
string, while the bow is drawn across it. The singing sound, therefore,
is a Monotone, and the speaking sound a slide or inflexion. (p. 7)
This looks beyond what prominence is and on into where it goes, blending points into
intonation contours. They slide rather than leap, and so are motion and inflection.
Walker describes the voice as having five properties (p. 8): his musical
Monotone (

); rising and falling slides (. ); and two circumflex slides (

).
This in-line marking is less costly than printing musical or tonetic information above
the line, and for some purposes the details that it ignores are not important. These
marks are part of what will become a standard tonal toolbox (cf. Sweet, 1890: 2f,
1892: 228f; Kingdon, 1958b: 2; Palmer (2.3)), and they reflect Hart (1551).
These are schematic patterns of relative motion without specified sound, and
all speaking sounds are held to move in one of these five ways. Amongst the other
properties of voice which may accompany these slides are pitch (high or low),
loudness (loud or soft), duration (quick or slow), and some sort of emotional
charge (passion, forcible or feeble). While pitch and duration (as such) are noted
to have precise musical counterparts, all of these other qualities are denied definite
scales when it comes to reading, speaking, or singing (p. 10), only relative proportions.
These are marked as parenthetical asides, such as, Arguing; a cool, sedate, middle
87
tone of voice (p. 26). In addition to accented and unaccented syllables within words,
he also lists the emphatic syllable within a sentence, which is but the accented
syllable made louder (p. 11). This equation of emphatic stress and loudness draws far
less criticism than Bloomfields associating general stress with loudness.
Odell (1806) minimizes his own contribution, saying that On the subject of
accents, I have done little more than comment on the text of Mr. Steele (p. vi), but his
commentary itself is not so little. While Steele illustrates speech by emphasizing its
similarity to song, Odell wants to drive home two differences, denying both discrete
and level tones in speech. He says that accurate observers notice:
...that an acute accent was indeed an elevation, and a grave accent a
depression of the voice; but that, in this elevation and depression, the
tone of the voice was varied, not as in singing, by distinct intervals, but
by a continued motion, gliding up and down, in a kind of undulation.
(p. 58)
It is evident, that in such a movement, whatever the interval from
tone to tone, through which the voice may glide, it can never dwell for
an instant on any tone whatever. (p. 59)
In addition, Walkers Monotone is dismissed as merely a succession of accents of
small dimension, and nearly of one uniform pitch (p. 121), so while speech can seem
monotonous, it cannot be monotonous.
Odell supports these claims with ancient Greek and Latin passages describing
speech, proving not only that those texts can be interpreted in a manner consistent with
his own descriptions of language, but that they need not be interpreted as if they
supported the prevalent conception of Greek and Latin as a sort of musical oration. In
88
order to turn that need not into should not, Odell argues that neither Greek nor
Latin could have been strictly musical, resting his contention on four points. First,
Odell suggests that Greek and Latin never really were strictly musical, and that their
mistaken characterization was all just a matter of scholars misinterpreting those
passages which portrayed language in terms of tone. Second, since there are still
human descendents of Greeks and Romans, but no trace of strict musicality in their
descendant languages, it is morally impossible for there ever to have been a quality
to leave a trace (p. 56). Third, even had the Greeks and Romans been utterly wiped out
at some point, the organ-based human ability to have such a musical language would
not, and therefore the modern lack of musical languages is testament to the fact that
they never could have existed. Fourth, and finally:
It is at the same time a fact, of which we have the attestation of our own
ears, that, in modern speech, anything like a musical tone is so far from
being either significant or pleasing, that we turn from it with universal
disgust: nor can we endure the smallest mixture of speech and song....
That which is universally disgusting now, can never have been pleasing
to either Greeks or Romans. (p. 57)
Odell is intending that disgust is registered relative to a sensory aesthetic rather than
a moral one, and so he is suggesting that this sort of sensory appeal must be universal
among humans. All in all, Odell is supporting Steeles association of speech and song
insofar as accent is concerned, but is firmly denying the strict musicality of language
by emphasizing the inflection of speech as continuous, as well as the innate human
antipathy for mixtures of song and speech.
89
The path traced to this point is natural: Burnet tosses a gauntlet by challenging
musicians to disprove his contention that speech is no more musical than that it has a
louder or softer beat (which is similar to Sheridans description); this prompts Steele to
go to great lengths to support as musical a portrayal of speech as possible; Walker then
simplifies this notation, allowing for the analysis of long contours into sequences of
five shorter contours, and paving the way for those who will want to know where those
contour pieces go (2.3); and, finally, Odell emphasizes the critical differences
between speech and song as far as strict musicality is concerned, while leaving a
feeling of predominant similarity with Walker.
So, whether seen in terms of punctuation or musical notation, prominence is a
matter of alignment, and its contour (as a whole or broken into critical points) settles
over a landscape determined by an utterances sense. It is treated as systematic and
rhythmic, like music, with the proviso that the tone of a spoken voice is a continuous
motion, while a song is a string of jumps along a sequence of discrete, domesticated
notes. The answer to the question, Where does prominence go? is not, Wherever it
wants to, but more like, It just follows an 800-pound gorilla named Sense. My own
analysis shows how Sense was first trained for communication with power and
precision, and then tamed with elaboration and revelation for language.
As linguistic analyses become more sophisticated in the 1900s, holism goes
out of vogue in favor of iterative analytic dissection. Increasingly smaller utterance
parts will create additional moorings for smaller contour parts, namely: component
tune peaks and phonemes (2.3); syntactic units (2.4); or beats (2.5). Volitional
90
prominence will stop being portrayed as a broad emotional outburst that lifts the whole
contour, or as an inflation of a contours peak that draws other parts along with it
elastically, and start to be seen as having a local effect. The meaning of prominence
starts out well under user control, but withers as this decomposition continues, and it is
no small wonder that these methods of placement become predictive by default. When
the language user gives prominence no power to go where it makes sense, prominence
simply gets placed where the language makes it mark time.
By this time, the question of what prominence is is no longer physiological,
but structural and functional, and so it starts to become an integral part of the question
of where it goes. In effect, all of the studies from this era resort to answering the
general question of where prominence goes in essentially the same way:
In a sentence, those words are said to be stressed which are pronounced
with greater breath force than the others. These are the words which are
felt by the speaker to be important; if he feels one idea only in a
sentence to be important, he stresses the word embodying the idea; if
many ideas, he stresses many words. (Armstrong and Ward, 1926: 3)
They have their own pet names for Sense. Palmer (1933) says that his tone-pattern
placement depends upon the words which the speaker feels to be the most prominent
(p. 5), or upon what the speaker wishes to give the maximum of prominence (p. 8).
The rest of the studies in this section make similar statements.
This laissez-faire attitude can be blamed upon the end of World War I, as the
massive influx of foreign students of English into Britain motivated teachers to devise
an intonation notation that would safely ignore irrelevant details while generating an
91
organized list of intonational phrases. Such a list would be cumbersome, but
necessarily limited, and the suggestion of patterns in the data (both in form and
meaning) would make the material seem easier to memorize. Contours recorded as
vectors at critical points on a graph are much easier to organize than detailed,
continuous contours. These patterns can then be viewed as cognitive routines, and
since the notion that humans resort to routine or rhythm is well-established, these
analyses certainly capture that aspect of prominence well.
2.3 Prominence GOES on Peaks of Tunes
Jones (1909a) uses analog curves to try to hit a point between 1) pitch
diacritics, which are too schematic for his tastes, and 2) kymographic tracings of voice
vibrations, which are too elaborate. Of the diacritics, he says that:
Such marks may give a rough idea of the kind of intonation required...
but they fail to show with any sort of accuracy the precise points of the
sentences at which the changes of pitch begin and end, and they do not
profess to indicate the absolute pitch, or the subtle variations of pitch
which are perpetually occurring in speech. (p. iv)
Of the kymographs, he says:
The vibrations... may be measured, or the number occurring in short
units of time counted and the results plotted on squared paper, the
variations of pitch being thus expressed by curved lines. Such curves
are, however, inconveniently large and elaborate, and the phonetic
symbols to which the various parts of the curves correspond have to be
placed far apart and at irregular intervals, thus rendering the text
difficult to read. Besides, the work of preparing curves by this method
is so laborious, that no one has ever yet analysed texts of sufficient
length to be of any practical value to language students. (p. iv)
92
In compromise, Jones records speech on paper as analogous to song, just as it was in
the earlier century. He starts with recordings and phonetic transcriptions of English,
French, and German conversation. A musical staff is printed above lines encoding
vocal quality, upon which Jones marks duration (: or..), loudness (with a stress mark),
and vocal pitch as a musical note or dot. He gets this note by removing the needle from
the record at the same place a number of times, and then comparing his impression of
the note at that point with the absolute pitch rung on a tuning fork. When these note-
dots are connected, intonation curves appear, and prominence becomes a curves peak.
Intense prominence is then simply a statistically higher-than-normal peak.
Coleman (1914) uses a digital system derived from Jones (1909a), translating
intonation curves into numbers representing relative pitch heights during an utterance,
with his 1-9 scale projected onto a musical staff over the phrase. The contour peak is
simply the highest local number in the sequence. Prominence (sentence stress) is a
local contour peak, and when intensity falls on that word, its height is exaggerated to
further differentiate it. When intensity falls elsewhere, a distinctive intonation is
probably never absent (p. 15, 26), but it can be marked by any number of changes,
amongst them a crescendo of stress rather than just loudness, or extra slowness, extra
quickness, length of word, additional words before the intensified word to gain
attention by keeping one waiting, pauses with the same object, and other devices, such
as repetition or additional words generally (p. 15, 26). Also mentioned are the use of
an afraid voice when talking about huge or awful things, or saying leetle for
something especially little.
93
Klinghardt (1920) uses a string of dots to represent a contour for a sequence of
syllables, where darker dots (point size) indicate stronger stress, and rather than
resorting to numbers, he uses dot height (vertical spacing) to indicate shriller pitch.
The pitch range is indicated by what amounts to the upper and lower lines of a musical
staff, without the intervening lines. An apostrophe or comma at the end of a contour is
used to mark a rising or falling final fillip. Prominence is continuous at the level of
stress and pitch. This system of notation rests on such assumptions as that stress is
iconically related to size, just as pitch is to height. Analyses of intonation which appeal
to this system of notation (often to illustrate a new system) include Kingdon (1939),
Trim (1964), Chappalaz (1964), and Crystal and Quirk (1964), but the most immediate
successor is Armstrong and Ward.
Armstrong and Ward (1926; A&W) devised a popular two-tune toolbox, and
wrote A Handbook of English Intonation as a manual for foreign students (p. 1).
A&W felt that intonation was not simply acquired naturally over time, and that its
study was typically neglected. If the student realized that correct speech melody is as
important as correct speech sounds, they would devote more time and energy to this
essential characteristic of our language (p. 1). It is apparent that these students were
supposed to learn some sort of standard English intonation, because A&W specifically
omit more elaborate schemes of English melody that vary according to region or
individual speaker (p. 1). What we have here is a detailed, taxonomically classified set
of observations intended to describe how intonation appears, and not proof of why
intonation works the way it does.
94
Avoiding other analyses, A&W credit only Klinghardt (1920), who is cited in
the preface only as the originator of their adapted intonational marking system. Their
encoding is simply a sequence of segments, any of which can be: 1) a pitch-contoured
line which is as short as the syllable it marks; or 2) a dot over an unmarked non-final
syllable; or 3) a pitch contour over a final syllable. Pitch marking is only placed on the
syllables of grammatically important words, such as nouns, adjectives, principal
verbs, and adverbs; and by squishing together the timing of the dotted syllables, the
stressed syllables are able to maintain a regular spacing, which gives English its
characteristic rhythm (p. 7). The segments are not drawn connected, but their actual
pronunciation should flow together continuously.
If you can picture the great number of patterns that can be generated under this
system of lines and dots, even when the pitch gradient tends to be monotonic, then you
can start to imagine the number of intonation contour skeletons that are classified in
this book. The drawings representing these patterns run all the way from a single
curved line, right on through patterns with over a dozen segments. The upper limit on
the number of syllables is defined by the limit applied to the length of a sense-group,
which is paralleled by an intonation group (p. 25). According to the range of data
presented by A&W, a sense-group is well familiar as a next-size-larger-than-single-
word-idea-unit, and while a sentence is typically used to express only one such
complex but readily circumscribed idea, there are also conjoined series of intonation
patterns that are used for sentences which express two or more likewise linearly linked
ideas.
95
A&W manage to wrestle this broad array of patterns down to a few sets of
variations on two basic tunes. Tune I is a generally falling contour, varying with the
initial rise or fall of the first few segments on their way to meeting up with the middle.
Tune II looks about the same, except that the last couple of end segments can show
some independence and rise a little bit, but usually not very high. These variations are
subclassified primarily according to the number of stressed syllables which appear in
the string, or whether the sentence expresses more than one sense-group, and therefore
requires the conjunction of sequential contours. In addition, intonation patterns of both
tune types can appear in either emphatic or unemphatic forms.
The big change found in Palmer (1922a&b, 1924, 1933, some editions edited
with Blandford) is a shift from the two standard tunes with their intractable variation
to six basic tone-patterns (p. 16; all references are from 1924), as follows:
Figure 2-2: Palmers Six Tone-Patterns
Each of these tone-patterns is built around one of five nucleus-tones (p. 14), either a
high fall (acute, ), a low fall (chronic, ), a full rise (acute, ), a low rise
(chronic, ), or a rise-fall-rise (acute, ). Pre-nuclear variation is handled by
one of three monotonic heads (p. 15f), the pitch of which either: 1) falls to meet the
Cascade Dive (Ski) Jump Wave Snake Swan
96
nucleus (superior,

); 2) ends no higher than the start of the nucleus (inferior,

);
or 3) rises above the starting point of the nucleus (scandent, ). The patterns
tail is not transcribed (p. 15), but is simply understood to be matter of extrapolating
the direction of the end of the nucleus, or interpolating between two nucleus-tones.
As early as 1922, Palmer had abandoned the intralinear tonetic dot notation for
his new intratextual transcription of lines and arrows, although for the sake of clarity
he listed 48 patterns in dot notation for reference in later editions of 1922b. These
marks were written in-line, right where the tone started. By 1933, these functional
head and nucleus units were fused into tone-patterns. Although it obscured certain
details, Palmer felt that the newer system was better for teaching because it was less
technical, easier to read, and concise. The patterns were named after the shape of the
resultant contour (listed in the same order as in the above figure). Theres the Cascade
(high head, low fall), the Dive (high head, high fall), the Jump (rising head, low fall;
called the Ski-jump in 1933), the Wave (high or low head, full rise), the Snake (high
head, rise-fall-rise), and the Swan (rising head, low rise). Some of these patterns may
repeat or alternate in an utterance.
None of the six tones has a unified meaning or function. Tone 5, the Snake, is
close with only two subfunctions, so-called contrastive and emphatic utterances. Both
of these revolve around the isolated prominence of one word, so they could be seen as
uniting to make Tone 5 the pattern used to mark focus. Tone 4, the Wave, is a type of
question intonation, also with one word isolated for prominence. In both 4 and 5, then,
the word in focus would be marked by the peak of the nucleus-tone, the difference

97
being that the intonation does not drop back down in 4 because it is a question (hence
the Wave), while in 5 it does (likewise the Snake). What this all comes down to is that
while Palmer has classified the data according to head-nuclear combinations, the data
falls out more naturally when classified simply according to prominent or not.
Kingdons (1939) intent is to develop a unified transcription system for stress
and intonation, because the two are so interdependent that to indicate one without the
other is unsatisfactory (p. 60). He dismisses Palmers in-line premodifier system
because the tone-pattern symbols are too prominent to be placed in the line of the
text (p. 60). Kingdons own superscripted premodifiers identify stress ('level,
falling, and rising) and intonation (fall, low-rise,

(rise-)fall-rise, and

rise-fall),
which are doubled for emphatic forms. While these symbols certainly are smaller than
those used by Palmer, they are just shrunken representations of the nucleus-tones that
Palmer used early on, before he fused them into tone-patterns.
Kingdon goes on to transcribe 44 different patterns for I cant find one,
where variables in the generated table include routine and emphatic stress on cant,
crossed with routine and emphatic tunes on find. (Four patterns are missing because
routine falling stress on cant evidently doesnt cross with routine tunes on find.) The
remaining eight patterns are routine and emphatic forms of cant and find intoned
alone. (Trim (1964) adapts this system for German.) Each pattern is assigned a
contextual meaning along the lines of apologizing for not having found it in the place
indicated, according to context may express either mystification or exasperation, or
unemphatic contradiction, made deprecatingly (p. 62).
98
By 1958, the tones split again, as the system accommodates: 1) the separation
of (rise-)fall-rise into fall-rise and rise-fall-rise; 2) the two subsequent types each of
fall-rise and rise-fall-rise (divided and undivided amongst syllables); 3) the three
subsequent types of rise-fall (one-, two-, and three-syllable versions); and, 4) the
doubling of everything to take into account the possibility of its being spoken in a
generally high or low pitch. The pendulum has simply started to swing back. If
Kingdon were just to write these superscripts outside the line of text, hed have
reinvented an intralinear tonetic system like Klinghardts.
In these tune peak studies, prominence continuously varies by pitch and stress,
with a resulting proliferation in systems designed to record this information for later
classification of these contours. The first step is to break contours into critical points,
with everything in between accounted for by an expectation of interpolation. All that
any given individual point needs to reveal is a direction and a strength, so its just a
vector. Direction can be just a matter of interpolation (i.e. just aim at the next dot), or
actual direction (i.e. aim up until further notice, such as when reaching the upper edge
of the staff). Strength is either the darkness of the dot or can be conflated with the
height of the dot. Variation comes in when analysts fuse patterns of critical points into
clumps of different sizes, and then classify by dot-clump rather than dot-sequence.
That explains the pendulum effect: if none of these systems is truly satisfactory, or if
their intended function changes (from teaching to analyzing and back again), then
developers will simply fuse distributed systems or analyze holistic ones, back and
forth, looking for a compromise which pleases everyone.
99
This treatment of prominence in terms of tune peaks seems to be a peculiarly
British game, as isolated as its American counterpart, namely the phonemic analysis of
intonation.
Ripman (1922) is the liminal stage between treating intonation as contours or
music and the phonemic analyses of pitch below. He uses a 13 scale for recording
pitch over syllables, where 1 represents what we may call the level of indifference,
the note on which we utter words to which we attach no special importance; 2 is a
higher note, 3 higher still. 3-1 will indicate a fall from 3 to 1 (p. 68). Only rarely does
Ripman resort to a 4 or 5, though he indicates that more numbers may be used to
widen the pitch range. There are also three pauses, namely short ( | ), longer ( || ) and
longest ( || ). While he makes no actual phonemic claims (complementary
distribution, and so forth), these are not just contour points. The number of pitch
levels, once determined, is fixed, as opposed to Coleman, for example, in which they
can just keep on going up, but still, his system is based on how many levels are
adequate, rather than necessary.
Bloomfield (1933) is the one who manages to congregate (if not strictly unify)
the segmental and prosodic processes of English into one system with his principles of
phonemic analysis. The trick turns out to be as simple as calling the segments primary
phonemes, since they behave consistently under phonemic analysis, and then
identifying the contours and stresses as secondary phonemes, because although they
are not as readily susceptible to phonemic analysis as the segmental material, they still
comply after a fashion, rather than not at all.
100
Bloomfield would like to find complementary distribution for whole contours
according to their meanings, but settles for partial sentence-final contours according to
grammatical type. He admits that there is no steady correlation for given contours and
specific grammatical types, but he is willing to work with trends, which is entirely
reasonable since thats as close as hes expecting to get. He derives the following
prosody-grammar associations, using punctuation for symbols since hes only working
with utterance-final pitches: a period is a falling or final pitch; a question mark is a
rising pitch at the end of a yes-no question; and an inverted question mark is a rising-
falling pitch which can have a lesser rise at the end (p. 92), which is identified with
other-than-yes-no questions (p. 114). An exclamation point is a non-distinctive and
gesture-like distortion of these other pitch schemes which is used to express anger and
surprise, or for voice types like sneering or calling (p. 115). There is also the
unassuming comma, which: 1) is used for a pause-pitch; 2) is said to be preceded often
by a rising pitch; and 3) promises continuation (p. 115). Prosodic material is seen as
no more than a modification of the fundamental segmental material, such that stress
and intonation are taken as varying in otherwise identical forms (p. 114).
The levels of phonemic stress are more straightforward: Our highest stress ['']
marks emphatic forms, usually in contrast or contradiction; our high stress or ordinary
stress ['] appears normally on one syllable of each word; our low stress or secondary
stress [
|
] appears on one or more syllables of compound words and long words (p.
111). The function of stress is only brought up so that having one can be used to
separate abnormal from normal stress, where normal stress only has a location.
101
Bloch and Trager (1942) are unconcerned with the intonation notation, saying
that Tone levels (higher and lower) and tone contours (rising, level, falling, etc.) may
be indicated by accent marks over the letters, by superior numerals with assigned
values, or by other devices (p. 35). They adopt Bloomfield, though, adding symbols
for suspensive and contrastive pitch, as well as secondary phonemes for juncture,
which differentiates the likes of night-rate, nitrate, and dye-trade (p. 36).
Harris (1944) expects a phonemic analysis improper to be the right approach:
the components described in this paper are not complete physical events; therefore,
they cannot actually be substituted for each other to see if any two of them are free
variants or repetitions of each other (p. 201); but, pitch and stress... constitute
morphemes by themselves, independent of the rest of the speech, with which they are
simultaneous (p. 182). Stress creates allophones, for example, we... consider pitch 2
to be an allophone of pitch 1 in stressed position, and occurrences of relative high
pitch 4 at one or more places... will always be accompanied by a loud contrastive
stress... and can therefore also be considered an allophone of pitch 1 (p. 189). It is the
pitch patterns and not the pitches themselves which are suprasegmental (as opposed to
Wells, Pike, and Trager and Smith, below), Since... components are not restricted as
to length... each of these pitch sequences is a single component whose length is that of
a whole utterance or phrase (p. 190). Pitch patterns 112, 113, and 114 are allophones
with increasing stress placed at the end of 112. The actual number of pitches is not a
matter of necessity, but of adequacy, where he shows examples of pitches 0 through 6,
but there could obviously be more if there were simply time for a longer rise.
102
Others researchers have treated suprasegmentals such as stress and juncture to
a phonemic analysis, or have analyzed whole contours into pitch phonemes, but Wells
(1945) wants to subject the component pitches to a phonemic analysis, which draws
the familiar apology:
In proving the existence and distinctness of the four pitch phonemes by
minimal contrasts, we ought to consider the same string of segmental
phonemes with minimally contrasting pitch contours imposed on it; but
it is difficult to find one such string equally well adapted to two
minimally contrasting contours. In most of the examples, therefore, the
contrast is minimal only so far as the contours are concerned. (p. 32)
This brings him to compare the numerical pitch sequences with no attention paid to the
utterances upon which they sit. Of all of the possible combinations of four pitch
phonemes (1 low, 4 high) taken up to four at a time (341 if you include no
phonemes), he finds 19 sequences which can be matched to English utterances, and
distributes these sequences over 29 example sentences. Minimal contrasts are
identified in such examples as
3
I
1
did it and
3
I
2
didnt ask you, did I? (between 1
and 2), or in
2
The
4
tel
2
ephones ringing! and
3
How did it
4
hap
2
pen? (between 2
and 3). He accommodates Bloomfields analysis, such that a question mark becomes a
23 sequence, and the appearance of an exclamation point means that there is a 4 in the
string. He assimilates Palmers contours, but makes the critical remark that tying up
pitches with stresses in one system engenders not merely repetitions but difficulties
(p. 69). He does not explain why it is okay, then, for Wells himself to say that stress
always falls on a 3.
103
In Wells, suprasegmental phonemes arent simultaneous (4 3 + 1), because
the length of a pitch phoneme is variable.... Its effect will continue until replaced by
another pitch phoneme, or until the end of the utterance (p. 31). Wells also says that:
The pitch phonemes... are organized into meaningful sequences called
pitch morphemes, which are the strict analogs of segmental morphemes
composed of segmental phonemes. But pitch morphemes are so few in
number that it seems unnecessary to regard them as in turn organized
into syntactical arrays. (p. 34)
He only talks about two sequences, the first of which is 231, which means normal
sentence stress, and the second of which is 4, which indicates surprise (p. 35).
Pike (1945) doesnt treat glides in terms of their directions (falling, rising,
and so forth), but rather analyzes them in terms of the pitch heights of their end points.
This is necessary to distinguish between, for example, two rising glides which differ
only according to their end points. Pike determines that four pitches are all that is
necessary to distinguish the end points that differentiate all of the contrasting pitches.
The resulting sequences describe contours which can contract to drape over
one syllable, or expand to cover a whole utterance, and so they are not dependent upon
particular grammatical structures, but rather the speakers attitudes:
The utilization of any specific contour is not determined by the
structure of the sentence upon which it is to fall, but by the attitude of
the speaker utilizing that construction. Any specific construction may
have superimposed upon it any of the English contours, provided the
speaker has the requisite attitude when he does so; this holds true even
when the meaning of the words conflicts with the meaning of the
intonationbut irony, or jest, or some special innuendo results from
such semantic conflicts.... Intonation contours cannot be defined in
terms of the grammatical constructions with which they occur. (p. 163)
104
While this is the same thing that the British researchers were saying in their section, it
has taken until Pike for the same sentiment to be expressed once the Americans got on
their phonemic kick.
These sequences are independent of the dictates of the grammatical types
which they are normally assumed to accompany, which frees them to have meanings
more sophisticated than goes on a yes-no question, where lexical and intonational
meaning are differentiated as follows:
English words have basic, intrinsic meanings; these LEXICAL
MEANINGS are the ones found in the dictionary.... [They] are
indicated only by the requisite consonants, vowels, and stress, and a
context where such a meaning is possible; in that sense, the lexical
meaning is intrinsically a part of the word itself and not dependent
upon extraneous phenomena such as pitch produced by emotion.
The intonation meaning is quite the opposite. Rather than being a stable
inherent part of the words, it is a temporary addition to their basic form
and meaning. Rather than being carried by permanent consonants and
vowels, it is carried by a transitory extrinsic pitch contour. Rather than
contributing to the instrinsic meaning of a word, it is merely a shade of
meaning added to or superimposed upon that intrinsic lexical meaning,
according to the attitude of the speaker.... In English, then an
INTONATION MEANING modifies the lexical meaning of a sentence
by adding to it the SPEAKERS ATTITUDE towards the contents of
that sentence (or an indication of the attitude with which the speaker
expects the hearer to react). (p. 21)
The meanings assigned to the contours are sometimes no more specific than, Rising
contours generally imply that the speaker considers them to be INCOMPLETE by
themselves, and NEEDING SUPPLEMENTATION of some type, by himself or by the
hearer (p. 51), where that supplementation is another clause, but rising clauses are
also noted to sometimes be polite or cheerful. A 23 contour is hesitant when non-final,
105
but implies endearment when final. Pike also mentions that tone overrides lexical
choice (p. 22), in that we believe the message conveyed by tone if it doesnt match up
with the words.
So much for the reliability of good old question intonation.
Haugen (1949) suggests that only successive sound features be referred to as
phonemes, while simultaneous ones be called prosodemes, even if this means giving
up attempts to demonstrate that stress is phonemic in the sense of being distinctive. In
general, any significant sound feature whose overlap of other features is temporally
correlated to syllabic contour should be called a prosodeme, and should be treated by
itself in a manner appropriate to its special nature (p. 282). Even though this seems to
fling the door wide open to demonstrating that stress is prosodemic, the term seems
to have died here, evidently losing out to suprasegmental, due in large part to Trager
and Smiths adoption of the generative grammarian perspective.
Trager and Smith (1951; T&S) provide a segmental inventory of English along
with organized observations about structure at various levels. Their stated purpose is
to show that linguistics, a social science, is not immune to the scientific method. It is
intended as a new foundation for further study, and so a few references are cited (such
as Pike and Newman, below), but no analyses are actually examined. The intent is to
start a tradition rather than follow one.
The study is of English, both broadly (US, UK, South Africa, Australia, New
Zealand) and narrowly (older generation Middle Western American), including a
lengthy comparison of the authors vocalic inventories. There are few actual instances
106
of data presented, because they are demonstrating a method of analysis rather than
defending its results. Phonetic data is generally omitted, syntax is treated sketchily
at best, and metalinguistic issues are only hinted at (p. 8). (Trager (1964) updates
this material, crediting Wells (1945) and Pike (1946) as the basis for the T&S outline.)
Their view of prominence is straightforward. One of three stress phonemes is
applied to each syllable of a word, as constrained by typical rules (one primary stress
per word and so forth). Internal plus junctures are added to distinguish the likes of
test tickle from testicle, as well as to link words in a phrase. Some primary stresses
become secondary to distinguish Lng sland from lng sland, and this new stress
becomes a fourth phoneme. A words pattern of stresses and junctures is a morphemic
superfix, and word-superfixes combine linearly to form phrase-superfixes. Pitch
phonemes (four) and terminal boundaries (three) are assigned, forming intonation
patterns which are dependent upon phrase-superfixes. These morphemes are assigned
schematic meanings along the lines of word with one primary stress, or phrase with
pitch and boundary pattern x. It is unclear at this point whether this is simply an
inventory, or a theory developing seriously around the linear structure of prominence.
T&S argue in favor of normal sentence stress being final, and some variations
are treated as being susceptible to analysis, while others are not. At first, they list some
possible strategies used for contrast, such as changing the placement or increasing the
pitch height of the primary sentential stress, or changing the ordering of two of the
higher pitch phonemes in the same phrase, or increasing the loudness in general over
some range of the utterance. They then note that:
107
...this is then the point at which we draw the line between
microlinguistics and metalinguistics: the phenomena that are
segmentable were analyzed as phonemes of one kind or another; the
phenomena that transcend segments are now stated to be
metalinguistics, matters of style, and not part of the microlinguistic
analysis. Here, then, phonology ends. (p. 52)
Later, as another variation, they posit a series of incremental shift morpheme
superfixes (p. 73), one for each number of syllables that the morpheme is to be moved
back away from the normal position. In their book, linguistic analysis is segmental
analysis. While they make such a distinction, they also support the need for what they
categorize as metalinguistics, which for them involves studies of meaning in general,
and intonational meaning in specific.
Bolinger (1951: 205n14) notes that T&S have discarded from consideration
those patterns of intonation which they classify as emotionally non-neutral, coupled
with their categorization of such patterns as not regular, which they use to suggest that
they are studying only normal, intellectual meaning. Bolinger illustrates his concern
by noting that while cold is the absence of heat, the cold of physics (only negative) is
not the same thing as the cold of physical experience (treated as having a positive
value), and that the types of instances that T&S would like to dismiss as emotionally
neutral will be treated as having positive values in their own right. Bolinger uses this
to show that the whole field needs to be looked at before any values are determined.
More importantly, it is in this same study that Bolinger really brings down the
curtain on phonemic analyses of pitch. He shows that those analysts who indicate that
there are a discrete number of relative phonemes (four for both T&S and Pike 1945)
108
must not mean that they are purely relative, but only variable within mutually
exclusive ranges. Having clarified that this is what such analysts must mean (the other
option leading to absurdities), he goes on to show that such an analysis will not stand
up to testing. Plotted intonational contours rise too steadily to lend themselves to being
cut up into discrete segments in anything other than an arbitrary fashion. There are
also intonational contours which under testing demonstrate differences in meaning,
but whose difference in contour is too slight to be described as one pitch phoneme
reaching into the range of another. Intonation is then best seen as continuous. For
Bolinger, the important part of prominence will always be as simple as up or down, but
not how far, which is just gradience.
Having said that, it should be noted that humans hate liminal things, wanting
them to belong to one class or another (Huntington and Metcalf, 1979). Red and green
traffic lights are no problem, but anticipating the onset of the amber light can cause
genuine anxiety. Similarly, there are a lot of ways to dispose of a dead body, but no
culture normally just lets one lay around in the middle of the street. A corpse is not
alive, but its not really typically inanimate either, and we dont want the rotten thing
getting up and chasing us around, thats for sure. We have to do something (Mitford,
1963).
Language is the same way. Its partly grammar-driven, and partly user-driven,
and a lot of linguists are very concerned about (not) being able to force the whole
package one way or the other. Prominence is a deeper liminal case within the liminal
case of language in general. What this all comes down to is that phonemic analyses of
prominence have all the earmarks of trying to pigeonhole liminal material. The ready
109
susceptibility of segmental material to phonemic analysis, and possibly the illusion of
pitch phonemes generated by the avoidance of the subglottal resonance, contribute to
the desire to pull prominence all of the way over onto the ordered side of the boundary.
These failed attempts to discretely portion the form and meaning of prominence make
me wary of identifying the behavior of any of the data in my analysis in too tyrannical
a fashion, and it should be plain that even to the degree that I use labels like primary
or rhematic, they are intended as poles along continua, and the spaces between them
are not genuinely devoid of data.
2.4 Prominence GOES on Units of Syntax
The division of intonation in terms of syntactic units is not musical, but it is the
logical successor to the phonological ordering of prominence seen in the preceding
section. Prominence is still not serving any real function here, but rather this research
is bent on determining where it goes strictly on the basis of information encoded
elsewhere, namely the structure of the syntactic constituency. Rhythmic studies (2.5)
will maintain this notion of an intonational default, but they will return to an appeal to
musical structure.
The works reviewed in the center of this section have a peculiar importance to
my analysis, because much of the data that I have gathered is of precisely the type that
these studies were not designed to handle. Some of them actively rejected examples
displaying volitional prominence as abnormal, or treated them as if they were beyond
the reach of linguistic analysis. As far as I was concerned, such taboo, pristine data
was irresistible. I would like to make it clear that my intent is not to portray these
110
previous analyses as wrong simply because they excluded data of this type, but rather,
I want to show that this data can now be treated as continuous with theirs, and that
while such examples may in some ways be statistically atypical, they are still frequent,
and they by no means constitute abnormal or deviant language behavior. In fact, they
fall prey quite nicely to an orderly linguistic analysis.
These analyses of syntactically predictable stress placement display a distinct
shift in this perspective over time. Newman (1946) starts by showing that two linear
strings of words differ in meaning if their stress patterns differ, and that this meaning
change can be labeled in common grammatical terms. Chomsky, Halle, and Lukoff
(1956) specifically label prominence patterns falling beyond the predictability of their
rules as deviant and abnormal, and suggest that such examples will only be tractable
when new, not phonemic, methods of linguistic analysis are developed. Chomsky and
Halle (1968) continue with the promotion of a grammar that can generate all normal
stress placement, providing a somewhat more specific and less pejorative metric for
the segregation of the data, at which point Bresnan (1971, 1972), Lakoff (1972), and
Berman and Szamosi (1972) argue about how well this transformational cycle actually
works, deciding that even normal stress must be positioned early by the user rather
than being predicted by the grammar. Bolinger makes this contention from as early as
his response to Newman (1958), consistently suggesting that the language user makes
an assignment based upon information and predictability (1972, 1987, for starters).
In Newman (1946), stress accent is presented as defined for any lexical item
or affix primarily as articulatory force, where that forces strength varies with
111
accentual contexts like loudness and pitch. Stress accent also varies widely with a
superimposed expressive accent, which modifies the stress through changes in force,
pitch, quantity, and aspiration. While stress accents are individually meaningless,
prosodic expressive accent patterns can convey a range of conventionalized meanings
for the users intent, mood, and personality. He gives examples such as a confident
tone of voice or mumbling, and social factors such as those which demand that a
church chant be monotonous. Newman sums up these observations by saying, The
system of stress phonemes and the expressive prosody constitute two accentual planes
which are constantly found to be interwoven and blended (p. 171). He would like to
show that once expressive accents have been factored out of the picture, the placement
of the remaining stress accents is predictable according to user-independent criteria.
Before starting his phonemic analysis of stress accents, Newman describes two
familiar types of expressive patterns, the first of which is a contrast of references in a
predicate. (We chased them.) This places extra force (articulatory intensity) on one
referent, shifting the peak from the end of the intonational unit back to that word, no
matter the units contour. (A unit-right-end tone peak is taken as the established norm.)
These changes are not absolute, but can be modified by expressive factors such as
those arising from emotion. The second pattern, rhetorical accent, does not shift the
intonation peak, but it does use extra force and quantity (size iconics) for elaboration.
(Look at the bones.) This is described as reflecting such equally big evaluations on
the part of the user as awe, amazement, and admiration. The choice of target in both of
these cases is held not to be predictable since it is entirely user-dependent.
112
The placement of a stress accent, however, is held to be predictable according
to its phonemic level, namely: heavy (nuclear, subordinate); middle (full, light); and
weak (sonorous, pepet). A word has a clearly defined heavy stress that appears when it
is pronounced in isolation, and When no expressive accents disturb a sequence of
heavy stresses, the last heavy stress in an intonational unit takes the nuclear heavy
stress (p. 176). Intonation and expressive accent turn a neutral, subordinate heavy
stress allophone into a prosodically heightened nuclear heavy form. Subordinate stress
placement is therefore predictable by definition, and nuclear stress is predictable by its
position in an intonational unit. Given Newmans choice of what constitutes an
intonational unit, this last part comes across as tantamount to saying that the end of an
intonational unit occurs when you think you hear a very heavy stress.
Some of these intonational units have no constituent counterparts in any theory
with which I am familiar: he brught ranges_| pples_| and peches
\
(nuclear heavy
stress on ranges and pples, the nuclei of enumerative units, and on peches, the
nucleus of a declarative unit) (p. 176). It is not that syntactic constituency should be
used to make such a determination, but rather that later studies which use Newman as
proof of the syntactic predictability of the placement of nuclear stress are mistaken in
their interpretation of Newman, as detailed shortly hereafter.
With middle stress placement, Newman does mention one construction which
appears to appeal to syntax (cf. Bolinger, 1958), namely: I have instrctions to leve
(I have instructions which I need to leave); and I have instrctions to leve (I have
been instructed to leave). The strong/weak stress pattern is associated with the noun
113
as the logical object of the verb structure, just as the strong/strong stress pattern is
tied to the verb stands in the relation of complement to the noun structure (p. 179).
Even if they are taken to support a syntactic placement theory, it should be noted that
these structures allow for a unit whose rightmost element has other than a heavy stress
phoneme (I have instrctions to leve), which means that it is a mistake to cite this
as if it supported the notion that a heavy stress has to be on the right.
Finally, while Newman decides that Expressive prosody is not necessarily
capable of the same type of systematization as that which is applicable to the usual
kinds of morphemes (p. 172), he evidently does consider them to be unusual
morphemes that are subject to a different systematization. Such notions as the prosodic
generation of the nuclear heavy stress and so forth suggest that Newman would not
consider that different analytic framework to be anything other than linguistic.
Chomsky, Halle, and Lukoff (1956; CH&L) analyze data from Newman and
T&S, intending to establish the predictability of stress from accent, and leaving out
pitch (1959: 69). They propose a single stress phoneme (accented/unaccented) which
is defined for every vowel in a word. There are two juncture morphemes, one internal
(word), and one external (phrase), which are placed at morpheme boundaries and
assigned hierarchic levels to render the desired phonetic effects. Rules assign phonetic
stress, the particular degree of loudness with which a vowel is pronounced (p. 69),
based on phonemic accent, cyclically from smaller to larger constituents. The upper
limit to the number of stress levels is user- and style-dependent. CH&L end up in part
redefining their dissatisfaction with boundaries and junctures as readjustment rules.
114
CH&L are proponents of the notion of objective normal stress, stating,
We have specifically excluded from consideration all forms of
expressive stress, including contrastive stress. In language, expressive
elements are deviations from the normal pattern. The possibility for
such deviation is, of course, enormous, and almost any stress
arrangement can occur under special circumstances. Therefore, if this
distinction between normal and expressive stress is not made at the
very outset, the number of significantly different stress levels is only
limited by the capacity of the vocal apparatus, and any hope for a
systematic account has to be abandoned. (p. 78)
This counters any suggestion that the word normal is only used here as a convenient
term for the data which works with their system; they really intend to imply that the
instances that work with their system are in fact normal objectively. They say:
It should be noted that as a consequence of our decision to exclude
contrastive stress from consideration we do not provide for the normal
stress patterns of such utterances as This is the brown house, not the
white one, where there is extra heavy stress and extra high pitch on
brown and white. The description of such utterances poses many
problems which have never been adequately handled. We feel that
these utterances are best regarded as being in a special sense deviations
from the normal pattern, and that a satisfactory description of them will
require the development of methods not currently in use in
phonemics.
11
(p. 78)
The footnote indicates that the predicted methods should be like those found in
Chomskys thesis (1955), which would seem to anticipate Chomsky and Halle (1968)
were it not for the fact that such utterances are ruled out there as well. Its not that I
have any qualms about CH&L having circumscribed a set of data for study, but I am
wary of the lack of definition for the metric used to implement the segregation, namely
115
that in their estimation some patterns are simply contrastive. This would not be a
problem were it not for the fact that no characterization of contrastive is given which
is independent of the applicability of rules, in addition to which the two types of data
are too similar to go without some sort of explicit, isolated differentiation (e.g. I like
the bok you wrote versus I like the book you wrte). This brings me to question the
appropriateness in this case of vocabulary like normal and deviations. I am not
saying that CH&L in specific is circular in its reasoning, but it does seem evident that
there is an insidious potential for circularity lurking underneath the common practice
in linguistics of using the same brains to generate the data as are used to figure out
how that data was generated.
In The Sound Pattern of English (1968; SPE), Chomsky and Halle intimate that
the user determines the position of emphatic stress before the surface structure gels,
and that the actual phonological form for that stress must be assigned after that by a
necessarily user-free rule, which, for unspecified reasons, they never specify. The
notion that the user should be able to assign different intensities of stress is not
addressed. Like the users choice of lexical items, the assignment of an emphatic stress
position is a matter of performance rather than competence, and SPE is designed in
congruence with the assumption that it should neglect matters that [C&H] have
assigned to the theory of performance (p. 25n13). Any explanations involving
discourse or context are also ruled out, because reference to a stored background
(other than an examples printed form on a page of SPE) is a memory restriction, and
so is not solely a linguistic function, but a parameter on all cognitive functions.
116
The definition of the Nuclear Stress Rule (NSR) in SPE puts primary sentential
stress on the rightmost element within a major constituent that has already been
assigned primary word stress. Now, Newman says that if a heavy stress phoneme is at
the end of an intonational unit, then it takes on the value of its nuclear allophone, but
SPE cites his view mistakenly in support of its contention that a major syntactic
constituent has a rightmost element which must be assigned primary sentential stress
(p. 90). There is a significant difference between talking about what happens if a heavy
stress phoneme is placed rightward, and insisting that such a phoneme must be placed
there. Newmans study does not say anything of the kind that SPE suggests. This
citation is all the more puzzling in view of the fact that Newmans formulation works
for his data, such as the instructions to leave pair, while the NSR as stated clearly
does not.
Bresnan (1971) tries to integrate many of the exceptions to the NSR by placing
restrictions on its ordering. The formulation of her ordering hypothesis in its 1971
version is: The Nuclear Stress Rule is ordered after all the syntactic transformations
on each transformational cycle (p. 259). Bresnan also states this less formally in the
articles abstract, saying that the NSR is ordered within the transformational cycle
after all the syntactic transformations (p. 257). In addition, she gives some proof for
the cyclicity of Relative Clause Formation and Question Formation, after which she
says that, From this demonstration and the fact that the NSR precedes these
transformations while following other cyclic transformations, it can be concluded that
the NSR is indeed cyclic... (p. 277). It would seem reasonable, therefore, to conclude
that Bresnan intends to characterize the NSR as cyclic.
117
It seems odd, then, that Bresnan 1972 should label Lakoff (1972) and Berman
and Szamosi (1972) not only as mistaken in their interpretation of the 1971 version of
her ordering hypothesis as cycle supportive, but that she would call their objections
simply inapplicable on the basis that they assume incorrectly that the NSR must,
under the ordering hypothesis, be a cyclic rule... (1972: 327, my emphasis). Bresnan
1972, however, is clearer than Bresnan 1971 in two ways. First, she says that the NSR
is not cyclic, but Cyclic (my capitalization). It is outside and after all of the
transformations in a cycle, but within the transformational domain (Cycle) which
houses that cycle, applying after the postcyclic transformations at the end of the last
cycle. That Bresnan refers to both this domain and the cycle it houses with the word
cycle only fosters confusion, and Bresnans evident surprise at the misunderstanding
it causes is in itself surprising.
Second, it comes to light that Bresnan 1971 was in no way defending the NSR,
but was simply using it to demonstrate that no matter what method was used to assign
stress, call it Method X, that stress would be assigned in deep structure, and that
Method X would be the last thing to apply to the output of any given Cycle, not cycle.
Be that as it may, as the material generated by the critics of Bresnan 1971 is covered in
the next few paragraphs, keep in mind that they are responding to the apparent
formulation of Bresnan 1971, and not its later clarification in Bresnan 1972.
Lakoff (1972) uses his treatment of Bresnan 1971 to expound upon the virtues
of global rules, noting that Once one sees that global rules are necessary, many of the
old-style transformational arguments for rule-ordering go out the window (p. 287n).
118
Lakoff states the NSR globally because even though it gets information from early on
in the list of transformations, it only applies to the surface structure, not earlier. So, the
cycle itself is supported here, and syntactic predictability is not questioned.
One small but strong point of interest for me is that although Lakoff identifies
Bury the mn the you killed, as contrastive (with chipmunk), and ...man you
klled as non-contrastive (p. 286), Bresnan (1972) does the opposite, saying I had
observed the possibility of two stress contours in examples like Mary buried those
mn she killed and Mary buried those men she klled; but I had concluded, because of
the meaning, that the latter was contrastive (1972: 337n). All of these are good
options, and none of these necessarily has a contrastive meaning. A sentence like
Bury the mn the you killed could be said in parallel to Clean up the mss you
made, where mess is not in contrast to something like sandwich. As will be shown
in the analysis proper, part of the problem here is that the same notation is being used
to mark two different levels of stress, only one of which is intense.
Berman and Szamosi (1972; B&S) step in and claim that Bresnans ordering of
the NSR makes incorrect predictions when applied to data other than that presented in
her article, and that generalizations about English prosodic stress cannot be made by
the ordering hypothesis.They maintain that Bresnan explains only her data, and so her
examples are mere non-counterexamples. B&S also provide example sentences which
suggest that stress assignment must be at least in part semantic (In 1556, what kngs
reigned? vs. In 1556, what kings bdicated?). They provide more examples which
have alternative normal stress readings, such as What book was banned? (p. 313).
While they analyze only normal, non-contrastive, non-emphatic stress (p. 312), they
119
note that the difference between a contrastive reading and a non-contrastive reading
is not so clear-cut (p. 314n). In sum, they find syntactic structure to be involved but
not sufficient as a stress placement determiner, and even at that, only at the surface,
and not in deep structure.
Bresnan (1972) then argues that Lakoff and B&S fail to see that their examples
do not counter her ordering hypothesis, but merely present arguments against the NSR
being cyclic. She agrees with B&S that the NSR itself, cyclic or not, will not work as it
is stated; for example, it places stress on shining, rather than sun, in the sentence,
The sun is shining. Bresnan proposes a rule which applies before the NSR, assigning
primary stress according to topical stress (p. 328), which is the material determined
by the user to be the new information conveyed in the utterance (which is traditionally
the comment, and not the topic.) After this rule applies, the environment for the NSR
does not exist, and so it will not generate B&Ss exceptions.
Bresnan states that B&S are wrong about surface stress assignment, because a
deep-structure topic might be deleted on the way up, and a word stressed at the surface
might not be the real topic. The classic Newman phrase instructions to leave then
has one form stressing a surviving deep topic, and one which is a consequence of the
deep topic having been deleted before reaching the surface. Bresnan ignores the fact
that B&S are addressing actual semantic (lexical, contextual) information being used
to assign stress, and concludes that they are wrong about sentence stress being
assigned at the surface, when the real argument is over what constitutes the real
topic.
120
A telling quote from Bresnan works well as a summary to this point, in which
she comments on some structurally identical sentences differing in their prominence
placement: They all argue that these are normal, non-contrastive intonations, which I
accept. But if so accepted, these examples constitute evidence not only against the
ordering hypothesis, but also against the possibility of any systematic structural
explanation of stress assignment (1972: 337, my emphasis). Precisely. There is a
consensus that while prominence is used to mark information which is important to the
language user, just what it is that the user will hold to be important is not mechanically
predictable, syntactically or otherwise, although when utterances are strung together in
context, notions revolving around new and old information can provide some strong
hints. These descriptions of information will be provided some detailed definitions in
the section on what prominence does (3).
So, to end the section with, intonation does not react well to being broken up
into a small, limited number of isolated building blocks when they have no way of
being truly continuous contours (although some amount of discontinuity might be
helpful for analyzing prominence subfunction patterns), and it does not submit itself to
lexically composed meanings which are consistently associated with specific
grammatical constructions. Even the best attempts have only managed to eke out a few
schematic meanings which end up being driven entirely by the user. Having failed to
force intonation to conform to segmental behavior patterns, researchers decided to see
what it would do when left to its own devices, which leads us to the rhythmic analyses
of prominence.
121
2.5 Prominence GOES on Beats of Rhythms
Human activity in general will tend to fall into a rhythm if left undisturbed, and
the studies in this section appeal to just these sorts of default prominence patterns. I
want my analysis to show that volitional patterns are continuous with the routine ones,
and these are the analyses which offer up rhythm as the routine extreme.
Liberman (1975; Liberman and Prince, 1977) develops a metrical analysis of
intonation, creating prominence patterns on a rhythmic basis and expecting them to be
disturbed by the users desire to move things. Liberman says this metrical NSR means
that you put the strong element on the right in any given metrical constituent, if you
have no good reason to do otherwise (p. 244), recognizing that there are a number of
good reasons to do otherwise. This view is derived from his portrayal of intonation as
ideophonic in origin, becoming metaphoric. Tying routine stress to strong metrical
elements is consistent with my analysis, where routine behavior is an attenuated form
of that which links activity iconically to prominence. It is entirely plausible to analyze
strong rhythmic feet as derivative of other forms of rhythmic articulation.
According to Pierrehumbert (1980; JBP), every well-formed pattern of English
intonation can be generated by properly spacing instances of only two levels of tones
over an utterance according to the metrical strength of the syllables. Models of English
sentence stress patterns (SPE; Liberman, 1975; Liberman and Prince, 1977) generate
the relational targets or metrical grids along which to aim the tones which make up the
intonation contours. JBP was brave enough to do the following: 1) demonstrate that
any contour can be described in terms of precisely two primitive tones; 2) develop a
122
system for generating all of the possible two-tone patterns which might underlie the
maximal set of intonation contours; 3) adjust the tone generator to produce only those
patterns which underlie well-formed contours of English intonation; 4) provide rules
for tune-text alignment; and finally 5) show how to phonetically realize the underlying
patterns in terms of pitch (fundamental frequency) as transformed into prominence
(units above the lower limit of a speakers voice, or baseline). I will digest JBP in
accord with this sequence, and then discuss JBP in regard to my own analysis.
JBPs has two elemental tones, High (H) and Low (L), which are differentiated
as follows: 1) L is always lower than H in the same context; 2) Ls phonetic value is
inversely proportional to its prominence (as defined below), while Hs phonetic value
is directly proportional to its prominence; 3) the intonation contour between L tones is
interpolated as a straight line, whereas between H tones it dips. A more concrete pitch
height value for H or L is computed according to: 1) the relation of H or L to preceding
tones; 2) the relation of H or L to the baseline (Ss lower voice limit); and 3) the degree
of prominence that the user assigns to H or L. My own work systematizes some of the
variations in user-assigned prominence that affect this height value.
JBP uses these two tones in the definition of three types of accent, namely
PITCH ACCENT, PHRASE ACCENT, and BOUNDARY TONE. A pitch accent can either be an
isolated tone, which is aligned with a metrically strong syllable, or it can be a pair of
tones, where one of its two halves is aligned with a strong syllable, and the other half
precedes or trails this strong alignment. An isolated tone representing a pitch accent is
marked with an asterisk (H* or L*), as is the strong member of a bi-tonal pitch accent.
123
Bi-tonal pitch accents can be any heterogeneous L/H pair, linked with +, with
a superscripted on the non-central half. H*+H

is supported by cases in which it

contrasts with H*, as when the tonal spread from the trailing H

protects an intonation
plateau that would dip between a sequence of H* tones. The other patterns are possible
theoretically, but there are no contexts in which they are contrastive with any others.
Unnecessary, they are treated as nonexistent by JBP. A phrase accent (L

or H

) cannot
be bi-tonal, landing on the main phrase stress immediately after the pitch accent. It is
an expressions last pitch accent, which controls the intonation contour from itself to
the boundary tone (not bi-tonal, L% or H%), which goes at the expressions end no
matter what the metrical structure of the text. It can be used as a medial utterance
boundary when an expression is divided into one or more phrases (breath groups,
essentially), and in this guise it appears at the beginning of an expression after a pause.
An expression can start with a boundary tone after a pause, picking up one or
more pitch accents, then a phrase accent, and a final boundary tone, which terminates
the expression or separates it from the next phrase (as adapted from JBP, p. 13):
Figure 2-3: Tone Pattern Generator
... [Boundary Tone] + [Pitch Accent(s)] + [Phrase Accent] ...
[...........H*..........]
[..........L%..........]....[.......L

+H*.......]....[.........L

..........
[........H*+L

........]
[........H

+L
*
.......
[........H*+H

........]
[...........L*..........]
...........H%.........]....[.......L* +H

.......]....[.........H

.........
124
Here are two single accent choices for two contours of Anna (p. 146, fig. 1.1 A-B),
and one multiple pitch accent choice for Another orange (p. 148, fig. 1.2 A):
Figure 2-4: The Tone Pattern Generator in Action
The intonation pattern for tag expressions requires that one of two modifications be
made to the tone pattern generator, the first of which allows an optional additional
phrase accent to appear before the boundary tone (p. 202, fig. 2.46):
time
200
250
f
r
e
q
u
e
n
c
y
300
Anna
0.4 0.4
Anna
350
| |
H* H* L

L%
L% H

+
H*
L

L%
H* L

L%
H

+
time
150
200
250
f
r
e
q
u
e
n
c
y
Another orange
0.4
H*
L

L%
H*
H* L

L% H*
| |
125
[2-8] Im sorry, Benjamin.
|
H% L* L

H%
The original and the additional phrase accents are L

in this case, but in others they can

be any L

/H

combination. The second change allows for an additional phrase accent

and another optional boundary tone to follow an earlier boundary tone (p. 51, ex. 17):
[2-9] Good heavens, Joe muttered.
| |
H*L

H% H*L

L%
Again, the two boundary tones can come in any combination.
This tune is aligned with a metrical representation of an expressions syllable
strength (Liberman, 1975; Liberman and Prince, 1977). Pitch accents go with strong
syllables, and the phrase accent goes with the syllable following the pitch accent that
falls on the main phrase stress. Boundary tones are simply placed at both ends of the
expression. This example gives possible tune-text alignments where the placement of
the main phrase stress is dependent upon user highlighting (p. 19, ex. 26a-d):
[2-10] (a) Its organized on the model of a gallon of worms.
| |
H* H*L

L%
(b) Its organized on the model of a gallon of worms.
| |
H* H*L

L%
(c) Its organized on the model of a gallon of worms.
| | |
H* H* H*L

L%
(d) Its organized on the model of a gallon of worms.
| | | |
H* H* H* H*L

L%
126
This next example uses bi-tonal pitch accents (p. 26, ex. 44):
[2-11] I really believe Ebenezer was a dealer in magnesium.
|+ |+ |+ | +|
H* H

+L* H

+L* H

+L* H

+L*L

L%
The alignment rules are the same for single and bi-tonal pitch accents, except bi-tonal
pitch accents specifically align one tone of their pair with a metrically strong syllable.
The question, then, is: what actual sounds do these representations represent?
To begin with, JBP defines a number of rules designed to interpolate a contour
between accented syllables when fed their assigned tones, and three of these rules
stand above the rest. The first is a rule of tone spreading, where any T

(any trailing
tone or phrase accent) spreads to the right across any intermittent unaccented syllables
towards a tone that is of equal or greater value. The second rule maintains that the
contour between L tones is a nice, straight line. The third and final rule requires the
contour to dip between H tones.
This last rule is motivated by processes of low-tone elision in some African
languages, when the low tone triggered a downstep. The dip between Hs is
underlyingly an H*+L

H pattern, where the L

triggers a downstep and evaporates.

This rule applies in a case such as the following high-mid calling contour:
[2-12] ______
..
.....
...______
...Tommy
...H*+L

L%
127
The pitch accent has a high center and a low trailing tone, where that same L

tone is
required by rule to lower the following H tone before that very L

tone is deleted.
On top of these rules is a transformation where intonation as measured in terms
of pitch (fundamental frequency) becomes represented as perceived PROMINENCE, the
value of which is given in UNITS ABOVE THE BASELINE (p. 213, fig. 3.10 A-B):
Figure 2-5: Baseline Units
Every speaker establishes a baseline frequency (lower voice limit) at the beginning of
every phrase, and the contour will approximate this baseline asymptotically over time.
Several H* tones over time might decline in pitch but have the same (or an increasing)
prominence.There is, however, research which suggests that this DECLINATION applies
inconsistently or not at all to some speakers (Lieberman, 1986: 255).
As far as the relevance to my own work is concerned, JBP shows that English
intonation contours can be analyzed into precisely two tones assigned to an utterances
salient elements (strong syllables, post-peak syllable, and boundaries), as long as you
time
70
120
170
220 f
r
e
q
u
e
n
c
y
1.5
1.0
0.5
0.0
b
a
s
e
l
i
n
e
u
n
i
t
s
128
have the right rules needed to interpolate a contour between those tones. (This might
just project hmans mathematical certainty onto a linguistically real system.) The
metrical grid defining those elements is a null hypothesis, to be applied (in its creators
own words) only if you have no good reason to do otherwise (Liberman, 1975: 244).
My work systematizes some of these good reasons, but JBP is not really interested in
how an element is identified as salient so much as locating elements that have already
been so identified. The results of my work would merely provide JBP with a different
source of information for locating salient elements for tonal assignment; likewise,
JBPs work would be more relevant to my analysis if I were trying to figure out how to
generate a tonal contour given a point of volitional prominence.
Hayes (1995) is the perfect study for rounding out this section on the rhythmic
patterning of prominence points as an intonational default, beginning with his central
claim that stress is the linguistic manifestation of rhythmic structure, and that the
special phonological properties of stress can be explicated on this basis (p. 1). Simply
put, stress goes on the beat. This flows well into the paired contentions that:
pitch is directly determined by the intonational system, but rules linking
tones to texts refer to the position of stress. As a result, pitch can serve
as a powerful phonetic cue for stress location. However, in locations
where the intonational system places no tones, pitch cannot serve as a
cue for stress, and other cues such as duration take over.... (p. 11)
and
My view instead follows that of Pierrehumbert 1980, in which phrasal
stress is an independent domain, and pitch accents are constrained to
attach to the strongest available stresses. (p. 370)
129
In other words, stress not only goes on the beat, but if there is an available pitch, it will
locate a point of stress, otherwise duration, loudness, and the like are the most readily
identifiable associates of that pitchless stress.
As far as physiological correlates of stress are concerned, Hayes says that, if a
breath pulse is present, it is probable that stress is present, but there are many stressed
syllables produced without a breath pulse (p. 6). In addition:
Naturally, certain phonetic correlates serve more readily as cues for
stress than others: it is only natural that strong rhythmic beats should
coincide with breath pulses, with greater duration, and with raising of
pitch. But these are only tendencies, and since phonology serves many
ends other than rhythmic ones, they sometimes override the natural
correlation between strong rhythmic beats and particular phonetic
phenomena. (p. 9)
As usual, the physiological correlates of stress are portrayed as highly variable (while
the fact of its perception is stable, 1.3). As far as levels are concerned, Hayes finds
three necessary, motivates four and five, and suggests that there may be more, subject
to the constraint that too many levels will end up being indistinguishable, or the
differences between them will be masked by other sounds (as in SPE, cf. 2.4.).
All of the rhythm researchers would be in agreement with Hayes when he says,
the [NSR] represents... only the phonological default, and may be overridden by
many factors, including focus marking and predicate-argument structure. Non-default
phrasal stress assignment is an enormously complex area (p. 369). Rhythm is then a
default with an enormous scope: It has been conjectured... that the principles of
eurhythmy are invariant across languages, and that they may extend beyond language
130
into other cognitive domains (p. 372). It is consistent with the findings of this study to
treat speech as gesture and motion, and to expect that intonation will be susceptible to
the same sort of rhythmic defaults which are applicable to other human activities such
as walking and breathing.
2.6 Prominence GOES on the Beat unless it is Making Sense
Since the 1500s, intonation has been treated as a contour whose critical points
are aligned according to rhythm when not disturbed by sense. Attempts to analyze
intonation into a static set of subcontours never does any better than to suggest that the
primary contour has a beginning, a middle, and an end, just like the sense which it
follows, in addition to which these subcontours keep sliding towards an approximation
of points of sense or rhythm. When it comes to where prominence goes, volitional
prominence would not be a matter of level so much as placement for reasons other
than the rhythmic default.
This is one of the reasons that the image appealed to me of the high-tension
lines draped over towers dominating a rolling landscape, because although towers are
placed at regular intervals over level ground, that pattern will be broken in order to
maneuver lines around obstacles, or to place towers on high ground in order to lift the
lines clear of intervening terrain features. Its tempting to just relax and enjoy this
bucolic landscape, where intonation is no more complex than up and down, and where
prominence simply goes here or there, but this rest is premature because sense has
yet to be given an adequate definition.
131
3 What prominence DOES
The research in the previous section treats prominence as redundant, as if it
provided no information that had not already been encoded through the function of
some other set of elements in an utterance. In this section, prominence is actually
providing or altering information not otherwise made available, and so prominence is
meaningful. The description of this meaning has been thoroughly refined over time,
being portrayed in turn as: semantic (3.1); contextual (3.2); pragmatic (3.3);
discourse functional (3.4); and finally in terms of outward and inward focus (3.5
and 3.6). No matter how the function of prominence is characterized, the common
ground is that the user appeals to volitional prominence when the grammatical or
rhythmic routine does not clearly convey the right sense, where sense is an
unmolested congruence of meaning from S to L. (Trust me.) These are the studies that
showed what volitional rather than routine prominence was needed to mean.
3.1 Prominence DOES Semantics
Bolinger has been an advocate of semantic primacy since at least as early as
1957, where he warns against overextending intonation segmentation studies, quoting
Pikes (1945) similar reservations. He then detaches syntax from stress, walking
through a number of counterexamples to the syntactic predictability claim, concluding
with the statement that, The encounters between intonation and grammar are casual,
not causal. Grammar uses intonation on those frequent encounters, but intonation is
not grammatical (p. 37). Prominence acts as a clue, but not as a specific identifier.
132
Bolinger (1958) continues along the same line in disagreeing with Newmans
analysis of stress accents, showing that there is a positive correlation in Newmans
examples between the syntactic structures that he identifies, and the informativeness
of the item which receives stress. Bolinger says, It is true that in the instances cited,
the stress tells us something about the syntactic relationship. But is this identifying
function essential to it, or is it like a green hat worn by a thief, which may help us to
identify him, but which he may change at any time? (p. 7) The intended meaning of
an utterance should be apparent through some means other than stress, and stress
functions other than to merely identify a particular grammatical construction.
Bolinger then identifies several cases of the strong/weak stress pattern rather
than the strong/strong stress pattern which are also used with what Newman labels as
the noun as logical object of the verb structure, showing that there is not a unique
correlation between syntactic patterns and stress patterns. This draws Bolinger to the
conclusion that when it comes to Newmans examples:
it was not the construction, but the informativeness, that determined the
stresses. In the narrow but frequent contexts typified by Theres a...
to... and Ive got a... to... we usually find the thing pictured as
constituting an incentive to perform the action that would normally be
performed upon it anyway (bread has little other purpose than to be
eaten, whence eat in such a context is relatively more redundant than
bread) or that it is foreknown to speaker and hearer as likely to be
performed.... (p. 8)
This identifies relative informative weight as the determining factor in the placement
of stress, and it dissociates any specific syntactic pattern from any specific stress
pattern. A stress pattern is seen as being only indirectly correlated with a particular
133
syntactic structure, if at all. It is correlated only if the relative informativeness of an
item happens to be linked to a reasonably predictable position in a particular structure.
Bolinger has long used accent instead of stress to avoid propagating those
problems which continuously arise from confusing various types of stress with one
another, and stress is only used to refer to a potential, namely the syllable in a word
that would normally be accented were that word to receive accent. Stress belongs to
the lexicon. Accent belongs to the utterance (1972b: 644). This is why Bolinger
claims that the placement of pitch accent is not syntactically predictable, but says that
the placement of stress is explainable in this way. He is not suggesting that the
placement of prominence is syntactically predictable, even though stress is.
Whereas Bresnan focusses on when prominence gets assigned, Bolinger
(1972b) is concerned with what gets pitch accent, and how it gets it, maintaining that
accent is assigned according to user intent, and not according to syntax:
The Chomsky-Halle Nuclear Stress Rule and its modifications by
Bresnan, and to some extent the criticisms that have been leveled at it,
have in common an attempt to account for accent in terms of syntax.
Instead, accent should be viewed as independent, directly reflecting the
speakers intent and only indirectly the syntax. Accented words are
points of information focus. (1972b: 633)
The distribution of sentence accents is not determined by syntactic
structure but by semantic and emotional highlighting. Syntax is
relevant indirectly in that some structures are more likely to be
highlighted than others. But a description along these lines can only be
in statistical terms. (1972b: 644).
The syntactic and semantic views are not really at odds, given that Bresnan assigns
prominence on the basis of topic rather than the NSR, where Bolinger points out that
134
Bresnans topic is more commonly called comment or newer information. As far as
when pitch accent gets assigned, Bolinger does not address the ordering hypothesis
directly. He notes that the speaker makes word choices very early, but his early does
not mean deep. These word choices help determine later accenting, which supports
prominence being considered early.
Gussenhoven, Bolinger, and Kjeispur (1987) collect articles exchanged by
Gussenhoven and Bolinger in the Journal of Linguistics (ca. 1983), and expand it to
include one more volley apiece, with final commentary by Kjeispur. Gussenhoven has
his Sentence Accent Assignment Rule (SAAR) break a sentence into candidate focus
constituents (based on old versus new information), one of which the user selects for
focus, after which the accent within the constituents is mechanically determined by the
SAAR. Bolinger has no need for intermediate focus domains of any size, much less
multiple domains of varying size, saying that the user places accent directly on words,
expressing focus or other concerns, and so there is no need for a mechanical rule of
sentence stress assignment. For example, Bolinger (1986: 95) differentiates a focus of
interest from a focus of information, which do not always overlap because the most
informative material in an utterance is not necessarily the most interesting (Wh did
he lave?), in which case both foci are accented. So, Gussenhoven portrays accent as
conveying meaning at best indirectly, while Bolinger has the user applying accent to
assign meaning directly.
Kjeispur tries to reconcile the two by shrouding their focus differences in their
mutual appeal to the use of old and new information to determine focus. Here is one
135
example she uses in illustration: Whats that fellow doing? Hes looking for BREAD
to eat. Gussenhoven would put the whole phrase bread to eat in focus because it is
all new relative to the first sentence, and then his SAAR would mechanically assign
accent to bread alone. Bolinger would suggest that perceptions of old and new track
through the context, and that once you hit bread, the notion to eat is expected, and
since to eat is old or predictable, only bread has an accent, with no recourse to
an intermediate focal constituent. Both theories rely upon the users evaluation of the
age of the information in an utterance to determine focus, and it is interesting to point
out that this reliance has been a disregarded similarity; however, because there is no
need for the SAAR other than to reach beyond a limit placed on the users influence by
the SAAR, and because Bolinger supports the users choices right through the direct
placement of the accent, there is a difference between user-driven and grammar-driven
accent which leaves their incompatibility significant.
Relative to Bolinger, sense keeps the user from surrendering to a syntactic
default, marking elements that are interesting, informative, or otherwise deserving of
special attention, not as specifically tied to anything else in the discourse. Bolinger
suggests that when contextual information places prominence, volitional prominence
proper is not needed, even when that context is not the current discourse, but the
Archive (bread has little use other than to be eaten). In that sense, the user never really
defaults to syntax, since there is always some background information influencing
routine prominence placement. Marking this information preserves Ss meaning to L,
where a syntactic default might permit an inadvertent misalignment of attention.
136
3.2 Prominence DOES Context
This section marks a switch from prominence described as marking relatively
important points in what amounts to an isolated sentence, to its doing so relative to the
information in a series of sentences. This is not the same as assigning an actual
discourse function peculiar to prominence, which happens in subsequent sections.
Gunter (1966, 1972) develops a context grammar [which] is the study of
features of the sentence that reach back to context, thus binding sentences together into
larger coherence (p. 165). This provides the basis for analyses of ellipsis, reference,
the temporal relation between sentences in a dialogue, and the use of accent placement
as a grammar signal to indicate whether a sentence is pointing to an overt or non-overt
context. Although he is not studying points of prominence, he still finds contextual
comparisons equivalent to some types of revelation, but no elaboration analogs, and
emotion is trivialized. So, there is at least a partial, clear, similarity between context
grammar and my analysis, particularly in the motivations for subtyping revelation by a
discourse counterparts explicitness. In specific, Gunter states (all in small capitals):
Any variety of a sentence in live speech is connected through accent
placement... to a context, which may be overt or non-overt. If there is
connection to an overt context, the accent placement specifies what the
connection is; if there is a connection to a non-overt context, the accent
placement signals that fact, and simultaneously limits the form and
content of that non-overt context. Accent placement, like all other
features of context grammar, reaches back to a foregoing sentence, and
shows the relevance of the response to its context. (p. 167)
While I developed addition and substitution (his REPLACEMENT) before reading Gunter,
his notion of RECAPITULATION did not occur to me until I came across his examples. In
137
recapitulation, L repeats Ss entire sentence exactly, with a shift in accent placement
for confirmation of one word. My data only show brief examples of this, just short
strings of words. That might be due in part to the lack of any questions in my data,
much less questions which follow Gunters precise format.
In Gunter (1972), intonation signals contextual relevance. All that matters is a
contours gross shape (falling, high-rising, falling-rising, or low-rising), and not pitch
levels defining their endpoints (e.g. 42 = 41 = 21). Contours fit monosyllables, or
are stretched to fit longer utterances. Gunter presents a taxonomy in which each of the
four contour types is associated with a specific function, where variations on a contour
do not change the function, but rather are seen as rendering unstable emotional flavors.
Such forms of expression are stated to be a proper object of linguistic study (p. 203),
but just not his study.
So, Gunters analysis does tie prominence to context; however, 1) it only
studies process and not point prominence; 2) it dismisses emotion or expression (while
noting that they are influenced by gesture); and 3) it does not address volitional
prominence levels. Even at that, the notion of sense does start to take on one of its
more important qualities, that of preserving meaning congruence across contexts,
where different contexts are equated with the meanings understood by S and L.
3.3 Prominence DOES Discourse
Bardovi-Harlig (1983; BH) analyzes major stress beats in intonation contours,
not the contours themselves, and not the phonetic characteristics of those beats, only
their placement. English sentence stress is demonstrated to be context sensitive, where
138
the sentences functional organization determines its placement by fitting the utterance
into the discourse. BH finds two stress-and-discourse-function associations, namely
that primary stress is always placed on the RHEME, and secondary stress is placed on
the THEME, when there is one: the theme relates the utterance to preceding discourse;
the rheme is the portion of the utterance which most advances the discourse (p. 7).
Almost as much of her analysis is devoted to making the definitions of these functions
less brief as it is to reassimilating stress patterns which have traditionally been cast
aside as matters of contrast.
BHs strategy follows these steps: 1) quickly and quietly dispatch any analyses
which conclude that sentence stress is syntactically predictable; 2) critique the limited
success of previous functional analyses of sentence stress in order to synthesize
stronger definitions of theme and rheme; 3) propose a formal stress assignment rule
for theme and rheme which explains previously problematic data in a unified fashion;
4) use this rule to dispel myths surrounding prominence (such as that it marks focus,
that it is influenced by syntactic categories, that it cannot mark theme or pronouns);
and then 5) identify contrastive stress as a category which is not supported by the data,
a garbage can into which intractable patterns have been tossed for decades.
The first of two functional analyses of stress which BH cites as influential is
that of Dane (1960), who says that intonation and prominence function to tease an
expression into position in its discourse context. Dane proposes two functions: THME
(theme); and PROPOS (rheme). The rheme occurs at the sentences end, and stress is the
CENTER OF INTONATION, falling on the last word of that rheme. This overgeneralization
139
primes a sense of circularity when the rheme is not found at the end of an utterance,
because the rheme is determined to be wherever the center of intonation has fallen.
There can be more than one center of intonation, the last one of which is hierarchically
superior to the rest, unless of course the real center (the one which identifies the
rheme) happens to be the first one in the utterance, in which case the whole phrase is
emphatic. Finally, further emphasis can be placed on a word in final position.
BH has five complaints: 1) Dane studies rheme to the neglect of theme; 2) too
much emphasis is on word order, as reflected in the rule placing the rheme normally at
the end of an utterance; 3) this emphasis disallows a more general rule of functional
stress assignment; 4) any non-final stress assignment is contrastive; and 5) there is no
part of the definition of rheme which is not dependent upon the definition of stress,
resulting in a circular relationship between the rheme and the center of intonation.
The second influence is Schmerling (1973), who has four (Roman-numeraled)
stress-assignment rules: I) stress cannot fall on any items which S deems relatively
insignificant (p. 73); II) arguments are stressed over predicates; III) the last among
equal stresses is the strongest (which adjusts the output of Rule IV); and IV) both topic
and comment are stressed when they exist. Rule II applies only to news sentences, just
as Rule IV obviously applies only to topic-comment (t-c) sentences.
BH expresses some preference for Schmerling over Dane because Schmerling
attends to theme, even though she characterizes it no more strongly than as pragmatic
aboutness. BHs complaint is that the stress pattern of a sentence is used to decide
which of Schmerlings rules must have applied. For example, the generic sentences:
140
[2-13] (a) Great aks grow from little corns.
(b) Great aks from little corns grow.
are both identified by Schmerling as news sentences because they follow Rule II,
despite their typically being regarded by other researchers as t-c sentences. BH spends
some time showing how increasingly complicated it would be to alter Schmerlings
system to accommodate this distinction between news and t-c sentences (such as
having sub-Rules for transitive t-c sentences), and at the peak of having created a real
mess, she mentions just how much easier it would be to use a system which already
works in a straightforward manner (such as her own) rather than to waste time trying
to patch up Schmerling. A more detailed comparison between Schmerlings rules and
BHs stress assignment principle appears later in her analysis, and in this review.
BH switches from a review of functional analyses of sentence stress to those
which develop the functional framework in general, beginning with the work of Kuno
(1972, 1977, 1980). Like Schmerling, Kuno characterizes theme in terms of pragmatic
aboutness, and where Schmerling has t-c and news sentences, Kuno has thematic
sentences and neutral descriptions, but the similarities end there.
BHs review of Kuno reveals a proliferation of terms that were left inconsistent
in accord with the philosophy that it is better to have loose definitions that work, rather
than tight ones which dont. Kuno ends up with seven overlapping theme types arising
from pragmatic aboutness, namely: 1) contrastive; 2/3) un/predictable; 4) prominent
(primary); and 5/6/7) latent/hyper-/secondary theme. Focus is defined as newer
information, but the focus can also sometimes be the topic, and the topic can also
141
sometimes be the theme, which is not new at all, but rather is defined as being
discourse anaphoric. BH complains that Kuno pays attention to sentence-initial stress
and theme in neglect of rheme (opposite to Dane), and so he does not explain
comment or primary stress.
As representative of the Prague school and their functional framework, Firbas
(1964, 1966, 1975; see also 1992, post-BH) comes closest to what BH is looking for,
studying the whole sentence in context, thus paying attention to theme and rheme, as
well as to the functional elements which link the two (called TRANSITIONAL elements).
The rheme has a high degree of COMMUNICATIVE DYNAMISM (CD; BHs FUNCTIONAL
VALUE), and the theme has a low value. CD is influenced by: context (in)dependence,
the distinction between given and new, predictability, semantic relations (presentation/
specification), and definiteness (p. 15). The rheme is newer information, that which is
the least predictable on the basis of the information already present in the context. In
support of this definition BH cites Prince (1981), Chafe (1974), and Firbas (1975,
1981). Rhematic material tends to be high in context independence.
The theme is context dependent, where the definition of dependence draws
upon: 1) the familiar notion that some information is already GIVEN; 2) Bolingers
(1972) PREDICTABILITY (dependence is higher for predictable items); 3) Kunos (1972)
DISCOURSE ANAPHORIC material; and 4) Kunos (1977, 1980) TOPIC/THEME/FOCUS. Prince
(1981) is also cited in reference to material which is dependent upon the information
in the context from which it is EVOKED or INFERRABLE. The theme is also likely to be
definite or generic, and its unmarked position is as the subject of a sentence.
142
BH draws upon two semantic scales distinguished by Firbas to identify some
of the characteristics which are used to measure a words degree of CD, and to set up a
standard of RHEMATICITY against which to compare some examples later on in her
analysis. The first is the appearance/existence (A/E) scale:
Table 2-2: Firbas Appearance/Existence Scale
Here are examples where each [bracketed] rheme is the phenomenon which appears or
exists on the scene, and so has the highest degree of CD on the A/E scale (p. 17):
[2-14] Over the veranda, prone on the paddock, flung over the fences
were [exhausted-looking bathing dresses and rough striped
towels].
[2-15] In the center of the room, under the chandelier, as became a host,
stood [the head of the family, old Jolyon himself]. (1975: 50)
On the appearance scale, the verb expressing appearance or existence is transitional,
the item representing the phenomenon appearing on the scene is the rheme.... The
rheme may be either the object... or the subject (p. 16). The clothes appear or exist
over, prone, and flung (lower CD elements which set the scene), and old Jolyon
appears or exists in, under, and as, portraying both the swimming attire and the
man with high CD values relative to the lower CD material describing them in context.
Degree of Communicative Dynamism
low medium high
element setting the
scene (setting)
element expressing A/E
on the scene
element representing the
phenomenon A/E on the scene
143
The attributive scale measures CD as follows:
Table 2-3: Firbas Attributive Scale
On the attributive scale, the attribute or the quality expressed is the rheme, where,
Adjectives, noun phrases, and verbs may all express qualities (p. 17):
[2-16] Mary is [beautiful].
[2-17] His father was [a musician]. (permanent quality)
[2-18] His mother played the piano [extremely well]. (just temporary)
BH goes on to say, Further specification of a quality, if present, will be the rheme; as
in example [2-18], but if it is not present, the quality or attribute is the rheme, as in
[2-17]. The theme bears a quality, and that quality is the rheme.
Firbas (1992) presents a new version of the functional sentence perspective,
proposing two CD distributions in an utterance, the first of which is its non-prosodic
CD distribution (p. 148) as determined interdependently by its linear modification,
semantic content, and context, and the second of which is the placement of prosodic
prominence (PP) or at least four degrees of PP (p. 143), specifically differentiated
as: 1) a lack of stress; 2) a stressed syllable inside the prehead or tail (stress without
accent); 3) a stressed syllable inside the head (stress with accent); and finally 4) the
one such stressed syllable in the head which acts as the nucleus of the utterance.
Communicative Dynamism
Low Lower Medium Higher High
scene (setting) bearer of
quality
quality specification
of quality
further
specification
144
The alignment of these distributions affects the interpretation of the overall CD
of the utterance. When in perfect correspondence (p. 148), prosodic prominence
reflects and amplifies the non-prosodic CD, giving additional meaning which
Firbas describes variously as intensification, emotive coloring, and special
emphasis (p. 154f). These distributions lack correlation when prominence provides
selective non-reevaluating intensification (p. 156), assigning a theme (as identified
by non-prosodic factors) greater prosodic CD than some rheme (likewise), where this
overly important theme remains just a theme. When the prosodic CD imbalance is so
severe between a non-prosodic theme (with high prosodic CD) and a non-prosodic
rheme (with virtually no prosodic CD) that the theme is re-evaluated as a rheme, then
the CD distributions misalign due to re-evaluating prosodic intensification (p. 159).
Firbas stresses that these prosodic and non-prosodic factors are equally interdependent
when it comes to determining CD, but prominence has the final vote.
By synthesizing the best parts of the previous analyses of sentence functions,
BH is making good progress towards providing strong, unified definitions for rheme
and theme. In addition, every complete utterance is defined as having a rheme, while a
theme is optional. The catch is that theme and rheme are not functions as such, in that
they do not cause specific effects when applied to a word (such as turning it into a
topic or comment, or bestowing upon it a particular level of prominence). Rather they
are roles in the sentence, which other linguistic functions help to identify by
determining their CD; for instance, when a word provides new information, this can
help to identify the rheme. Firbas attributes a critical part in this role-identification
145
function to prominence, in that once these other factors have provided a word with the
potential to be a theme or rheme, then prominence can sort them out. So, definitions of
theme and rheme which give the impression that they are functions in themselves
should be understood to mean that they are identified by other factors as roles. It is that
identification which is used to maintain sense as it has been described to this point.
3.4 Prominence DOES Theme and Rheme
Having established substantial definitions for theme and rheme, BH goes on to
show why they are better than the alternatives.
To begin with, Schmerlings Rule IV assigns stress to both the rheme and the
theme (without a theme, Rule II applies), and then Rule III makes the stress on either
the rheme or the theme primary, whichever is sentence-final (the rightmost of equally
strong stresses). In contrast, BHs PRINCIPLE OF FUNCTIONAL STRESS ASSIGNMENT (PFSA)
simply assigns secondary stress to the theme when there is one, and primary stress to
the rheme always, without regard to its position. The PFSA can be supported over the
Rule {(IV + III), (II)} system with examples of utterances in which an early rheme has
the primary stress, and a late theme has the secondary stress, such as in, Who hit
Judy? Pnch hit Jdy. At best, Schmerling would need to introduce sub-Rules.
Schmerling identifies different stress patterns for t-c and news sentences, and
so has different stress assignment Rules for each of them. The PFSA treats both as
context sensitive, and it is just one rule which applies to both equally, assigning stress
according to their functional organization. The examples used to make this distinction
in both studies are as follows:
146
[2-19] (a) Trman ded. [Schmerling: t-c > (IV + III); PFSA: died =
rheme] (part of an on-going discourse about Trumans health.)
(b) Jhnson ded. [Schmerling: news > II; PFSA: Johnson =
rheme] (out of the blue.)
Schmerling needs two rheme stress rules, one with and one without an accompanying
theme stress rule, and it is only by determining the sentence type that the appropriate
rheme rule is accessed (II or IV). The PFSA is stated as if it were one rule, but it acts
as two, being applied as the rheme and theme are identified. Schmerling goes through
sentence type to get at function, needing a redundant rule, and the PFSA economically
goes straight for the function. Schmerlings stress-pattern division is artificial, and the
real division is simply whether or not the thematic function exists in an utterance.
BH relies on Liberman (1975) for the precise assignment of stress within a
rheme which is multisyllabic (Pnch hit [Jdy]) or multiword (Pnch hit [Judy and
the Dvil]). Liberman suggests that the metrical system should be applied only if
you have no good reason to do otherwise (1975 [1979: 244]). Rheme stress only goes
so far, after which there is no good reason to do otherwise than resort to a relational
assignment of stress within a rheme.
BH also notes that any word can be the rheme, so any word is a candidate for
primary stress, however, the nature of discourse tends to restrict this tolerance. No
stress pattern is disallowed outright, only in a particular context, namely one in which
the proposed candidates for stress would not be rhematic in context. In the absence of
context, such as in a typical presentation of linguistic examples, BH once again applies
stress according to the null hypothesis, or metrical grid.
147
BH goes on to reanalyze data approached in Bresnan (1972), showing that
Bresnans problems came from trying to identify stress patterns in syntactic terms
rather than according to functional organization, and that her successes only occurred
when the two happened to overlap. Bresnans use of a functional rule (Topical Stress
Assignment) to take care of exceptions to her ordering hypothesis is an admission that
syntactic rules only work when there is an overlap with function, because an appeal to
a functional rule is required when no such overlap exists.
This is a recurring theme throughout BH: discourse functions are independent
of word order, but they tend to be distributed theme-left and rheme-right, with the
degree of CD likewise tending to increase as the sentence progresses in alignment with
function, not syntax. It is this tendency that allows syntactic analyses of sentence
stress to often be correct by coincidence, and it explains why many of them appeal to
semantic or functional explanations when the syntactic base collapses, that is to say,
when the functional order and the syntax do not coincide.
Now BH is ready to use her system to demythologize stress.
Myth #1: Syntactic Category Influences Stress
Truth #1: Prominence doesnt Prefer Content over Function
Ladd (1980) proposes a hierarchy in which grammatical category determines
accentability, where nouns are the most accentable of content words, and content
words are more accentable than function words. This continuum exists in the same
system in which broad focus is unmarked or normal compared to narrow focus. This
leads to inconsistency, as in the following examples (ignoring any theme stress):
148
[2-20] (a) (Why are you so terrified?) My prents are coming.
(b) (I think your parents will stay home.) My parents are cming.
[2-21] (a) (I like one of the Beatles in particular.) Jhn is wonderful.
(b) (Im so in love!) John is wnderful.
[2-20](a) has noun stress and broad focus according to Ladd, and yet the noun stress of
[2-21](a) leads to a narrower focus than the adjective stress in [2-21](b). If normal is
measured in terms of the broadest focus, then syntactic category cannot be used
identify normal.
Schmerling says that such examples are problematic for information theories:
[2-22] (a) John is a wnderful man.
(b) John is a wonderful mn.
If man adds no information beyond that specified by John, but still gets primary
stress, then the primary/new association is wrong. There are two functional solutions:
1) man can stand alone as the rheme when it specifies a further quality of John, as
in, John isnt just a wonderful linguist, John is a wonderful mn; and 2) when
wonderful man is a multiword rheme, man gets primary prominence metrically.
It-cleft sentences are often used to support stress as following mobile syntax:
[2-23] It is Jhn who writes poetry in the garden.
Prominence hasnt followed John as the formerly-rightmost constituent, but it goes
on John the rheme. In context, prominence could just as well go on other elements.
For the functional approach, these are all just cases of theme and rheme stress.
149
Myth #2: No Theme Stress
Truth #2: Prominence doesnt Avoid the Theme or Topic
Here is BHs answer to the claim that the theme or topic cannot be stressed:
[2-24] (a) The statues had is mssing.
The theme gets secondary stress, and the myth is easily countered by this data.
Myth #3: No Pronoun Stress
Truth #3: Prominence doesnt Avoid Pronouns
Her answer to the claim that anaphoric pronouns cannot be stressed is similar:
[2-25] Punch hit Judy and then sh hit hm.
Once again, the data belies the myth.
Myth #4: Prominence is Contrastive Stress
Truth #4: Prominence doesnt do Contrastive Stress
BH presents several arguments in favor of strongly curtailing the use of the
term contrastive stress, the first of which is that researchers can take a look at
precisely the same examples and come to contradictory classifications, as follows:
[2-26] (a) Bury the man that you klled.
(b) Bury the mn that you killed.
As mentioned earlier, Lakoff (1972: 286, cf. p. 118) finds [2-26](a) to be contrastive,
but not [2-26](b), whereas Bresnan (1972: 337n12) does just the opposite.
150
BH goes on to note that contrastive stress tends to be abused as a label for
intractable data, where these exceptions tend to fall into specific categories, namely:
1) non-rightward placement:
[2-27] (a) Jdy hit Punch.
(b) The sn is shining. [vs. The sun is stting.]
2) non-content placement:
[2-28] (a) Id go, but theres no one to go wth.
(b) Yu get it.
3) equally good stress patterns for the same syntactic structure:
[2-29] (a) Just think of all the jnk we sold.
(b) Just think of all the junk we sld.
and 4) abnormal utterances with no normal placement counterpart:
[2-30] Even a lnguist has more sense than that.
Purely syntactic or rhythmic accounts handle such examples by treating them as
exceptional, but a functional analysis integrates them all as normal utterances,
disallowing the need to refer to contrastive stress.
For example (from [2-27]), qualities attributed to highly rhematic material tend
to appear rightward in English, but when it appears leftward, primary prominence
follows the wayward rheme without creating an exception to the general rule,
151
integrating the abnormal non-rightward cases. In addition, examples which used to
be dismissed as matters of non-content placement (such as [2-28]) are in BHs view
simply cases in which a function word expresses content in a prominent context,
even though in other situations it tends to express little content.
Because the notion of rheme is not bound by syntax, any word in an utterance
potentially has a measure of rhematicity without regard to its structural position, and
so in context, some words are close to one another in rhematic level, which means that
an utterance might have more than one equally good primary prominence placement
(as in [2-29]). This not only obviates the need to chose one prominence placement as
normal while the rest are identified as matters of contrast or narrow focus, but it also
makes it unnecessary to find a normal or non-shifted counterpart for any sentence
which is traditionally held to only have either a so-called contrastive form or one with
a shifted center of focus (e.g. [2-30]).
BH finishes off her dismantling of contrastive stress by commenting on
analyses which target those stress patterns which typically accompany parallel
structures, such as the following:
[2-31] (a) His name wasnt Bll, but Gorge.
(b) Prtt roasted a pg in the fireplace last year, and Whtney did it
with a gme hen.
Any contrast in meaning is encoded by the contrastive statements being parallel in
structure, and not by some specific stress pattern being contrastive. As long as that
clarification is understood, BH does not object to its being called contrastive stress.
152
Of course, Bolinger treats stress as a matter of accent potential, so his version
of contrastive stress is used to highlight contrasting syllables in otherwise equivalent
words, such as in [dport vs. xport vs. mport]. Bolinger calls this a resolution of
homonymic or near-homonymic conflict (1961a: 93). So, a phonetic definition can
be applied to this contrastive stress, but it is not predictable, and although its use can
imply contrastive accent, the reverse is not true.
Myth #5: Focus is Equated with Stress
Truth #5: Prominence doesnt do Focus (as it is ill-defined)
BHs says that studies equating focus and stress are limited simply by having
to depend upon poor or no definitions of focus. Chomsky (1971) assigns focus to
stress or a phrase containing the intonation center (p. 205), and then labels the
remainder as a presupposition, and does not talk about stress assignment (assuming
SPE-type rules). Jackendoff (1972) and Ladd (1980) go the other way around and
assign stress to focus. Ladds failing, according to BH, is simply that he does not
define focus well enough.
Ladd treats portrayals of focus by Chomsky (1971) and Bolinger (1972) as if
they were compatible, but BH not only denies that compatibility (Bolingers definition
is not syntactic in nature), she also indicates that they amount to no definition at all.
Other than that, Ladd only says that focus is: 1) an ill-defined concept, but intuitively
not hard to grasp (p. 108); and 2) an independent semantic phenomenon (p. 111).
That is a pretty vague set of descriptions compared to BHs work on theme and rheme.
Ladd does, however, distinguish broad and narrow focus (cf. Halliday, 1967):
153
[2-32] (a) Im leaving for Crete tomrrow. (narrow focus acc. to Ladd)
(b) Im leaving for Crte tomorrow. (broad focus acc. to Ladd)
Ladd identifies [2-32](a) as narrow focus, and says that there is a context for [2-32](b)
in which the whole sentence is in focus, such as in answer to the question that BH
proposes, namely What are you doing this summer? BH finds it hard to interpret
these differences in scope, suggesting that the answer lies in the identification of the
focus constituent, which is waiting for a better definition of focus.
Jackendoff (3.5, below) defines focus as information which the speaker
assumes is not shared with the listener. (I am not substituting S and L here because
they allow for cross-modal analogs which might not be included in Jackendoffs
original intent.) Focus is allowed assignment anywhere, and its location determines
the assignment of primary stress. The rest of the stresses fall prey to SPE. The problem
here is that this multiple focus is described as topic and comment, which means that a
point of focus can be either topic or comment, which means that only one kind of
focus, namely topic, determines the location at which stress is assigned. BH would
prefer that these terms be sorted out, and that a better definition be provided for focus
or topic before expecting them to support an equation with stress.
In summary, BH concludes that English stress is sensitive to context, where the
functional organization of the utterance determines the stress contours major beats.
Stress identifies the theme when there is one, and primary stress identifies the rheme
always. Care is taken to define theme and rheme well enough that BH does not fall
prey to the same problems as those which up to this time had tried to equate stress with
154
focus. In the absence of context, rhythmic principles are used to assign stress. The
limited success of syntactic analyses is attributed to the tendency for functional
organization to be aligned with syntactic structure. Finally, BH treats formerly
problematic data in a unified fashion, discarding appeals to contrastive stress.
3.5 Prominence DOES Outward Focus
The predominant complaint about linking prominence to focus is the lack of
any really good definition of focus. Another issue that needs to be settled is the scope
of focus, which is an issue that has been around since Walker distinguished two kinds
of emphasis; namely, emphasis of passion, and emphasis of sense, which refer to
sentential and word focus, respectively (1774: 18f).
Jackendoff (1972) is a good place to start, as he clearly explains the interaction
of prominence, focus, presupposition, and assertion for those sentences which have
one focus, and then likewise for two foci. The problem is that the definition of focus
gets all tangled up when the two accounts are tied together, as BH complained about.
Focus and presupposition are first defined in terms of the sentence information, where
S assumes that L shares presuppositional information, but not focal information. Given
a routine sentence contour, wherever the main stress and highest pitch are located, that
is where the focal word is found. A focus marker can be attached to any node in the
surface structure, and prominence falls on the syllable of the (non-function) word in
that constituent which would normally be assigned SPE primary stress.
A sentences semantic representation is derived in three steps: 1) identify the
Focus and a function Presupp
S
(x); 2) construct the presuppositional set xPresupp
S
(x),
155
and 3), form the actual presupposition and assertion (p. 245f). The lambda operator
simply means the property (or class) of x which are such that ...x..., as opposed to
the one individual that... (Carnap, 1956: 3). In sum:
Figure 2-6: Focus, Presupposition, and Assertion
The Focus simply consists of that semantic material associated with surface structure
nodes dominated by the marker F, and the predicate Presupp
S
(x) is made by replacing
this Focus with an appropriate semantic variable x in [the semantic representation]
(p. 245). The replacing information and the Focus overlap in semantic content only at
the variable, and differ elsewhere (e.g. cactus and grass can overlap at the variable
green or plant, without having to overlap at barefoot)
Appropriateness is defined in terms of a coherent class of possible contrasts,
meaning that the information represented by the variable could have been used in
place of the focus in context, having the same functional semantic (but not necessarily
syntactic) form. Jackendoff notes that:
{
}
is a coherent set in the present
discourse is well-defined
is amenable to discussion
is under discussion
Focus xPresupp
S
(x)
Presupposition
Assertion
156
The problem of defining appropriateness is amenable to linguistic
analysis only to a limited extent. Beyond a rather quickly attained
point, however, it becomes clear that the solution deeply involves
conceptual structure and knowledge of the world, which we are
(arbitrarily) not investigating here. Hence we will be relying heavily on
unanalyzed intuition in certain aspects of the discussion. (p. 242f)
Such conceptual structure and knowledge of the world is investigated in this analysis.
The presuppositional set, xPresupp
S
(x), is the set of values which, when
substituted for x in Presupp
S
(x), yield a true proposition (p. 245). This set is used to
construct the actual presupposition and the assertion, as in the figure above. As far as I
can tell, this is the same formal object that Rooth will shortly call a p-set.
Words such as even, just, and only (Horn, 1969) are ASSOCIATED WITH FOCUS. If
the focus falls anywhere within a words individually defined RANGE, then that word
affects the assertion, and not the presupposition. Evens range is as follows: If even is
dominated by a node X, the range of even includes X and all nodes dominated by X to
the right of even, plus the subject if X is an S (p. 251). The range of not is the same as
for even, but it is only optionally associated with focus. When not pursues this option,
the assertion is negated, and the in the figure above becomes . Another semantic
element that associates with focus is the Yes-No question [which]... obviously changes
the form of the assertion (p. 257) from a declaration to a yes-no question.
Jackendoff analyzes sentences with two foci, where the assertion has one pitch
contour per focus. Bolingers B accent (high pitch, abrupt drop, rising end) marks the
independent variable or topic, and the A accent (high pitch, abrupt drop, falling end)
marks the dependent variable or comment. When there is only one focus, it takes the B
157
accent, and so affirmation-negation is chosen to serve as the dependent variable, as
allowed by association with focus.
This is where BH has problems with Jackendoff. At first he says, we observe
that emphatic stress occurring anywhere in a sentence attracts the focus (p. 234), but
then he says, In the theory we have been discussing, emphatic stress is assigned
according to placement of the focus marker F (p. 259). The first of these definitions
might have been intended in context to be an informal description of the behavior of
the stress within a specific set of examples, rather than a statement designed to actually
describe the direction of causation between stress and focus.
The real confusion comes in when focus as unshared information gets tangled
up with topic and comment and (in)dependent variables in discourse. At first, focus is
defined in terms of the information which the speaker assumes not to be shared with
the hearer, but now both topic (shared) and comment (unshared) are types of focus,
and they both draw prominence. The way things are set up, the freely chosen topic
demands pitch accent B, which defines an independent variable. The comment is
constrained to saying stuff about the topic, and it takes pitch accent A, which defines a
dependent variable chosen in order to make the sentence true. A comment can be an
entire focus, including a presupposition if there is one, and the second variable.
This notion of free choice leads to convolution. Jackendoff sets up examples
where the topic of a second sentence is freely chosen from any material contained in
the first, but the comment is restricted to anything in the vast presuppositional set. So,
the topic is freely chosen from a very limited set, but the comment is constrained to
158
being chosen from a virtually unlimited set. The freely chosen topic is restricted to
shared information, and the comment, though not free, is unrestricted in that it is
unshared. I think that BH is right in suggesting that the definitions could be tidied up a
bit, as in the following research.
The next example (derived in part from Krifka, 1992) shows how Rooths
domain selection theory (1985) represents an expression (a) in terms of an abstract,
stressless MEANING (b); a set of ALTERNATIVE MEANINGS or p-set (c) as generated by
the expression when in focus; and a MEANING POSTULATE (d) for the FOCUSSING
OPERATOR (only). The formal representation for the expression (e) is created by
substituting the meaning and the p-set for the two arguments defined in the meaning
postulate. The whole formal representation has a paraphrase (f):
[2-33] (a) John introduced Bill to [SUE]
F
(b) M = introduce(j,s,b)
(c) A = p-set = {p|x[xALT(s) & p=introduced(j,x,b)]}
(d) ONLY(M,A) iff true(M) & p[p & true(p) p=M]
(e) ONLY(introduce(j,s,b), {p|x[xALT(s) &
p=introduced(j,x,b)]}) iff true(introduce(j,s,b)) &
p[p{p|x[xALT(s) & p=introduced(j,x,b)]} & true(p)
p=introduce(j,s,b)]
(f) Out of all the people John might have introduced Bill to, there
was one and only one person he actually introduced Bill to: Sue.
Notice that Rooth portrays an expressions usual meaning (b) as one without a focus.
Rooths alternative meaning encodes information about the focus. This encoding is
explicit (but indirect for entities), in this case a set of alternatives to SUE, which in a
lack of other context might be the set of all individuals capable of being introduced to
159
someone by someone else. The meaning postulate (d) for the focussing operator
ONLY states that the only true element of the set of alternatives A to the meaning M is
M itself. Applying this meaning postulate (d) to the alternative (c) and usual meaning
(b) of the expression (a) renders the formula given in (e), which has the paraphrase in
(f). Rooths work is characterized by von Stechow (1989) as alternative semantics.
This is the same example with the same focal word, but a different focal scope:
[2-34] (a) John [introduced Bill to SUE]
F
(b) M = introduce(j,s,b)
(c) A = p-set = {p|P[PALT(x.introduced(x,s,b)) & p=P(j)]}
(d) ONLY(M,A) iff true(M) & p[p & true(p) p=M]
(e) ONLY(introduce(j,s,b), {p|P[PALT(x.introduced(x,s,b))
& p=P(j)]}) iff true(introduce(j,s,b)) &
p[p{p|x[xALT(s) & p=introduced(j,x,b)]} & true(p)
p=introduce(j,s,b)]
(f) Out of all the things that John might have done, there was only
one thing that he actually did: introduce Bill to Sue.
The alternatives here are not to [SUE], but rather to [introduced Bill to SUE]: same
focal word; different focal scope. The portrayal of prominence such that: [semantic
scope phonological scope] was already floating around, but only the phonological
part of the inequality had been addressed. For example, BH took care of such cases by
resorting to rhythmic defaults. The formal difference presented here takes on the
semantic half, suggesting that word scope is not manipulated in the same way as a
broader focus of attention, in part because the greater breadth can come to encompass
operators rather than just a single argument. As will be shown in more detail below,
this significantly changes the way that the prominent material in the foreground is
understood to match up with the background.
160
Krifka (1992) moves on from Rooth to develop a framework for representing
FOCUSSENSITIVE QUANTIFICATION. Krifka refers to his system as DYNAMIC because
rather than simply identifying the focus of an isolated expression, it is used to track
focal elements and their alternatives across successive stages of discourse, including
any anaphoric binding with respect to those focal elements.
Krifka begins by analyzing instances of quantification in terms of a function of
the form QUANTIFIER(RESTRICTOR, MATRIX), identifying as prototypical the following
example with a relative quantifier:
[2-35] (a) Most frogs croaked
(b) MOST({x|frog(x)})({x|croak(x)}),
with MOST = XY[#(XY) > #(X)]
This next expression is given as an example of quantification over situations s as well
as objects/entities x. It will be important that Rooth (1985) only quantifies over
situations:
[2-36] (a) Mostly / Most of the time, if a frog is happy, it croaks.
(b) MOST({<s, x>|frog(x) & happy(x,s)})({<s, x>|croak(x,s)})
Utterances are turned into formulaic equivalents by motivating SEMANTIC PARTITIONS
(Diesing, 1990), where clues to the proper partitioning of an utterance are found in
sources such as phrase structure, syntactic markers, and morphological markers. In
this particular study, focus is the partition source of greatest importance, where focus
is typically marked by sentence accent in languages like English (p. 216).
161
When focus is used to form semantic partitions, expressions such as [2-37](a)
or [2-38](a) divide into their respective restrictors ([2-37](b); [2-38](b)) and their
matrices ([2-37](c); [2-38](c)):
[2-37] (a) [In St. Petersburg], OFFICERS
F
always escorted ballerinas.
(b) EVERY({s|xy[escorted(x,y,s) & ballerina(y)]})
(c) ({s|xy[officer(x) & escorted(x,y,s) & ballerina(y)]})
[2-38] (a) [In St. Petersburg], Officers always escorted BALLERINAS
F
.
(b) EVERY({s|xy[officer(x) & escorted(x,y,s)]})
(c) ({s|xy[officer(x) & escorted(x,y,s) & ballerina(y)]})
Lines (b) and (c) go together in each case to make a single assertion. I would like to
ensure the clarity of Krifkas statement that expressions that are in focus are mapped
to the matrix (p. 216). Notice that the single words in focus in either [2-37] or [2-38]
do not give rise to their respective matrices alone. These matrices require information
found in the whole utterance, but it is not the whole utterance that receives the actual
focal prominence. So, when Krifka refers to expressions here, he is talking about an
entire utterance which has some element in it which is in focus, and not just the
prominent word alone. In that sense, matrix partitioning actually treats the scope of
focus as irrelevant, with the result that the utterances above have identical matrices
despite different focal scopes.
Krifka intends for this system to encode the focal constituent (of whatever
scope) onto the matrix along with everything else, which is entirely reasonable and
consistent if only non-focal material is mapped onto the restrictor, which it is. The
whole utterance is mapped onto the matrix, and the scope of the focus is only apparent
162
as the material which appeared in the matrix but not the restrictor. Just in case it is not
clear, what matters is the relationship between the restrictor and the matrix. While only
the focus OFFICERS
F
is missing from the restrictor of [2-37] when compared to its
matrix, only the focus BALLERINAS
F
is missing from the restrictor of [2-38]
likewise. This is what portrays the focal material as the variable for which other
semantically appropriate material might have been substituted.
There is a reliance in Krifka upon syntactic constituency to determine the
range potential of the focal scope:
...focus is represented by a feature F that applies to syntactic
constituents and may be spelled out by sentence accent on certain
syllables of certain words of the constituent in focus. The constituent in
focus may be associated with a focusing operator such as only that c-
commands its focus. (p. 216)
Some instances of focus are well known not to obey syntactic constituent boundaries,
such as Newmans list of fruit, or parallel structures. After acknowledging them,
Chomsky and Halle suggested that resolving this behavior was complex and beyond
the scope of SPE. The solution is to treat such examples as multiple individual
instances of focus, some of which work in a systematic, coordinated fashion, and
others of which are simple coincidence. I analyze such linked volitional prominence in
chapter 5.
Krifka complains that Rooth fails to represent Most of the time, Mary took
[JOHN]
F
to the movies, as possibly meaning that Most of the time, Mary took only
John to the movies and no one else. Quantification over situations instead of entities
163
leaves [JOHN] inaccessible for portrayal as unique. Krifka rejects solutions involving
the nesting of focal operators such as MOST and ONLY on the basis of even further
meanings that they fail to capture. In addition, Krifka notes that variables in the matrix
and restrictor remain unbound in expressions like:
[2-39] (a) Most of the time, a frog that sees a fly tries to CATCH it.
(b) MOST({s|x,y[frog(x) & fly(y) & see(x,y,s)]})({s|try-to-
catch(x,y,s)]})
Alternative semantics fails to capture anaphoric reference. Finally, a sentence which
has no situation argument over which to quantify presents problems for Rooth, as in
Most of the time, a three-colored cat is INFERTILE (p. 219, ex. 16).
Krifka synthesizes four semantic representation systems into his dynamic
framework, which operates as follows: 1) focus is used to partition an expression into
background and focus elements; 2) which are fed as direct arguments to focal
operators; 3) where quantification can be expressed over cases, and not just situations;
and 4) dynamic semantics allows for the projection of information about background
and focus, such as that involved in anaphoric binding, through obstacles such as
complex foci, multiple foci, and the stages of discourse. Prominence then becomes a
matter of helping to align S and L not just over the immediate context, but across the
stages of discourse as well, including the tracking of anaphor.
Now we come to Lambrechts take on focus (1994), who uses ALLOSENTENCES
(p. 6; after Dane, 1966) to illustrate the differences in three types of sentence focus
structure. The following examples are adapted from his work (p. 223):
164
[2-40] (a) My car broke dwn. (What happened to your car?)
(b) My cr broke down. (What happened?)
(c) My cr broke down. (I heard your motorcycle broke down?)
[2-40](a) is an example of predicate-focus, where the car (topic) is already a part of the
discourse context, the proposition is the comment about that topic, and the prominent
word is the focus of that proposition. Example (b) displays sentence-focus, where a
whole new set of referents is introduced into the discourse, so: 1) there is no topic
(except the speech event and its participants by default); 2) there is little formal
pragmatic presupposition (something happened); and 3) the assertion coincides with
the focus. Neither one of these strikes me as representing an analog of volitional
prominence.
Then there is the argument-focus in example (c), which has the presupposition
that Ss x broke down (p. 228), and the assertion that x (the presuppositions focus)
was the car. Lambrecht does not mention the additional intensity of car in (c), but
he does specifically mention that, The assertion made by uttering [(c)] is therefore not
merely the identification of X with the speakers car but also the correction of a
mistaken belief on the part of the addressee (p. 229). This sounds like substitution to
me. Finally, the prominence pattern differences between argument-focus and
predicate-focus structures evaporate when predicates are made intensely prominent,
because such predicates approach being treated as if they were just as referential as
arguments.
165
While Lambrecht pays attention to routine patterns of prominence, I am mostly
interested in their volitional forms, on top of which, Lambrecht is concerned with
prosody only inasmuch as it serves to mark contrasts in the INFORMATION STRUCTURE of
sentences, and specifically not with its use for other kinds of semantic or pragmatic
purposes such as SPEECH-ACT DISTINCTIONS or SPEAKERS ATTITUDES (p. 238f). So,
my concern is not routine, and his is neither volitional nor word-internal; nevertheless,
there is overlap. Argument-focus seems to be a type of outward focus or revelation
whose volitional prominence simply did not merit Lambrechts attention; similarly,
predicate-focus is used for revelation when its prominence is volitional.
To top this off, I really had not expected Lambrecht to say anything at all about
the iconic relationship between prominence and communication, much less something
specifically supportive, but he does:
One of the tasks in the description of sentence prosody must therefore
be to show how prosodic prominence as an iconic information signal is
converted into informational meaning by being mapped onto
grammatical structure, which is an essentially non-iconic system for the
expression of meaning. (p. 242f)
Lambrecht reveals prominence functions at their most advanced state of adaptation
away from their origins in sensation, portraying the relationship between grammatical,
normal prominence and communication as only PARTIALLY ICONIC (p. 242);
however, the functions of volitional prominence studied here reflect an earlier stage of
adaptation, and their iconicity is greater because they are closer to their completely
iconic origins in sensation.
166
3.6 Prominence DOES Inward Focus
These previous analyses of focus deal primarily with shifts in the placement of
prominence that select one out of a set of alternative values to the variable or focal
element. For example, cactus would be interpreted as cactus in particular versus all
alternative values of the variable plant, such as grass or palm tree, and so on. Shifts
in meaning also accompany changes in the form of prominence, but they generate a
different kind of alternative to the focal element. In such a case, cactus might be
interpreted to mean one of those dangerously spiny kinds of cactus in particular,
rather than some sort of succulent which is sometimes called a cactus, even though it
really isnt one.
Alternatives of the first kind present the focal element (cactus) as the specially
chosen member of a set which contains an indefinitely huge number of things (plants)
which are as similar (succulents) or as different (orchid) from that element as
Jackendoffs semantic appropriateness will allow. Alternatives of the second kind
present the focal element as a select member of a set (cactus) which contains a very
small number of things which are so absolutely similar to it (cactus-like things) that
they would usually be ignored by the conversants (succulents are included as cacti),
were not prominence deliberately brought to bear on the differences that distinguish
that element from the others (only spiny cacti). Where external focus compares the
typical meaning of a word to that of others in its context, internal focus draws
attention to shades of the words meaning which would normally be overlooked. This
inward focus is ELABORATION.
167
Just as the formal semantics version of revelation is outward focus, elaboration
or inward focus finds analogs among Lasersohns (ms.) pragmatic slack, where
slack regulators tighten the routine tolerance for falsehood around the denotation of
an utterance (p. 8). To begin with, Lasersohn uses the example, Mary arrived at three
oclock (p. 3) to point out that it usually doesnt matter if Mary actually arrived a
minute (or so) earlier or later than 3:00. Suggesting that Mary arrived at 3:00 when it
was actually 3:02 would be saying something which is literally false, but close
enough to the truth for practical purposes (p. 8). 2:59 and 3:01 are usually so similar
to 3:00 that people ignore the difference, allowing for what Lasersohn calls pragmatic
slack in meaning, which is a truth-theoretical counterpart to what I have been studying
as conceptual or ideational INTOLERANCE.
Lasersohn portrays the arrangement of the primary alternatives to the utterance
as arrayed in a circle around its denotation in a pragmatic halo (p. 15). The
pragmatic context determines how many of these alternatives are associated with the
event without any of them exceeding the limit of irrelevant detail. it might be expected
that the slack allowed in the mention of a particular minute (3:03) might include at
least 20 or 30 one-second increments to either side (plus an increasingly large number
of increments with increasingly smaller size), and that a minute which coincides with
the top of the hour (such as 3:00) might normally be used as an approximation of that
hour so often that a few minutes to either side might well be ignored.
Words such as exactly and precisely function as slack regulators (p. 8),
and the process of tightening this slack is an elimination from the halo of those
168
elements ordered furthest from the core formed by the expressions truth-theoretic
denotation (p. 17). This reminds me of radial categorization (Lakoff, 1987: 91), and
while Lasersohn would not likely take exception to the radial part, I would not want
to project the ideational associations of categorization onto this truth-theoretical
representation. (Although Lasersohn specifically does not promote halos as mental
representations, some sort of mental principles are involved in determining what goes
into the halo and what does not.) Even at that, I am interested in the similar facilitation
of functions which is allowed by a radial construal of the alternatives, even if in
Lasersohn this spatial arrangement is not supposed to reflect mental structuring.
Finally, both of these analyses define the halo of a complex entity as a
combination of the alternatives to its individual components, but Lasersohn intends for
this to be taken as syntactic phrasal componentiality (NP-halo + VP-halo = S-halo),
and I prefer an interpretation in terms of the syntagmatic combination of semantic
units (cf. Langacker, 1987: 82).
In comparison to Lasersohns treatment, I interpret Mary arrived at 3:00 as
one activation pattern distributed through a web of potential semantic substructures,
where that web in its entirety is composed of all of the possible alternatives to each of
the utterances subunits over the course of its combination. The assignment of an
utterances strongest prominence, volitional or otherwise, defines what the primary set
of alternatives will be which surrounds that utterance, but the secondary shadows must
radiate at a greater remove, because that is part of the way that conversants navigate
their way through discourse. Somewhere around Mary arrives at 3:00 there lurk
169
alternatives to Mary as we usually know her (different moods, physical appearance),
and alternatives to arrived ( la Dolly Levi), and alternatives to the actual time (and
to at, for that matter), all of which occur within some pragmatically unimportant
margin of the definition in the utterance, and all of which are crossbred as individual
alternatives to the utterance as a whole.
Routine prominence indicates which of these radiations is the primary set of
alternatives, whose variations are being ignored, and volitional prominence brings
attention either to that or to some other set of alternatives such that this margin for
error is no longer overlooked. Revelation draws attention to a contextual alternative,
and elaboration signals a word-internal alternative. Now, a proper pragmatic halo only
surrounds a denotation which plays a role in determining truth and falsity, and so while
most of what I talk about as sets of meaning alternatives are not halos at all, halos still
represent a type of tolerance. I interpret the reduction of word-internal tolerance to be
a type of elaboration (and instantiation), which has been examined in great enough
detail by Langacker that I was easily able to identify this inward focus function of
prominence as elaboration (to begin with: 1987: 68; 1991a: 61; also described at the
beginning of chapter 3 in this dissertation).
3.7 Prominence DOES Focus, both Outward and Inward
Prominence is characterized so far as performing two discourse functions,
which until now have been called inward and outward focus. As this analysis
progresses, I will differentiate each of these functions internally, and in light of the
increasingly detailed information exposed about their behavior, they will be
170
rechristened elaboration and revelation. There are a few places in the literature where a
similar pairing of functions has been proposed, and there has even been some mention
of the possible components of these functions, but no analysis has ever pulled it all
together before.
The earliest similarity I found to revelation was Sweets (1890) negative
emphasis, which gets its name from the fact that once any given utterance is stripped
of information which is old in the context, and then also has its function words
removed, then whats left must be the logically prominent wordsthose which are
most indispensable for expressing the sense (p. 28). Sweet gives the example (in
phonetic transcription), I got wet (p. 28f), which when stripped of old information
(I), and then function words (got), would leave only wet designated as the
logically prominent word with negative emphasis. Note the implied equation drawn
between Sweets sense and the part of the meaning to which S and L should both be
paying the most attention. If, however, wet were used to mean very wet, then that
would be positive emphasis, or emphasis proper (p. 29), or elaboration, which sets
up the two functions as a pair.
Colemans (1914) division into prominence and intensity is the same thing:
14. We have not thus far defined the meaning of emphasis. It is
however desirable to give a definition before preceding further. The
first kind of emphasis may be defined as that manner of utterance
which marks any word or phrase of greater importance than its
neighbors. This kind of emphasis may be termed Prominence.
15. The other kind, which may be termed Intensity, may be defined
as that manner of utterance which imparts an added degree of intensity
to some part of the idea represented by a word. (p. 11)
171
So, while prominent black is opposed to dark blue, intense black is opposed to
regular old black. While Coleman finds this to be similar to Sweets Negative
Emphasis and Positive Emphasis (p. 29), respectively, he still finds Sweets
characterization of those categories difficult to follow, given a disagreement over
Sweets identification of contrastive emphasis as positive (p. 7n1).
A&W (1926) attach increased pitch and optional deaccenting in the surround
(Special Prominence) to the expression of contrast (revelation), and combinations of
increased breath force and pitch changes (Intensity) to prototypical or peripheral
meanings (p. 43f; elaboration). Admittedly, this sets up all emphatic sentences as
being based on unemphatic forms, but at least A&W do not throw out emphatic
utterances altogether, which some later analysts seem all too eager to do.
Jones also grazes the precision subtype of the elaboration function, which
displays a characteristic reduction of tolerance: The stressing of this, these, that
(demonstrative), those depends upon the amount of demonstrativeness it is desired to
suggest. Sometimes they are equivalent to little more than the definite article the, and
in such cases are unstressed (1918: 266; 2.968). In addition, he says that emphasis
can be used either for intensity or contrast, where a word like enormous would
come to mean particularly large (1950: 109; the power subtype of elaboration).
Then there are Newmans (1946) two types of expressive accent. One used a
peak shift and extra articulatory intensity for a contrast of references in a predicate
(revelation), and the other used extra force and quantity (size iconics) to augment the
meaning of the targeted word (elaboration).
172
Bolinger (1986) posits two classes of accent patterns, namely accents of power
(ch. 6) and accents of interest (ch. 7). An accent can be placed on an individual word to
highlight its interest, and there are patterns of pitch accent called PROFILES (ch. 8)
which throw portions of varied scope into prominence. When a profiles accents are
shifted to the outer edges of an utterance, the impact or power of the message as a
whole is made prominent. Accents of power affect meaning broadly, involving the
release of an effusion of feeling (p. 83). My functions of elaboration and revelation
are both subsets of his accents of interest, depending upon the reason a word is
interesting enough to highlight.
Taken together, this gives the same sort of strength to the definition of inward
and outward focus that BHs work was able to attribute to theme and rheme, which
amounts in great part to putting the functions through their paces. Inward focus is used
by S to let L know that a word is being used with an unexpected variation of its typical
meaning, and this warning helps to preserve the sense of the utterance, where making
sense refers to preserving the meaning unmolested from S to L. Outward focus is used
by S to keep Ls attention directed towards the information that preserves or corrects
the alignment between their perspectives on the discourse, helping to make sense by
tying newer information to older stuff which has already proven stable and sensible.
3.8 Prominence DOES Proportion
The phonological levels of prominence have been described so far as having
four rough divisions, namely: intense; primary; secondary; and then the significant
lack of prominence (zero/weak). Prominence has also been portrayed as identifying
173
areas of semantic attraction along a CD continuum, specifically: focus; rheme; theme;
and then the little or no prominence that is assigned to transitional elements.
The appearance of any given phonological component or set during a specific,
individual instance of prominence has been treated as unreliable, but Wells (1986) at
least establishes a likely distribution of these components across prominence as a
whole. In addition to this, Wells goes on to show that each of four typical sets of
components are strongly correlated to the expression of four distinct levels of outward
focus in an utterance. (For the sake of clarity, I am going to substitute the vocabulary
which has been promoted in this analysis for those terms used by Wells, such that his
use of subsidiary is replaced by secondary, and so on.)
Wells presented each of 30 test subjects with a transcription of 23 sentences
which had been removed from the context of an audiotaped conversation. While
listening to the tape on which the sentences were separated by short pauses, each
subject was to underline the part of each sentence which the speaker [was] focussing
on as particularly important (p. 54). Multiple foci were to be numbered for relative
order of importance. Wells found that the test subjects appealed in common to four
distinct levels of focus which are analogous to the outward focus, rheme, theme, and
transition levels of prominence from the previous chapter.
Each focus level is consistently correlated with a specific phonological level
(focus = intense = 1, rheme = primary = 2, theme = secondary = 3, zero = weak = 4),
each of which is characterized by a consistent set of phonetic characteristics, as laid
out in the table below. There are three main groups of such characteristics which are
174
significant in the identification of intensity levels: pitch; loudness maxima; and tempo
changes. Pitch takes the form of 1) tone movement on the point of focus (any pitch
movement in an otherwise still surround), or 2) a pitch maximum on the point of focus
(either the pitch peak or maximum movement of pitch). Loudness maxima include
either 3) the loudness peak, or 4) a decrescendo (a step downward in loudness when
leaving the point of focus). Tempo changes are 5) virtually any change in the rate of
speech at the point of focus, either faster (allegro) or slower (pause/drawl). Duration is
not measured.
These findings are summarized in the following table:
Table 2-4: Prominence Components Associated with Focus
Each of the four phonological levels of prominence can be consistently identified by a
specific combination of these five features. An intense phonological level of
prominence marks outward focus (an intense semantic level) with all five phonetic
components (1, 2, 3, 4, 5). Primary prominence on the rheme is less intense in that it is
only marked by tone movement and either decrescendo or tempo (1, (4 or 5)).
Pitch Loudness Maxima Tempo
Tone
Movement
Maxima
(peak, movement)
peak decrescendo allegro or
pause/drawl
L
e
v
e
l
1 1 2 3 4 5
2 1 4 (or 5) 5 (or 4)
3 1 4
4 1 (or 4) 4 (or 1)
175
Secondary prominence for the theme is somewhat less intense yet by being marked
only by tone movement and decrescendo (1, 4). Weak prominence, when it is used at
all, only accompanies transitional elements with the presence of either tone movement
or decrescendo (1 or 4).
This supports the consistent, proportional variation in the intensity of the
phonetic characteristics of prominence from intense all the way down to the weak
levels, and there is also support for the intensity of the level of focus or interest marked
by prominence varying along a scale as well, from outward focus to rheme to theme to
weak. There is a direct iconic proportion holding between the phonological and
semantic levels of prominence intensity. I would like to introduce a symbol for this
relation, to be used as follows: focus ^ intense; rheme ^ primary; and so forth.
In summary to this point, Wells shows that prominence has phonetic feature
sets associated with semantic distinctions in what he calls focus levels. Bardovi-Harlig
was shown earlier to demonstrate that this association which determines primary and
lower prominence placement can be described in terms of two synthesized discourse
functional terms, namely theme, which gets secondary stress, and rheme, which gets
primary stress. Also arising from the comparison of prominence analyses is the
suggestion that there is a direct proportion holding between the phonetic and semantic
properties of any given level of prominence intensity. Wells supports such a contention
as portrayed in terms of focus levels, and goes one step farther, showing a few cases
which suggest that rhematic prominence might have been a continuous adaptation of
the meaning of volitional prominence. Together, these studies predict that just as the
176
phonetic characteristics of prominence consistently disperse from intense through the
lower levels, so might rheme and theme be weaker forms of the communication
function appealed to by focus prominence.
4 Prominence MAKES Sense
Physiologically, an increase in prominence is accompanied by a greater
expenditure of energy, which results in greater efforts such as the recruitment of
additional motor units for the movement of the lungs and articulators, with consequent
increases in subglottal pressure, and changes in the baseline values of features such as
formant and fundamental frequency, duration, and vowel quality. Language users have
this price so well internalized that hearers will attribute greater prominence to sounds
which require the greatest gestural effort to make on the part of the user (distal
evaluation), and as I keep repeating, Im sure that there are not only analogs in other
modalities, but even cross-modal universals. This realization is buried deep in the
metaphorical structure of our language, which equates greater size and anger and heat
and sound all together with a greater energetic power, as well as focus of attention with
greater energy put towards an intense precision of movement.
Similarly, an otherwise undisturbed, continuous expenditure of energy will fall
into rhythmic patterns which require no additional input from its generator, and any
deviation from that low-energy physical default state (other than to lapse into inaction)
requires the expenditure of greater energy. The greater the shift of an event from its
normal timing, the greater the effort required. Cognitive routines, conventional
177
meanings, and undirected attention are all representatives of low-energy mental
default states which cycle just this side of sleep: the lower the activity, the lower the
processing, and the lower the energy expended. Inward and outward focus are used to
convey novel meanings in the service of preserving the sense of an utterance within a
discourse, or the sense of a words meaning within an utterance. The greater the shift
in a cognitive event from its routine timing, or the greater the distance from the central
meaning of a word, the greater the mental effort required.
For each of these types of prominence taken individually, linguistic or
otherwise, greater prominence takes greater effort. The rest of this analysis is devoted
to providing the linguistic functions of prominence with enough structure that all of
these types of effort can be demonstrated to be in direct iconic proportion to one
another.
178
CHAPTER 3
Elaboration
If a language permits a contrast in form to survive, it ought to be for a
purpose.
Dwight Bolinger, Meaning and Form
The analysis of volitional prominence for elaboration requires a clear portrayal
of what people do when they actually speak or sign about something, namely isolating
portions of their shared reality as: THINGS (food, nothings, Marilyn); TEMPORAL
RELATIONS (processes such as defend and graduate); or ATEMPORAL RELATIONS
(fairly, green, in). To begin with, a potential S brings to mind material about
which it wants to converse, thus making MENTAL CONTACT with a TARGET, after which S
describes that target so that L can contact similar material. S narrows down the range
of reality through which L has to search by defining a schematic TYPE SPECIFICATION
(Langacker, 1991: 53; TS). The TS describes properties of the target, identifying a set
which encompasses all of the entities which belong to the same class as the target,
without actually singling out either the target itself or any other particular entity in that
class. The process of describing this TS in greater detail, thereby helping L to locate
the target more easily, is one kind of ELABORATION (Langacker 1987: 68; 1991: 61). For
example, the type [PLANT] is less elaborate than the type [EDIBLE PLANT], which is
less elaborate than the type [CARROT], and so further elaboration narrows down the
material through which L has to search.
179
As a refinement of this tactic, S can use variations in prominence to ensure an
even closer alignment of Ls conceptual construals with its own. This is simply another
way of saying that S uses volitional prominence to make sure that L really understands
what S means. In this case, prominence works to correct or avoid those mismatches
between Ss and Ls conceptual construals that are provoked by a WORD-INTERNAL
difference in meaning. S presents L with a prominent phonological form, drawing Ls
attention to the fact that the full, fuzzy spectrum of meaning conventionally evoked by
that form is not what S had in mind. This internal meaning change is a special form of
elaboration, where the full set of a forms potential range of meanings is treated as if it
were a TS, and the individual, subtle variations of meaning within that set become
member instances of that type.
When S applies volitional prominence to a phonological form, L interprets that
effort in terms of a directly proportional ICONIC (physical force ^ nonphysical force)
change in its conventionally associated semantic representation (cf. ch. 1, 3.1; ch. 2,
1.3, 3.4). There are two interpretations of elaboration, divided according to whether
this increase is taken as signifying greater brute force or exceptional finesse. With the
brute force approach, L takes the specification of additional effort as promoting one of
the meaning variations which exploits the application of greater POWER in its construal
(1), such as larger things and intensified relations, including more forceful processes.
It is obvious that PRECISION (2) influences literally physical arrangements; just
go to any grocery store and try to find a 32-ounce plastic jar of Brand X reduced-fat,
no-preservatives-added, no-sugar-added, chunky-style peanut butter. In fact, your
180
thoughts about doing so reflect the existence of cognitive processes which are at least
compatible with the articulation of figurative spatial organization, even in the absence
of visualizing (or otherwise sensualizing) any spatial structures inside of your head.
The figurative space representative of such thoughts is familiarly portrayed in terms of
delimiters such as boundary placement and tolerance, orientation, or measures of
quantity, where L interprets the additional effort not just as raw power, but as finer
precision. The analysis of prominence as it represents exceptional finesse draws upon
the common attribution of a figurative spatial structure to the conception of relative
quantifiers, demonstratives, types and their instances, and some adverbs. The efficacy
with which this heuristic explains their linguistic behavior is taken as reflecting the
existence of cognitive processes which are at least compatible with the representation
of spatial structure, even if those processes are not defined as literally spatial in an
actual language user.
When this finer precision amounts to a closer alignment of boundaries, an
increasingly exact identification of a location, or a more accurate measurement of a
quantity, then precision is used with narrower tolerance (2.1). When precision draws
upon a spatial portrayal to characterize instances within a type as things located in
parcelled space, then L understands this precision to accurately isolate a prominent
instance within its type, so L interprets prominence as identifying a particular POSITION
within that type as special (2.2). Context determines whether an instance is pushed
right to the center of its type as a PROPER member, or forced outward toward the types
PERIPHERY.
181
In all of these cases, L interprets the increase in the phonological intensity of
the physical form as being in direct iconic proportion to the increase in the semantic
intensity, as represented by the conception of increased power or precision in the
forms associated meaning.
1 Power
Prominence for the attribution of greater power elaborates a words meaning
by emphasizing size or perceived imminence. Power can be inherent in an entity, or it
can be added adventitiously in accord with Ss emotional charge. If this potential
potency is built in, then it should essentially be universal across personal construals, as
well as virtually context independent (plunge, always, hero, huge). When not in
contextual competition with each other, such entities always elicit at least a respectful
rhematic prominence, and they are naturally prone to being assigned intense potency
by S. This power can be physical or nonphysical (mental, spiritual, emotional), or can
derive from indefinitely large quantity (big, lots), and it encompasses things which are
physically large, emotionally overwhelming (abject terror), valuable (gold), or deeply
personal (intimacy). Of course, the vast majority of entities have no great power, but
any of them can be treated as powerful when S makes them personally prominent in
context. This power reflects Ss experience with the entity in general, or Ss emotional
charge or investment in that entity in a specific context. Elaboration for internal and
external power can overlap sometimes, in which case an inherently or adventitiously
powerful entity is attributed even greater power by Ss emotional charge.
182
Capitalizing on the potential power inherent in most adjectives (as continua or
relations with open upper ends) is particularly popular:
[3-1] ...Cathy had stayed all winter too.... That was an awfully long
two weeks.
P15 0270
[3-2] [The dishes] looked so formidable, however, so demanding, that
I found myself staring at them in dismay and starting to
woolgather again....
R02 0350
Other examples include: terrible step
G49 0830
; real money
P10 0630
; big fight
K15 1220
;
swell party
P03 1540
; and, fast, fast muscle growth
E01 0200
. Prominence does not add
power that was not here before, but forces the relation toward the extreme end of the
scale. Some modifiers, though open-ended, do not describe properties which tend to be
associated with power, and so when made prominent, they tend to be treated with
finesse rather than brute force, as if precise or paragonal (Ben is so gentle).
Prominent adverbs elaborate with precision (2) in a context where they
approach a limit to within a narrow degree of tolerance, but they elaborate with power
when taken to mean simply that the limit is being approached (rapidly, inevitably), or
that it has in fact been reached or exceeded:
[3-3] It had gone like clockwork. Almost too smoothly....
L24 1170
[3-4] Families are very interesting.
N19 0970
[3-5] Of course hes in.
L14 0260
Other examples are too much....
N17 1290
, and the pair awfully good
R07 1130
and
awfully kind
R07 1140
by two successive speakers. As relations, adjectives and
adverbs dont have types to specify as such, and so they are not prone to elaboration in
183
the same way that an instance of a noun or verb type might be; however, these same
principles can be applied analogously in the sense that there are locations along the
relations continuum used to identify instances of greater or lesser power.
Just as with modifiers, some verbs have a greater tie to physicality than others,
but unlike the modifiers, a lack of such a tie does not as strongly suggest that
prominence will be used for precision rather than power. Verbs draw upon the power
of prominence so that the given process can fulfill a potential for expending great
energy, but some do so less generically than others:
[3-6] I wish so much someone loved me.
P14 0290
[3-7] I could go with him. He knows me as your niece, which, of
course, I am. But I am a slave! You own me. Its your decision,
said Juanita, holding her face very still, trying to contain the
bitterness of her voice as she enunciated her words too
distinctly.
K15 1710
[3-8] I drove him away.
P03 0240
[3-9] You could try. And if I ever hear you say Mist Laban again
Ill scream.
P03 0660
Other examples are: [prohibits]
J46 0900
; [stared]
P19 1400
;

and [come on]
N29 0710
.
The distinction that I am trying to draw here is that there is an inherent physicality to a
modifier like long or fast that is lacking in a verb like own. The power of own
comes not from size or speed as such, but from the potential potency of the process,
which determines how long the process will last, how many entities it will affect, how
strong that effect will be in its consequences, and so forth. As there arent that many
examples of prominent content verbs in the Brown corpus, Im going to illustrate this
point more clearly by drawing upon a few examples from the Other corpus.
184
So, some content verbs have greater inherent physicality than others (punch
vs. ponder), but the effect of additional power on physical verbs like bust
3.303
or
sails
8.-11
is just as clear in the 17 instances of the non-physical verb know:
[3-10] But though the chapter is ending, you know theyre coming
back next month, because they never tell you which one is
Victoria, and whats the big secret.
8.314
[3-11] A typical example of the occasional resistance mustered by
intuitive thinking against the clear conclusions of analytical
thinking is D.H. Lawrences opinion of the nature of the moon:
Its no use telling me its a dead rock in the sky! I know its
not.
17.192
[3-12] Cmon, she knows I love her.
8.302
[3-13] Crooks said gently, Maybe you can see now. You got George.
You know hes goin to come back....
6.127
The meaning comes to reflect a daunting depth of intuition, and a formidable strength
of conviction. Other examples are: [care]
5.202
; [live]
11.51
; [need (not vs. want)]
37.359
; and [needed (not vs. want)]
37.359
. There are also some cases in which the
repetition of a verb is used to enhance this power:
[3-14] REINER: Im a little queasy about this, telling tales about
Presidents and Presidents wives.
2000: Theyre all a little power-crazy, right? And they love to
do it. Lets face it. They love it. They love it. They LOVE it.
19.89
[3-15] Her mind echoed with Stephens voice, and, try as she might,
she could unlock no secret meanings from his words. She did
trypressing deep into those dim memories of her infancy for
words and phrases.
24.83
[3-16] I mean the whole thing with prostitutes and menI just dont
get it....
I just dont get it.
8.174
185
[3-17] The Boys would blink and stiffen, and then they answered him.
That was the moment when Frank knew he had control. They
answered him.
37 100
This repetition supports the contention that the intensity of the phonological form is in
direct iconic proportion to changes attributed to the intensity in the semantic structure,
particularly in the example which goes from love to love to LOVE. There was
also a repeated adjective earlier (fast, fast muscle growth
E01 0200
), but there is no
evidence yet which clearly suggests that nouns behave like this.
When it comes to intensely prominent nouns, it is best to begin with the regular
old nouns which are attributed greater size or imminence of threat:
[3-18] This brings us squarely to the problem of power, and the uses a
nation makes of power.
F23 0490
[3-19] In Inside Africa, John Gunther describes one of these [huge
cartels], the Societe Generale, as the kind of colossus that
might be envisaged if, let us say, the House of Morgan,
Anaconda Copper, the Mutual Life Insurance Company of New
York, the Pennsylvania Railroad, and various companies
producing agricultural products were lumped together, with the
United States government as a heavy partner.
A41 0450
[3-18] resembles the use of prominence for technical terms, and it would have been
disallowed if it had only mechanically set off a section of the passage, but in context
the intent was to make power come across as ominous. In [3-19], prominence
associates government with threateningly inhuman size and power, not only making
its actions seem ponderous and destructive, but isolating its impersonal nature, which
makes it seem indifferent to the populace, perhaps even malevolently oppressive, all of
which add up to portray the government as a shadowy, invulnerable Enforcer.
186
These examples easily resolve the shrimp problem that puzzles Schmerling
(1976: 43), undermining Ladds (1980) treatment of Schmerling in general, as well as
their mutual mistreatment of Bolinger. Suffice it to say that in this analysis, it seems
straightforward to suggest that the sheer size or volume (not necessarily physical) of
all these shrimp is enough to explain the prominence used on, Ive got all these
shrimp to {clean, sort, peel, devein, package, etc.}, and that predictability is not the
issue. One of the points here is that this same prominence pattern can not only be
applied to a lot of shrimp, but also to just one shrimp (Ive got this shrimp to...), just
so long as that shrimp (or other entity) is construed as having a more elaborate TS than
normal, such as if the particular shrimp were repulsive, vast, or more anything. The
shrimp doesnt have to be new or not predictable, just disturbingly conspicuous.
In English, classifiers are those nouns which often appear in constructions
and perform functions similar to those of the classifiers found in other languages, such
as lot (of), bunch (of), and mess (of). They can be elaborated for power because
their meanings are related to quantity and size:
[3-20] ...the hebephrenic patient... shows... laughter laughter which
now makes one feel scorned or hated, which now makes one
feel like weeping, or which now gives one a glimpse of the
bleak and empty expanse of man's despair; and which, more
often than all these, conveys a welter of feelings which could in
no way be conveyed by any number of words, words which are
so unlike this welter in being formed and discrete from one
another.
J31 1280
[3-21] Maybe because they have had virtually no radioactive exposure
and dont have any Rs stored up, they could take a lot without
harm.
M04 0840
187
The additional prominence increases the merely large diversity of welter and
quantity of lot to their greater welter and lot capacities. This behavior is
reflected in examples from the Other corpus as well:
[3-22] Try the game a few times, and you think youre doing really
well to score 800 points. Then you find out that lots of guys
have broken 100,000, and someone in Japan hit 300,000.
11.148
The context links you with guys, so prominence on the quantity lots emphasizes
the proportion against the implied singularity of you. The tie between players and
scores points out that 800 is pretty lousy compared to one or more 100,000 scores, and
prominence indirectly makes it plain just how many of those 100,000 scores there are
(possibly in proportion to the number of players and scores as a whole). So, where one
guy equals one score, you might still have saved face had there been a few aberrant
100,000 scores, if scores were normally in the 500s, but there are so many scores of
100,000 that the 800 points looks particularly pitiful. Alternately, its not just that a
number of guys have done better than you, but that the number in itself is huge. The
prominence on lots, then, emphasizes the number of 100,000 scores, and thus is an
elaboration for power. So, as welters go, that was one heck of an impressively large
welter, and amongst lots, these are no piddling lots, but vast lots.
Prominence also conveys a wide range of moods, attitudes, and charges of
passion, or EMOTION. The portion of that range which falls within the current scope of
this study is the application of adventitious power. In this section, power for emotion
will be distinguished from the rest of this general emotion. Bolingers wave metaphor
188
(1964) frames this distinction well, describing intonation as the ups and downs... of
the fundamental pitch of the voice, and likening it to ripples on waves on swells on
tides (1964: 282). The ripples are irrelevant, involuntary changes in pitch, such as the
first part of a vowels pitch being heightened after a stop due to the release of pressure.
The waves are the familiar patterns of pitch accent, and the tides are the broad surges
of emotion (with swells going undefined). The article in which this metaphor appears
makes the point that when it comes to making the (para)linguistic distinction, such as
in language teaching, ripples can be ignored, waves must be taught, and if tides are
addressed, it is likely that teaching their suppression will be important, given cultural
differences in accepted displays of emotion during speech. The range of emotion
conveyed by prominence in this study only amounts to that which is poured into one
italicized word. It does not deal with long impassioned speeches, or boiling tirades, or
any more widely distributed outpourings, except to the extent that one italicized word
is understood to occasionally take the brunt of such an emotional explosion.
Here is an example of this emotion being discharged with one word, rabbits:
[3-23] [Lennie] laughed delightedly. Go on now, George!
You got it by heart. You can do it yourself.
No, you. I forget some a the things. Tell about how its gonna
be.
O.K. Someday were gonna get the jack together and were
gonna have a little house and a couple of acres an a cow and
some pigs and
An live off the fatta the lan, Lennie shouted. An have
rabbits. Go on George! Tell about what were gonna have in the
garden and about the rabbits in the cages and about the rain in
the winter and the stove, and how thick the cream is on the milk
like you can hardly cut it. Tell about that, George.
6.30
189
The prominent use of rabbits is a strong discharge of Lennies excitement and
emotional investment, and it is not due to his stating that they will have some sort of
unusual or prize-winning rabbits rather than regular ones (position), nor is he showing
his confidence in his belief that they will have rabbits, as if it were specifically in
contrast to Georges belief that they will have cows (revelation). It is also not merely a
mechanical matter of speaking loud enough to be heard. The notion of rabbits simply
carries a strong emotional charge for Lennie, which he discharges with power.
Here is a similar example from the Brown corpus:
[3-24] They held the funeral the next morning from the crossroads
church and buried the little box in the quiet family plot. Kate
moved through all the preparations and services in a state of
bewilderment. She would not accept the death of such a little
child. God called her to Him, the minister had said. God
would not do that, Kate thought stubbornly.
K15 0180
In this example similarly, the word do is not made prominent to distinguish it from
hope or any other process, neither is it being isolated as odd, but rather the woman is
discharging emotion as pain and strength of conviction. The meaning of the prominent
word is changed in these examples such that it stands for an entity which holds great
importance. Neither the strength of rabbits nor that of do is compared to any
earlier use, not even to the typical emotional strength with which that or any other
word might normally be expressed. Just as a battery has a charge without comparing it
to the capacity of other batteries, or to the strength of its own earlier charge, it is useful
to think of such an expression as having a strong emotional charge or discharge
without any comparison or relative difference.
190
The source of this emotional charge is sometimes portrayed as a matter of
disbelief or sheer surprise, which displays variations such as indignance, delight,
bafflement, and repugnance:
[3-25] I was shown, to my surprise, into the kitchen.
R01 1690
[3-26] But do you know something curious? she added. I reached
into that funny little pocket that is high up on my dress. I have
no notion why I reached. And I found a radish.
R07 1170
[3-27] 2000: All right. After I eat asparagus and I make Number One,
theres such an odor. You know? Such a nutty flavor. I mean,
that really puzzles me, why there should be that
19.63
[3-28] His mother took out a tissue, spit on the tissue, and rubbed it
into the kids face. Im not making this up.... You know that if
babies could talk, thats the first thing theyd bring up. Hey,
dont do that. Its revolting. Would you like it if someone did
that to you? Okay, then.
8.330
Outright revulsion (down through mild distaste) also commonly generates an
emotional charge:
[3-29] That this abandonment [of outer guise to display inner self]
takes place on a stage, during an artistic performance, is
enough to associate Jacoby with art, and to bring down upon
him the punishment for art; that is, he is suspect, guilty,
punishable, as is anyone in Mann's stories who produces
illusion...
G15 0810
[3-30] ...Jenny could look forward to years of conflict with an animal
who disliked her intensely and showed it.
P03 0920
[3-31] Jenny wished now that she had.had Dr. Dunne, feeling that
somehow he wouldnt have allowed the dear baby to turn into
triplets. There was something not nice about triplets....
P03 1120
Personally, I experience a mild sympathetic scrunching of my articulators just reading
some of these passages, which motion I associate with one type of greater effort.
191
There are analogs of this emotional function mentioned often in the literature.
In Walker (1787), tonal slides were not only accompanied by pitch, loudness, and
duration, but also some sort of emotional charge, which he characterized as passion,
forcible or feeble (p. 10). Emotion is attributed a relative proportion, rather than a
definite scale such as that associated with pitch and duration, and Walker marks it in
the notation as an aside written in parentheses. Here are a few of my favorites:
Arguing; a cool, sedate, middle tone of voice. (p. 26)
Enquiry, with surprise; higher, and more forcible tone. (p. 26)
Hatred and detestation; lower tone. (p. 30)
Strong grief suppressed; lower tone. (p. 32)
Disturbed; broken lower tone. (p. 32)
Resignation; cheerful, solemn, lower tone. (p. 34)
Over-acted earnest entreaty; high tone. (p. 40)
Dread and terror; low tone. (p. 42)
Enumeration of particulars; firm, loud, lower tone, rising in
strength on each to the end. (p. 62)
Walker provides a good number of these sorts of meanings, but there are no consistent
forms said to accompany them. All of Walkers contemporaries said much the same
thing about this type of emotional expression.
These attributions of power, both the type which is inherent in the intensely
prominent word and the type derived from the strength of Ss emotional charge, are
best understood as a pair of poles along a functional continuum, rather than as two
discrete subfunctions of power. In retrospect, there seems to be little doubt but what I
peremptorily ruled out many cases of emotion on the basis of their having been
included in exclamations. Given this new framework, however, the next pass with
audio or video recordings need not be so negligent.
192
2 Precision
Unlike power, precision is not an effect of prominence which occurs freely
across grammatical categories, but rather it works most compatibly with the likes of
simple or complex relations, types and their instances, and deictics, because they are
most easily characterized in terms of figuratively spatial representations. When
prominence applies precision as INTOLERANCE (2.1), the spatial components in a
words conception are understood to display less than their normal degree of variation.
When the spatial components more precisely define the location of an instance within
its type, then prominence is used as precision of POSITION (2.2).
2.1 Intolerance
To begin with, members of closed classes are not often subject to elaboration at
all, tending to be treated as parts of conventional pairs or sets (chapter 4, 5.1f), as in
[this vs. that]. Except for the prepositions, there is a lack of physicality that makes
them unlikely to be elaborated in terms of power, and their conventionality tends to
keep them from being used for position, especially peripherally. Within elaboration,
that leaves intolerance.
Here are some intolerant or precise conditionals (look for if twice herein):
[3-32] Mr. Philip Toynbee affirms at one point that if he shared the
anticipations of Orwell in Nineteen Eighty-Four, if he believed
Communism was not only evil but also irredeemably evil,
then he might think it right to do anything rather than to take
the risk of a communist world. Even a nuclear holocaust is a
little less frightful to contemplate than a race of dehumanised
humans occupying the earth until doomsday.
D11 1160, D11 1700
193
This is a matter of intolerance to the extent that the set of conditions is being narrowed
down to almost nothing, and there are now very few reasons why Toynbee would
agree to destroy the world rather than live under communism. In fact, he will no longer
do so if communism only proves to be horribly evil, but specifically if and only if it
were precisely irredeemably evil. Thats how much more strict if is in this case
than if. Other examples are: if they're Japs...
N15 1140
(sic); provided the farm is
owned free of debt
F13 0450
; and except... toward impious folk...
R09 0670
. Note that
these are not examples of external contrast, such as [if vs. when].
Now lets take a look at some precise demonstratives. These are not to be
confused with the use of demonstratives as part of a conventional pair or set, as in [this
vs. that], but rather to be compared to them. The following all display intolerance:
[3-33] And that's what I'm going to tell Jim.
P15 1630
[3-34] That does it. That dog has to go.
P16 1380
[3-35] What difference does your batting average make? Or your
fielding average. Or even the way you run bases. I tell you
when it's necessary to hurt in order to win you won't do it.
That's what I mean by no heart for the game.
P24 1440
In none of these cases is that being used to correct the impression that this is what
is going to be told or done or meant, so its not revelation. In each case, that
indicates a severe if not absolute reduction in tolerance that used to be present at a time
when the person might still have been willing to use plain old that, but not any more.
Taking [3-34] as an example, perhaps that didnt do it (chewing up the sofa), and
neither did that (always barking at nothing), but that almost did it (peeing on the
194
carpet), and sure enough, that finally did it (gorging on a ham). In this case, that
not only isolates one instance out of many over time, but could also do so in space if
the residue of earlier offenses were still around (the dog was gorging by the couch).
Now, in [3-35], the coach had already told the ballplayer that he had no heart
for the game, and then told the ballplayer to sacrifice himself during a game. The
ballplayer simply wouldn't do it, counting on his impressive stats to keep him from
getting fired. In the example itself, the coach lists the stats (hitting, fielding, running)
in order to point out that they do not have anything to do with what he had been trying
to describe as having a heart for the game, which he isolates in terms of one quality:
self-sacrifice for the team and the game. The coach is saying, Having heart is not
being a good hitter, and its not being a good fielder, and its not being a good runner,
but being a team player. That is having a heart for the game.
A feeling of power haunts these examples in part because emotions run high,
but mostly because of the immediacy of that, and in any case not due to an actual
application of power. Suppose I were to put the open palm of my hand right in front of
your face, ever so slowly and gently. The imminence in itself would come across as
intimidating, which would feel like power. Similarly, in these examples, precision is
drawing two things closer together, where one of those things is the boundary of one's
fight-or-flight zone, which is an artifact of the relation between the function of deictics
and the composition of the ground. On top of that, in all of these examples, there is
some degree of identification with the person to whom that is being said. What this
comes down to is that prominence is actually used for precision in these cases, but it is
195
treated as having a side-effect of power because one of the boundaries involved is Ls
personal space. There are cases where power and precision do act coequally, rather
than casually, but this isnt it. Even at that, thats not the end of that (cf. Lakoff 1987:
527; the paragon-intonation construction).
The following demonstrative pronouns display an interesting variation:
[3-36] The big factories which are relatively near the centers of our
cities the rubber factories in Akron, Chrysler's Detroit plants,
U.S. Steel's Pittsburgh works often began on these sites at a
time when that was the edge of the city, yet close to transport
(river), storage (piers) and power (river).
J60 0410
[3-37] When that was broken up after the First World War,
[Constantinoples] name was changed once more.
E13 0230
[3-38] She could easily understand why the two men had been startled
to find a strange girl in the back seat of their car (she had
figured that out), but she couldn't understand their subsequent
actions.
L03 1090
In [3-36], the city site is a variable which takes on different values. (These algebraic
terms will be defined more rigorously for use with revelation in chapter 4.) Its value
starts out as the edge, after which it alternates over time between being the center and
the edge. In [3-37], the ancient city is a variable with a value identified by a sequence
of cultures: the indigenous population; then the Megarians; then the Romans; then the
Eastern Romans; then the Ottomans; and so on. (The whole quote is too long to put it
all here.) Example [3-38] is precisely similar in that she works at understanding one
action and then another. In each case, the value associated with the variable changes
over time, and while that can be used to refer to any of those values, that is only
associated with the current value. This change in meaning is precision.
196
At first, it seems reasonable to suggest that this is actually a word-external
change or revelation, where that is being swapped wholesale for a thoroughly
distributed counterpart (e.g. The big factories which are relatively near the centers of
our cities the rubber factories in Akron, Chrysler's Detroit plants, U.S. Steel's
Pittsburgh works), but this sort of exchange is what a pronoun normally does, and
so in this case it is not a function of the volitional prominence. Intolerance allows
that to be used as an increasingly precise word-internal meaning variation of that
in this context, which is elaboration.
As you might expect, regular pronouns behave this way as well, using
prominence to clarify an elaboration chain:
[3-39] Often, I heard my uncles and cousins speak of it when I was a
small boy growing up in Rabaul. They had never seen a
[jumping platform of death] but they had heard about it from
their fathers.
N21 0780
[3-40] Somehow I think that Watson paid more attention to me than he
otherwise might have because his foe, Colonel Van Hamm,
wouldn't touch me with a ten-foot blue pencil. I remember one
day when Mr. Hearst (and I never knew why he liked me,
either) sent the Hetman a telegram....
G40 1600
[3-41] This included Mamma, jolly, generous, and pretty, with whom
they all fell in love, just as Papa had first fallen in love with her
Mamma before he chose her....
G31 1640
[3-42] At the same time, the wallpaper strips themselves seem to be
pushed into depth by the lines and patches of shading
charcoaled upon them, and by their placing in relation to the
block capitals; and these capitals seem in turn to be pushed back
by their placing, and by contrast with the corporeality of the
woodgraining.
J59 0770
197
In each of these cases, this same chaining of values for the variable occurs, and the
elaboration of the pronoun indicates that it points with narrow tolerance precisely and
only at the current value of the variable, namely: 1) the older male relatives of the
older male relatives of the boy; 2) not the first person who is said to like S, but the
most recently mentioned person who provides the value for that variable (Mr. Hearst);
3) not the Mamma, but the Mammas Mamma, who provides the current value for the
variable of the person who was fallen in love with; and 4) the capitals which provide
the current value for the variable something which gets pushed back by its placing
relative to the value of something else. These examples are by nature convoluted.
As mentioned, some examples of revelation are similar, but in those examples
the exchange of the counterpart for the pronoun is a normal function of pronominal
reference even without the volitional prominence. In the current examples, the
switches of value are treated as mundane, a lot of them happen, and what is important
is being able to point to the most current value, whereas in the revelation examples,
what will be important is the fact that there has been a switch.
To analyze volitional prominence with relative quantifiers, I need the terms
GROUND (G) and REFERENCE MASS (M), drawn from Langacker (1987: 126; 1991: 82).
Think of S and L as trying to make mental contact with the same entities for the
purposes of conversation. The speech event and its participants make up the ground,
and S helps L to make contact with the target by using the ground as a REFERENCE POINT
from which to locate the desired entities. Nominal grounding predications such as
the, a, and those ground primarily in G, and only secondarily (if at all) in M.
198
M is the primary ground for relative quantifiers, and it is a conceptual construct
against which to measure the magnitude of the expanse designated by a thing. For
example, some water (as in There might not be much left, but I think that theres
still some water left in the pitcher) does not necessarily imply any connection to any
aspect of G, such as either S or L, but rather primarily compares the amount of water
designated by some water against M, where M in this case is some sort of maximal
extension of the category [WATER], such as all water, or all the water in the
universe. Relative quantifiers are distinct from the TRUE or ABSOLUTE QUANTIFIERS
(several, few, little, numerous, much, many, the open-ended set of cardinal
numbers), which describe an instances magnitude without such reference.
Crucial to this analysis is the assertion that under normal circumstances, some
tolerance is allowed in the conception of the borders of the compared instance relative
to M. For example, a statement such as, All politicians are corrupt, is taken with a
grain of salt, and at least a very few politicians are presumed to be honest, even though
all is specified; however, the phrase All politicians are corrupt leaves no margin for
error, and there is absolutely no such thing as a politician who is not corrupt.
As it turns out, the Brown corpus only contained one example of an intolerant
relative quantifier (no), so I am going to illustrate its behavior under volitional
prominence accompanied by instances from the Other corpus:
[3-43] Certainly, in analyzing an action which truly faced such
alternatives, it is never possible that no world would be
preferable to some worlds, and there are in truth no
circumstances in which the destruction of human life presents
itself as a reasonable alternative.
D11 0290
199
[3-44] Are all of you going? asked Wilbur. You cant all go. I
would be left alone, with no friends. Your mother wouldnt
want that to happen, Im sure.
29.180
(The first all is discluded
as a question.)
If no had not been made prominent in [3-43], the narrator might well have been able
to follow up the injunction with at least a hesitant exception, such as, Well, okay,
maybe if we all lived in a ghastly world where we all endured constant torture or
something like that, then maybe no world would be preferable to that one; however,
the prominence of no removes the option of any hedging like this. Without the
intense prominence on all in [3-44], Ls attention would wander to an emphasis on
some other element (you, cant, or go), which would allow for some tolerance
in the construal of the relative quantity of baby spiders who could or couldnt go, but
all leaves no question in Ls mind but what Wilbur is worried about absolutely every
last one of the little cuties leaving. (Its okay, three of them stay to keep him company.)
Not all relative quantifiers work this way. The ones which do are the four
PROPORTIONAL relative quantifiers (all, most, some, no), because they compare
the magnitude of a profiled mass against M, and the tolerance between the measure of
the two masses is prone to being construed with narrower tolerance.
Of the four UNIVERSAL relative quantifiers (all, every, each, any), the
three non-proportional ones (every, each, any) do not behave this way. While all
four of these quantifiers treat the profiled mass as coincident with M, all does not
designate one arbitrary instance in M. Every scans all the instances simultaneously,
and finding them equal in regard to some contextually critical feature, selects one as a
200
representative example. Each makes a similar selection as the result of sequentially
scanning the whole set of instances. Any chooses a random exemplar, so it is not
necessarily representative. When one of these three individual relative quantifiers is
made prominent, it is this mutually differentiating feature which is made conspicuous:
[3-45] It is interesting that it is not the getting of any sort of knowledge
that God has forbidden, but, specifically, the knowledge of the
difference between good and evil.
17.98
In [3-45], randomly selected knowledge is set against one deliberately chosen specific
sort of knowledge, which emphasizes the very randomness which differentiates any
from every and each; therefore, this is not a matter of intolerance, but rather a type
of revelation, specifically derivation by conventional set. So, prominent individual
relative quantifiers will be analyzed with the other conventional sets (chapter 4, 5.2).
Think of M as a timeline, and it is easy to understand why adverbials of time
are prone to precision. The following example is a particularly good one:
[3-46] Ekstrohm never slept. Some doctors had informed him he was
mistaken about this. Actually, they said, he did sleep, but so
shortly and fitfully that he forgot. Others admitted he was
absolutely correct he never slept. His body processes only
slowed down enough for him to dispel fatigue poisons.
Occasionally he fell into a waking, gritty-eyed stupor; but he
never slept. Never at all.
M04 1200
At first, never is used with the normal tolerance (i.e. Ekstrohm doesnt sleep well or
often, but certainly he must sleep enough to stay alive) which ignores any exceptional
intervals that are short and irregular enough to be easily forgotten; however, never
201
has an absolutely intolerant definition, which supports the later use of Never at all.
Prominent never denies any of the short gaps in never that he is purported to have
forgotten. This is not never as compared to some member of the conventional set of
adverbials of time, but an internal comparison of its precise meaning as used here with
its usual, more tolerant one. The elaboration makes it clear that there are absolutely no
exceptions. Whatever process Ekstrohm experiences may only resemble sleep in some
necessarily few regards. This intolerance is found in two other uses of never
J69 1030,
P22 0530
, with similar examples in only
J08 1490, J41 0710, J58 1300, D02 0490
and ever
P03 0650,
P03 1530
. While the negative adverbials tend to be used for precision, the positive
ones tend to be used for revelation in contrast to other members of the adverbial set.
With regular adverbs, intolerance is the increasing approximation of a limit:
[3-47] It can't be wrong, can it? Not really.
P22 0880
[3-48] [Kennedy is compared to Louis XVI].... not completely
virtuous, but completely incompetent.
B26 1740
[3-47] shows that something can be really wrong without being really wrong. In
[3-48], a type of precision is evoked similar to that for all, comparing how much of
something (virtue or incompetence) someone has versus how much of it exists. In the
case of virtue, the boundaries of the two quantities come as close as they can without
actually matching. In the case of incompetence, they match; if they do so exactly, then
this is precision, but if one is seen to forcefully overrun the other, then it might well be
power. These cases of precision merely look similar to the behavior that Lasersohn
describes as a regulation of pragmatic slack (cf. 3.6, p. 166).
202
These examples have emphasized the difference between intolerance signaling
a word-internal meaning change (e.g. always with some exceptions versus absolutely
always), and derivation by conventional pair or set correcting a mismatch between
the set member which S has in mind (e.g. always), and the one which S thinks that L
is likely to have in mind (e.g. some other adverbial of time, like never). In fact, there
would be no use for intolerance, or elaboration, if a language parcelled every words
range of meaning variations into conventional sets, as if every last one of the degrees
of always were a separate word. Intolerance is the more economical alternative.
2.2 Position
There are two salient positions that an instance can occupy within its type,
namely: 1) the center, as a typical, PROPER member, where some instances are virtual
paragons of their type, and 2) far outside of that center, on the PERIPHERY with other
examples which are atypical for one reason or another. Obviously, this elaboration for
position only applies to those classes of words which have TSs within which to be
proper members, such as nouns and content verbs, or those modifiers whose continua
define the central areas against which these proper nouns and verbs are related.
Simply indicating that this position is necessary to understand the meaning of a
prominent word makes that meaning more elaborate. Establishing the importance of
an instances position is a type of elaboration provided directly by prominence itself,
as if it were shoving an instance around inside of its type, but the direction of that
pushing is determined by the context; in that sense, marking an instance specifically as
proper or peripheral is not a direct function of prominence, but rather of context.
203
There are two further contextual variations of elaboration for proper position,
called simply POSITIVE and NEGATIVE. Positive position seems to establish a PARAGON:
[3-49] She had assumed before then that one day he would ask her to
marry him. Blanche couldnt remember when she had first
arrived at this conclusion. She thought it was sometime during
the second week she worked for Stanley. It was nothing that he
said or did, but it seemed so natural to her that she should be
working for him, looking forward to his eventual proposal.
L10
1280
The word natural is not used here in opposition to something unnatural, neither is it
iconically representing the level of Blanches emotional charge, which is calm indeed,
in addition to which she lacks the surprise, revulsion, disbelief, or the other irreal
overlays typical of elaboration for emotion. She is not currently disbelieving that she
had ever assumed anything so stupid as that Stanley would propose to her. This leaves
us with a critical assertion of purity or centrality, one which portrays this specific
instance of natural as utterly unadulterated by any improper characteristics. Blanche
is specifically drawing Ls attention to the fact that her assumption seemed absolutely
natural. It is a fantastic representative of natural amongst other uses of natural. Just
as there are variable intensities of green, there are variable intensities of natural,
and if natural isnt an actual paragon, it is certainly a contender for that position.
Parts of this context, such as nothing that he said or did, could be interpreted
as providing the source for deriving a counterpart to natural, such as designed, in
which case prominence would be used to elaborate a word which has been used for
revelation. Similar conflations will be analyzed in chapter 4.
204
[3-49] is the only positive proper elaboration in the Brown corpus, so I will
support it with the only two positive examples from the Other corpus. These describe
paragons for firm and good:
[3-50] His hands are firm. I wonder if sculptors dont have thick hands,
too?
4.333
[3-51] There were my first mountains, the Catskills; there were
Ichabod Crane and Rip Van Winkle visible on the blue uplands
or in the mountain gorges; George Washington rode all over the
place; and there were stone houses that, compared with the
frame cottages of a prairie village, seemed to me coeval with
the Acropolis and in considerably better repair. This, I
decided, oh, this was good; I was simply going to love the East,
particularly New York; love it and dominate it.
5.5
[3-50] is not a matter of simply contrasting [firm vs. flabby], but of portraying the
artists hands as firm amongst the firm. Likewise, [3-51] is not just describing a big
amount of good, or a big good thing in itself, but an intensely or precisely good thing.
It is a paragon of goodness, and it is accompanied by an aspiration of awe. These
positive examples are used by S in context to assert that something is in fact a paragon.
Negative position is more common, and S uses it to assert that something else
in context is not up to the standards of a paragon, and so in fact is not itself a paragon:
[3-52] Line? Line? But there is no line between France and Germany,
that is, no actual line...
G47 0540
[3-53] In my estimation, they were people who read Daphne du
Maurier, and discussed Kafka; well, not discussed him exactly,
but said, Kafka! reverently and raised their eyes, as if they
were at a loss to describe how they felt about Kafka, which they
were, because they had no opinions about Kafka, not having
read Kafka.
R02 1590
205
[3-54] Finally, there is the undeniable fact that some of the finest
American fiction is being written by Jews, but it is not Jewish
fiction....
G74 0730
Other examples from the Brown corpus are: wrong
L08 1350
, shape
J59 1520
, and
selective
B26 0270
. Notice that it is the context in particular which determines that the
important proper position is missed, with words like no actual and not... exactly
indicating that the material in comparison to the prominent word is not as proper an
instance of the type. This creates a context in which the phrase as such could
comfortably be added (no actual line as such, not Jewish fiction as such), which
acts as a good diagnostic for identifying this use of volitional prominence.
With peripheral elaboration, regular words fall short, and there is always a lack
of alignment between the meaning associated with the conventional form of a word
and the concept that S is trying to convey. Either the intended meaning of the word
only approximates the concept it normally conveys, or there are no words in English
which match the concept any more closely than the one chosen. Sometimes there is a
social reason for using a different phonological form to get to the associated semantics
in a circuitous manner (EUPHEMISM), and sometimes there is an available phonological
form which is conventionally tied to a semantic structure which is almost what S has
in mind. Sometimes the concept is just too different, strange, or gross for words.
The diagnostic for euphemism is so to speak, and I dont know what the
signed analogs might be:
[3-55] George W. Cable... agitated continuously the Southern
question.
G17 1550
[vs. saying the question of slavery outright]
206
[3-56] Turning in at the Flannagans driveway, he tried to remember if
he had ever met them. The [Flannagans'] name encouraged him,
because he always felt that he could handle the Irish.
K22 0810
[vs.
saying manipulate... or intimidate the Irish]
These are both cases in which a socially acceptable phonological form is understood to
be cross-linked to additional meanings which have poorer connotations than the one
with which the form is conventionally associated. This detour achieves its own degree
of conventionality over time, and S uses volitional prominence to tell L to access the
hidden meaning. They then both pretend that they are dealing with the nobler
sentiments, and not fouling themselves with the offensive, crude, or painful meanings.
In [3-56], handle might also be elaborated for some degree of power.
In the examples above, S is not only aware of the redirected meanings, but
willingly supports the use of euphemisms as a social convention in collusion with L;
however, in the examples below, contextual incongruities suggest that S wants to
expose the charade (if mildly) rather than surrender peacefully. The diagnostic for this
sort of euphemism is so-called:
[3-57] Where were the hardships she had expected? She was certain
now that it would be no harder to bear her child here in such
pleasant surroundings than at home in the big white house in
Haverhill. With childlike innocence she wrote of the Indians as
walking with fruit and umbrellas in their hands, with the tawny
children around them.... This is the most delightful trial I have
ever had, she decided.
G37 1240
[so-called trial]
[3-58] Anyway, Julia asks me to....
Julia?
Come on, Inspector, look alive. Julia Buck, the deceased,
Moore said, slipping me his smug, idiot-grin again.
L20 0380
[so-
called inspector]
207
In [3-57], the contextual incongruity is simply the mismatch between delightful and
trial, and in [3-58] it is a woefully ignorant detective. I dont know that these fig
leaves are considered proper cases of euphemism, but they fit here well as defined.
Sometimes the problem is not that S is trying to avoid the semantic structure
conventionally associated with a phonological form, but rather that S wants to reach a
particular semantic structure, but cannot. The concept is so alien to the language and
its users that no phonological form has ever been attached to a notion like it often or
long enough (if ever) for it to be able to achieve anything like conventional status. The
diagnostic again is as such (or for lack of a better word), and the prominence warns
L that the phonological forms conventional associations are an approximation, and
that the marked form is only being used because no closer link could be located. Here
are some examples of this almost apologetic use of euphemism:
[3-59] Detached from their prior statuses and social groups and exposed
to the pervasive stimuli of the university milieu, the students tend
to assimilate a new common culture, to converge toward norms
characteristic of their own particular campus.
G57 1210
[3-60] II didnt do it deliberately, just suddenly I was in a dream that
wasnt my own and I made it change.
24.30
In [3-59], S is trying to describe the homogenizing effect that cohabitation has on the
(allegedly) previously diverse behavior of two or three thousand young adults living in
the same dorm rooms for an academic year. There simply are no really good words to
describe what amounts to this new social structure, relative to the age of the language.
S resorts to using the word culture because thats as close as he can get, a concept
which typically involves a much greater number of people living together and raising
208
families over several centuries as members of the same genetic and geopolitical set.
This mismatch is the internal meaning change applied to culture. In [3-60], S wants
to find a straightforward link to the semantic structure that she has in mind, but she is
simply unable to locate a conventional phonological association with that structure in
a language whose users are virtually all nontelepaths. In these examples, the semantics
are peripheral not due to ostracism, but due to novelty.
In some such cases, the concept is just too different for words, and there is such
a strong feeling of inaccessibility that S just gives up and resorts to a pronoun. In the
following examples, God and the familiar thing are both at a distance:
[3-61] I am merely a channel for... something.
D17 0180
[3-62] Again there was something familiar about her, something...
P19
1440
The diagnostic here is still the feeling of grasping for the right word, any word, which
will explain what S is trying to encapsulate, accompanied by the feeling of giving up.
The peripheral elaboration means that the something in question is inaccessible,
elusive, deep, small, and isolated, whereas the revelatory use would indicate that it
was just accessible enough to avoid being swapped for nothing at all.
This next example stands out because on the one hand, prominence might be
used to point out that English has no word for all manner of Martian creatures:
[3-63] McAuliff, for good reasons, had a hypochondriacal view of his
cows; he suspected that all manner of Martian things were out
to get them, to make them lean, sick, and fitful in their milk
production.
35.16
209
On the other hand, prominence is definitely used to express revulsion here as an
emotional charge. My analysis leans towards a simple categorization as elaboration for
emotion without any overlap with periphery on the basis that the word things,
although schematic, is a perfectly good word for all manner of Martian creatures.
In peripheral elaboration, S lets L know that the semantic structure
conventionally associated with the prominent phonological form is not the one that S
has in mind, and so L aims at a target towards the edge of the words possible range of
meanings. Euphemism can be used to support or ridicule taboo semantic structures,
where S avoids phonological forms that are properly connected to a concept in favor
of ones which are only linked peripherally. Sometimes S wants to make a direct
connection, but cannot due to the novelty of a semantic structure, so S settles for the
closest appropriate phonological match, which might be no nearer than a prominent
indefinite or interrogative pronoun.
3 Conclusion
S wants L to understand exactly what S means, and so S necessarily wants their
construals of a given semantic structure to be as closely aligned as possible. Volitional
prominence for elaboration repairs or avoids those misalignments which are caused
when S wants L to associate a phonological form with a different semantic structure
than normal. L interprets prominence as a signal to draw upon a meaning which is
either more powerful or more precisely intolerant than normal, or whose relation to
other word-internal variations in meaning is crucially either more perfect or more
210
peripheral than normal. S physically articulates a change in the level of prominence in
the phonological form, which L cognitively articulates as a directly proportional
change in the level of prominence in the semantic structure. This direct proportion
between these physical and cognitive efforts is iconic. Elaboration as a whole, then, is
said to be a function which affects phonological forms and their associated meanings
in direct iconic proportion to one another.
211
CHAPTER 4
Revelation
How can I be mistaken in what I say, unbelieving traitor?
returned Don Quixote; tell me, seest thou not yonder knight coming
towards us on a dappled grey steed, who has upon his head a helmet of
gold?
What I see and make out, answered Sancho, is only a man on a
grey ass like my own, who has something that shines on his head.
Well, that is the helmet of Mambrino, said Don Quixote.
Miguel de Cervantes, Don Quixote (1615)
When engaged in conversation, people rely upon a significant overlap between
their individual sets of beliefs about reality, patching the occasional mismatch with the
strategic use of volitional prominence. As S updates an exchanged discourse space, or
unilaterally unfolds a monologue, S marks the most important additions with routine
prominence, and resorts to volitional prominence to pinpoint any necessary
corrections:
[4-1] One man remarked that if he had a hundred pounds, he would
give ninety of them to be back in England. Up spoke carpenter
Staffe, who said he wouldn't give ten pounds to be home.
F16 1310
To the degree that routine prominence marks new information (ninety in the context
of the first sentence), volitional prominence reveals new information about old
information ([ten vs. ninety] in the second), whence the term REVELATION. S is
revealing Ls mistake to L.
212
Revelation is usually a volitional shift of prominence specifically away from
its routine location, and it only raises the prominence level of the word it falls on as a
consequence of that movement, since that word would otherwise only have secondary
prominence at best. That is why such a volitional placement requires only a primary
level of prominence to uniquely identify a word as volitionally prominent, rather than
an intense level. In contrast, elaboration raises the prominence level of a particular
word without regard to its location, dragging the primary sentence prominence along
with it as a side-effect. The routine location of sentence prominence normally has a
primary level of prominence, and so an elaboration or revelation at that location will
need to be intense to distinguish it from the routine pattern.
Here is an example of revelation which results from volitional placement:
[4-2] Suppose it was not us that killed these aliens.
M04 0900
This notation represents the pattern, Suppose it was nt us that killed these aliens,
with a primary level of prominence on not, as opposed to Suppose it wasnt us that
killed these liens, with a primary level on the first syllable of aliens. Underlining
the whole word only indicates a volitional prominence level on one syllable, where
even though the change only needs primary prominence to differentiate the utterance
from the routine pattern, sometimes it gets intense prominence anyway.
In those cases where the placement of revelatory prominence does coincide
with the routine location of sentence stress, the level of the revelatory prominence will
need to be greater than that which normally associates itself with the primary sentence
213
stress. The revelatory prominence will need to be intense to make it stand out from the
primary prominence that would normally go on the routine location. For example, a
primary level of prominence would normally fall on the last word of the following
example, in accord with the routine placement of sentence stress:
[4-3] I was shown, to my surprise, into the ktchen.
Adapted from R01 1690
When this expression gets revelatory prominence, the same notation for volitional
prominence is used as was shown in [4-2]:
[4-4] I was shown, to my surprise, into the kitchen.
R01 1690
The difference here is that [kit] needs intense prominence, and so such an example
should be understood in this case to represent a phonologically intense level, as in, I
was shown, to my surprise, into the KITchen. The underlining in the examples,
therefore, does not represent an absolute level of prominence, but one which is at least
one increment higher than the routine level at that location (cf. chapter 2, 3.8).
This same utterance can also be interpreted as an elaboration for power, where
a person who had overly fastidious tastes regarded the kitchen as an odious room in
which to wait, and was surprised to be there rather than in the parlor. Both the
revelation and the elaboration in those cases require an intense level of volitional
prominence.
Volitional prominence is almost always marked on whole words, but it can also
be used to single out part of a word:
214
[4-5] Since then, and since the pure grain had gotten him divorced
from every decent and even indecent group from Greenwich
Village to the Embarcadero, he had become a sucker-rolling
freight-jumper.
N29 0040
[4-6] If it is an honest feeling, then why should she not yield to it?
Most often, she says, its the monogamous relationship that
is dishonest.
G13 1140, G13 1140, G13 1150
[4-7] He made many tasteless, irreverent and unfunny remarks, not
only about me in the title role, but about religion in general.
R03
1550
[4-8] And if youre the surprisee, its even worse
8.256
[4-9] I sensed that there was a deep tension between the two, but I
was used to behavior that repressed rather than expressed.
37.37
[4-10] The next step of mind and imagination is to grasp the fact of
long-term change in the larger design of an ocean shore or
continent.
38.28
These are the only such cases in either corpus (plus one pair listed in chapter 5), but
none of these prominent affixes is an overt grammatical marker, such as for subject-
verb agreement. This gap in the data set is not just because the appropriate contexts are
all that uncommon in spoken (and perhaps signed) language, but because writing
conventions tend to disallow forms such as I was painting the shed [vs. painted], or
Im not just going to have a pool, Im going to have pools (poolzzz or pool-ZUH).
In such cases, the whole word tends to be underlined. Even at that, there are no
prominent nouns, for example, which are used to make count/mass, number, or
common/proper distinctions. Examples will be given later which show what happens
in cases where this phonological dependence is a matter of wholesale integration, such
that revelations can be made for tense or aspect.
Thats all there is to the notation used for revelation.
215
Now, every revelation can be analyzed into a set of characteristic components,
namely: 1) a change; 2) a setting; 3) a counterpart; and 4) an omen. These all appeared
in the first example, repeated here for convenience:
[4-11] One man remarked that if he had a hundred pounds, he would
give ninety of them to be back in England. Up spoke carpenter
Staffe, who said he wouldnt give ten pounds to be home.
F16 1310
change = ten
setting a trip back home has a value
counterpart = ninety (explicit, thoroughly encapsulated)
omen = Up spoke carpenter Staffe (distributed)
To the degree that background or shared information is analogous to Krifkas matrix,
the SETTING is parallel to Krifkas restrictor (cf. chapter 2, 3.5). It is the coefficient of
Jackendoffs appropriate semantic variable, namely the portion of the background
which most immediately cradles the variable which has been assigned the value of the
newest information or focus (cf. chapter 2, 3.5). The setting in this case is a model
wherein a trip back home has a value. That value is variable among Ps, but because it
is archival knowledge, it gets treated as if the value should be the same among Ps. This
setting is that portion of the discourse which aligns so well with carpenter Staffes
beliefs about reality (or his current model of the discourse space) that the discrepancy
within it caused by the one man who remarked just stands out too strongly for Staffe
to ignore. Staffe believes that a trip back home has a low value, and he takes exception
to the glaring error made by the man who spoke up and suggested that it had a high
value instead. That man must be made aware of the error of his ways.
216
In its turn as S, P is motivated by its perception of the discrepancy to respond
with an utterance which uses a volitionally prominent word, one which identifies and
insists upon a CHANGE to correct the mismatch (ten). The error is represented by the
DISCOURSE COUNTERPART to the change (ninety). These counterparts vary both in their
EXPLICITNESS (their expression in so many words) and in their ENCAPSULATION (their
expression in how many words, explicit or otherwise). I have come across no cases in
which the counterpart was actually absent, although this configuration is theoretically
possible to the degree that a word might be made volitionally prominent to emphasize
its appearance out of the blue (INITIATION). The problem would be in distinguishing
examples of that sort from those in which the volitional prominence was intended to
convey an emotional value representing shock or surprise over the unanticipated
appearance of the marked word.
The OMEN is a qualifying word or phrase which heralds an imminent change
(Up spoke carpenter Staffe). Like a counterpart, an omen varies in encapsulation
(distributed across four words in this case), but unlike a counterpart, the absence of an
explicit omen is not uncommon, and specific omens are not left implied. Omens
constitute a subset among Fauconniers (1985) SPACE-BUILDERS, supporting the
characterization of discourse in terms of his MENTAL SPACES.
Changes have been portrayed so far in terms of one S making a statement to
which another S takes exception in turn, but in a common variation on this theme, S
anticipates that the other conversants will object to its portrayal of reality, and so it
heads off any opposition by insisting upon a change according to what S expects the
217
others to believe. In such cases, Ss utterance provides: 1) the setting (e.g. Like lots of
ignorant slobs, you probably think that tomatoes are vegetables); 2) which cradles the
counterpart that S intends to refute (vegetables); as well as 3) an omen that not only
heralds an impending change (Well, let me tell you); 4) but which characterizes S
as an authority (I know all about this stuff); and then 5) S tops it all off with the
change (and tomatoes are a fruit.). The notion is to silence any objection before it
has a chance to be expressed. Revelations by authors in their own books can usually be
attributed to their making these sorts of authoritative statements as they promote a
particular belief that they expect the reader to adopt.
This is simply one way to portray changes in belief as being associated with
the reception of new information. There is experimental evidence to suggest that initial
comprehension of new information entails belief, and that people by nature tend to be
gullible rather than skeptical, at least for a moment (Gilbert, Krull and Malone, 1990;
Gilbert, 1991; and Gilbert, Tafrodi and Malone, 1993). People automatically, and at
least momentarily, accept everything as true upon initial exposure ( la Spinoza), only
changing this marking later if the material proves false when it actually comes under
evaluation. In this system, unmarked material is true, and marked material is false.
The alternative disallowed by these experiments is one in which discourse and
other perceptual material are not evaluated for veracity during a persons initial
exposure, but are only explicitly tagged as true or false under later evaluation ( la
Descartes). This system would produce initially unmarked material which was neither
true nor false, then some material which was marked true under later evaluation, as
218
well as some which was likewise marked false. Explicitly marking only false items in
the first system is more efficient, and it is also more adaptive, allowing a creature to
react more quickly to perceptions of its environment, only wondering later whether the
sensation was a mirage.
These experiments indicate that people are likely to accept things as true, even
if only for a moment. They also show that interrupting a persons processing will tend
to block reevaluation, leaving the material categorized as true. I suspect that volitional
prominence helps to keep L off-balance or interrupted (with L often a willing victim in
the interests of promoting closer understanding), thus increasing the likelihood that S
will be believed (whether or not L actually adopts the belief, or simply acknowledges
Ss belief). While these studies do not go into detail about the mechanics of subsequent
reevaluation, my analysis would suggest that people tend to maintain their beliefs, and
so intense prominence is needed to add authority to the pronouncement. Keeping L
off-balance during this pronouncement might make them more prone to giving it
weight, just as an interruption would do in the initial stages of perception. In addition,
if comprehension equals belief, then anything that makes a perception loom large will
increase the likelihood of its comprehension, and this study suggests that this entails
an increased likelihood of belief, or at least that is what S expects.
This description provides enough of an overview of the salient components of
revelation to make a deeper discussion of its behavior navigable. The three main types
of revelation are differentiated below according to the explicitness and encapsulation
of their counterparts, and so I am going to continue this chapter with a more detailed
219
description of changes and their counterparts in terms of those two criteria (1). This
will be followed by a more thorough discussion of the setting (2), because its careful
characterization is crucial to the best understanding of the discourse counterpart. Each
of the revelation subtypes will then be discussed in its turn, namely: substitution (3);
addition (4); and derivation (5).
1 Change and Counterpart
Volitionally prominent personal pronouns provide excellent early examples of
discourse counterparts because not only are there plenty of instances of them in the
Brown corpus, but those instances are distributed across a wide range of explicitness
and encapsulation values, as shown in the following examples:
[4-12] The whole act is tailored to her pleasure, and not to theirs.
F08
0160
[4-13] Next on his program was a call to the Jackson office of Peerless
Business Machines to find out if Vincent Black was still with
them or, more specifically, still with us.
L07 1370
[4-14] But the only love I was giving him was the pure kind.
P22 0430
[4-15] His rock was to the right of a V-shaped inlet.
P16 0100
[4-16] Holy mackerel, that's the most unique dog I ever saw, she
said firmly.
P16 0950
Linked instances of volitional prominence are analyzed in chapter 5, but Ive included
[4-12] here as an example of a linked counterpart/change pair (her, theirs). Both
are as encapsulated and as explicit as they can get, because they are both single words
sharing intense prominence. In [4-13], the counterpart them is no less encapsulated,
220
but it is less explicit because it has weaker prominence. Counterparts with this lower,
but routine, level of intensity should be treated as typical (normally explicit). They are
almost five times as common as the intensely prominent, linked counterpart/change
pairs which appear in instances of linked prominence.
The counterparts for I in [4-14], namely both Johnnys wife and his lover (a
different lover than S), are also explicitly mentioned earlier in the passage from which
this example was taken, but they are not as thoroughly encapsulated because: 1) the
two of them are referred to with separate words, and therefore they are referred to in at
least minimally different places; and 2) those words are short phrases. In [4-15], the
implicit counterpart to his is understood to be any one of a number of possessive
pronouns representative of anyone else tied to the beach in such a way that they might
think of themselves as having their own rock. Finally, [4-16] comes out of the mild
blue, and so the counterpart to I can be interpreted either as 1) standing in contrast to
everyone else in the world, not represented pronominally, which is about as implicit
and distributed as it gets, or 2) only being in contrast to the other members of the
conventional set of personal pronouns, which is less distributed, but no more explicit.
2 Setting
Counterparts and changes reside in a setting, in that they are values assigned to
variables couched in their setting. Example [4-12] cashes in on the stereotypical
portrayal of sex where men are insensitive jerks who are by nature able to do no better
than to use women for their own selfish pleasure and so forth. This can be boiled down
221
into a pithy model which holds a variable that can take such values as her and
theirs, namely the setting sex is for only one persons pleasure. The analysis of
linked instances of volitional prominence in chapter 5 will identify this as a case of
DISSOCIATION, which divides the domain of pleasure into that which is her pleasure
and that which is theirs, as specified by the setting.
In [4-14], the setting is arrived at differently. It is true that one proposed setting
could be the equally ugly set of double standards for women which revolve sluggishly
around myths concerning fidelity and the looseness of their collective and individual
virtues; however, there is a specific, parallel line earlier in the context which describes
someone other than S as giving Ss boyfriend love of the not-so-pure or adulterated
kind. Schematically, this contextual setting is y gives him ((im)pure) love. In [4-13],
the setting is also not derived from an archival model, ugly or otherwise, but just the
contextual phrase still with x, even though the phrase at large has ties to a complex
domain which involves an employee being with a company.
[4-15] needs an archival setting because none exists in the context. It has to be
broad enough to house all of those other potential rock-havers, or perhaps a notion as
general as a person experiencing some typical sort of mild territoriality as a result of
becoming more comfortable in a particular environment. Other than the boy, no such
person is identified specifically in the passage from which the example was lifted, but
the fact that at least the boy has a rock proves that it can happen to a beach-goer, and
the presence of other people on the beach makes them all potential rock-havers, which
is what elicits a setting involving territoriality.
222
The setting for [4-16] is also broad, because even though the scene only really
encompasses the girl giving her estimation of the dog to its owner, the counterpart can
be taken to be anyone else in the whole world who might be able to have an opinion
about the dog, which makes the setting itself something along the lines of the vast
notion of people having opinions, as cast in terms of the idea that the value of one
persons opinion can be rated against anothers. The range of the counterpart to the
change depends upon the interpretation of the context by L, or if you prefer, upon the
intent of the interpretation by S. The change is I, so if the girls assertion is taken to
mean that she is specifically comparing her opinion to the boys, then the counterpart
is simply a reference to that boy (he, you, Jeff), but if it is her opinion against the
worlds, then the counterpart refers to the opinion-holders of her world.
These descriptions of change, counterpart and setting are now strong enough to
be used as tools to pull apart the rest of the examples of revelation from the corpora,
starting with substitution.
3 Substitution
Substitution uses all of the components and obeys all of the rules, defining the
hub of typical revelatory behavior from which the others extend. There is a setting, in
which one word acts as an explicit, unitary counterpart to the change. There is almost
always at least one omen of the oncoming change, and then there is the change itself.
There are two radiations of REGULAR SUBSTITUTION (3.1), namely RECAPITULATION
(3.2), in which the counterpart is a duplicate of the change, and REFERENCE AND
223
PRONOUNS (3.3). These sections will provide background conducive to a discussion of
REFERENCE AND PARALLEL STRUCTURES (3.4), in which volitional prominence signals a
detachment of reference, but does not identify the reattachment necessary for a full
disambiguation of reference.
3.1 Regular Substitution
Here are some typical examples of substitution from the Brown corpus:
[4-17] We were at a party once and heard an idealistic young European
call that awful charge glorious.
G75 0660
[4-18] But to say that at a moment in history something is new is not
necessarily to say that it is modern.
J57 1550
[4-19] Lacking the pioneer spirit necessary to write of a new economy,
these writers seem to be contenting themselves with an old one
that is now as defunct as Confederate money.
G08 0670
[4-20] Truman Capote is still reveling in Southern Gothicism,
exaggerating the old Southern legends into something beautiful
and grotesque, but as unreal as or even more unreal than
yesterday.
G08 0550
[4-21] Almost everything about the movies that is peculiarly of the
movies derives from a tension created and maintained between
narrative time and film time.
F33 0780
This is by far the most easily identifiable form of revelation, where the substitutions in
these examples are as follows: [glorious vs. awful]; [new vs. modern];
[new vs. old]; [more vs. as]; and [of vs. about]. Substitutions will often
be represented in this analysis with this particular form of the square-bracket notation,
where both words are in quotes because they both appear explicitly as single words in
their example, rather than having to derive the counterpart (5).
224
Whereas L has the ability to appeal to conventional pairs, sets, and domains in
order to derive the appropriate counterpart necessary for making mental contact with
the instance that S has in mind, substitution amounts to simply being told right up front
specifically what that counterpart is. Substitution has the advantage of being able to
establish its own conventions, such that while new can be used as a conventional
counterpart to old, as in [4-19], it can also pair up with modern, as in [4-18], or
even with green in the right context (I dont care if the cars green, just so long as
its new.)
Other substitutions include: [cascaded vs. channel-type]
J78 0270
; [feeling
vs. listening]
P09 1160
; [writes vs. assembled]
C14 0830
; [used vs. interpreted]
D02 0140
; [was vs. like]
P17 0840
; [risk vs. certainty]
D11 1080, D11 1120
; [need vs.
reason]
L10 1520
; [treatment vs. retention]
F11 1110
; [money vs. cash]
K22 0920
;
[mankind vs. it]
D11 0810
; [involuntary vs. voluntary]
F07 0680
; [complicity vs.
conditions]
F48 1590
; [basic vs. real]
J41 1560
; [human vs. inhuman]
G22 1620
;
[conduct vs. cause]
F48 0690
; [personality vs. type]
G08 1110
; [pace vs.
content]
J27 0860
; [even vs. especially]
M03 0900
; [try vs. can]
B23 1510
; and (as
seen earlier) [ten vs. ninety]
F16 1310
.
Here are some similar substitutions in which the counterparts (not the changes)
happen to be indefinite pronouns (someone elses marginally cheats as two words):
[4-22] Insofar as these nations claim to incarnate traditions and ways
of life which constitute ultimate, trans-political justifications for
their existence, such people are inevitably led to emphasize the
ways in which these traditions and ways are theirs rather than
someone elses.
D10 0890
[theirs vs. someone elses]
225
[4-23] I do not mean to suggest that these assumptions are self-evident,
in the sense that everyone agrees with them I do mean,
however, that I take them for granted, and that everything I shall
be saying would appear quite idiotic against any contrary
assumptions.
F23 0050
[I vs. everyone]
[4-24] From an initial investment of $1,200 m 1943, it has grown, with
no additional capital investment, to a present value estimated by
some as exceeding $10,000,000 (we don't disclose financial
figures to the public).
G22 1870
[we vs. some]
And of course there are cases in which the change itself is an indefinite pronoun:
[4-25] They couldn't have much dough, but then none of the freight-
bums Feathertop rolled had much.
N29 0610
[4-26] I've been ready a long time goodness, we all have
M01 0930

[vs. (all in Mikes commune)]
Notice that when all acts as an indefinite pronoun, it is taken as counter to a small
quantity such as one or few, and in this case I, whereas it runs counter to some
when it appears as a prominent quantifier, or is elaborated for precision when used as a
prominent proportional relative quantifier.
Sometimes, a series of substitutions is made in close succession with the intent
of drawing causal or consequential links between them but not this time:
[4-27] These rumors of permanent separation started up a whole crop
of stories about her. One had it that a friend, protesting her
snobbery, said, But, Gracie, you are an American, arent you?
and she replied, I was born in America, but I was conceived in
Vienna.
G67 1240, G67 1240, G67 1240
This is just a coincidence in this case, or close instances of prominence. Actual
instances of linked prominence are analyzed in chapter 5.
226
3.2 Recapitulation
As mentioned in chapter 2 (3.2), recapitulation is not a type of behavior that I
had considered until I came across a description of it in Gunter (1966), by which time
I had already gathered the Other corpus. Gunter portrays recapitulation in the
following way, where the whole first sentence acts as the setting, and the rest of the
sentences are speculated responses (as listed in a form more like my own notation):
[4-28] The man can see the by.
The man can see the by.
The man can se the boy
The man cn see the boy
The mn can see the boy
The way that Gunter describes recapitulation, it is used by L to confirm that S and L
are in agreement, just so long as the volitionally prominent material in particular is
understood by S to be correct. According to Lambrecht (1994), these should only
occur with volitional prominence when they have argument-focus structure (cf.
chapter 2, 3.5, towards the end on Lambrechts allosentences).
The problem is that nothing as strictly parallel as these examples ever came up
during the sifting of data from either corpus, except for these two examples from the
Other corpus, where the identical strings are only three words long:
[4-29] REINER: Im a little queasy about this, telling tales about
Presidents and Presidents wives.
2000: Theyre all a little power-crazy, right? And they love to
do it. Lets face it. They love it. They love it. They LOVE it.
19.89
[4-30] Heres what we do. I got it, I got itI got it. Heres the plan.
Ill get the chicken, and you get the salmon.
8.248
227
The problem is that L is not repeating S in order to confirm agreement, but rather S is
repeating S more powerfully to drive home a point. In any case, to the degree that the
first example uses successively greater levels of prominence to increase the amount of
force attributed to love, and to the degree that it works like elaboration in all ways
other than that the change happens to be repeated in the setting, I suspect that it is
elaboration (and so it has already been analyzed in that chapter). To the degree that in
the second example, increased prominence is used for confirmation rather than power,
I suspect that it is actually a type of derivation discussed below (5.3) which appeals to
the conventional positive/negative opposition.
Given the paucity of these sorts of examples, it seemed like a good idea to try
to adapt the function of recapitulation to work with the patterns that were more readily
available in the data, and so I tried to use the function to classify instances similar to
substitution, with the exception that the only repetition required would be that just the
change and its counterpart would be the same word. Even at that, there were only four
examples that came out of the Brown corpus, out of which this one instance was the
most clear:
[4-31] The major weakness of vocational training programs in labor
surplus areas is their focus on serving solely local job demands.
This weakness is not unique to labor surplus areas, for it is
inherent in the system of local school districts in this country.
Planning of vocational education programs and courses is
oriented to local employer needs for trained workers. All the
manuals for setting up vocational courses stress the importance
of first making a local survey of skill needs, of estimating the
growth of local jobs, and of consulting with local employers on
the types of courses and their content.
J38 1110-1130
228
There just doesnt seem to be any specific reason to identify this as anything other than
three repetitions of a derivation which is based upon the conventional pair [local vs.
remote < outlying, removed] (5.1). In other words, simply calling this behavior
recapitulation doesnt seem to serve any purpose. The other two examples are
another repeated derivation and a case of coordination, but their analyses have both
been put off until the section on conventional sets (5.2). Of course, to the degree that
these three close instances of local can be interpreted as having a consequential link,
[4-31] is an example of coordination, cf. chapter 5, 2. In any case, none of these
examples motivate the establishment of recapitulation as a type of revelation.
With the revised definition of recapitulation, I managed to find twelve more
likely candidates from the Other corpus, but half of them turned out to be examples of
elaboration for power, like this:
[4-32] Her mind echoed with Stephens voice, and, try as she might,
she could unlock no secret meanings from his words. She did
trypressing deep into those dim memories of her infancy for
words and phrases.
24.83
and the rest were derivations based on the familiar positive/negative opposition (three
with prominent content verbs, and three with prominent auxiliary forms of BE), like
this:
[4-33] Its a perfectly beautiful egg sac, said Wilbur, feeling as
happy as though he had constructed it himself.
Yes, it is pretty, replied Charlotte, patting the sac with her two
front legs.
29.145
229
Everything considered, the behavior in this analysis which most closely resembles
recapitulation is a type of derivation which uses an appeal to a basic domain in order to
find the appropriate counterpart to the change (5.3), which will be defined later as the
confirmative version of appealing to the conventional positive/negative opposition. As
this analysis stands right now, there is no application for the term recapitulation.
3.3 Disambiguation of Reference and Pronouns
There generally isnt enough semantic content in the TS of a personal pronoun
for elaboration to work (cf. chapter 3), and so virtually all of the volitionally
prominent pronouns are cases of revelation, either substitution or derivation, with one
instance of my providing addition. This is in comparison to demonstrative pronouns,
for example, which are often usefully elaborated for precision. Crucially, while
revelation draws Ls attention to the fact that the value of a prominent pronoun (as a
variable in the setting) has indeed changed, revelation does not actually disambiguate
pronominal reference because it does not identify the new replacement value.
The trick is that pronominal reference is similar in behavior to substitution
even in phrases which have routine prominence patterns. In a generated phrase like,
Tammy let Shelby crawl into bed with them, but Shelby drooled on the pillow, so Jeff
made her get out of bed, there is either a regular pronominal (but not prominential)
substitution of her for Shelby (a dog), or there is a pragmatically somewhat less
likely pronominal substitution of her for Tammy (a human). The prominence
pattern can be routine in both cases.
230
Now, when volitional prominence is added for revelation, as in Tammy let
Shelby crawl into bed with them, but Shelby drooled on the pillow, so Jeff made her
get out of bed, there has to be a mismatch related to her which gets corrected in the
interpretation, or else the additional prominence would not have been applied in the
first place. This is definitely not a disambiguation of reference, because the volitional
prominence is not used to indicate that her is intended to be the pronominal
substitution for Shelby rather than Tammy, or vice-versa. S uses prominence in this
case to disabuse L of the notion that Jeff might be the one who is going to get out of
bed rather than her, no matter who her happens to be. The prominence goes on a
referentially ambiguous pronoun, but the revelation is a substitution of her for Jeff.
These sorts of examples are not the same as the sequential value examples
from the chapter on elaboration. Volitional prominence for revelation emphasizes the
change in the value of the participant in a process ([her vs. Jeff]), but elaboration
emphasizes the change in the values of the participants in a process by precisely
selecting one value from among many.
In each of the following substitutions, both the change and its explicit,
encapsulated counterpart are personal pronouns, but the change and counterpart each
have different referents:
[4-34] The air, he said, was just right; a cigarette would taste
particularly good. I really didn't know what he meant. It was a
nice day, granted. But he knew; he sniffed the air and licked it
on his lip and knew as a vintner knows a vintage.
G05 0790
[4-35] His advice, his voice saying his poems, the fact that he had not
so much as touched her on the contrary, he had put his head
back and she had stroked his hair this was all new.
N19 0620
231
The prominent pronouns are not referentially ambiguous here, with the substitutions
[he (y knows) vs. I (x know
not
)] and [his (stroke y) vs. her (touch
not
x)].
The semantic similarities between the verbs are treated as significant, and their
differences are ignored (know know
not
; stroke touch
not
), as is the difference
between the possessive and personal pronouns in his hair and her. The setting is
that someone is participating in a particular type of process, where the variable holds
the identity of the participant, the value is a person in the context, and the pronoun
acting as that value is a POINTER which is substituted for a participant. The volitional
prominence does not substitute one participant for another, but rather it changes one
pointer for another as the value of the variable, and then by normal processes of
reference, one pronominal referent gets substituted for another. The point once again is
that prominence is not being used for a disambiguation of reference.
There are examples in which two pronouns share the same phonological form
before the addition of prominence, and in such cases volitional prominence is prone to
being mistakenly described as if it were placed on one of the two pronouns in order to
distinguish between them, as follows:
[4-36] He limps, and the man who hit you and took the cane, he
limped.
L13 0160
An argument might be made that volitional prominence is used here for precision to
elaborate the meaning of the word he, thus distinguishing it from the he at the
beginning of the sentence, but the fact of the matter is that the two pronouns only
232
match in phonological form because of the context. In a different context, any pronoun
could have been made prominent instead of he, such as you: He limps, and after
you tried kicking the door down, you limped. No one, however, is going to argue that
volitional prominence is used for precision to elaborate the meaning of he into
you. The confusion arises simply because the two pointers he and he have the
same phonological form before the addition of prominence.
Confusion of this sort also arises when two referents participate in the same
process, particularly BE:
[4-37] But my people Martians, I mean; I now grok you are my
people teach plants another way.
M01 0700
Heres the case: [you (are x) vs. Martians (are x)], where (x = my people).
Now here is the problem: because the value of the variable in each case (Martians,
you) is equal (are) to the same thing (my people), there is an impression that
volitional prominence is somehow disambiguating this reference by marking one of
the two equations as the most newly true. Volitional prominence is used to imply that a
correction is being made, and that you should be the value of the variable in the
setting rather than the one which is there (Martians in this case), but it is the context
which defines that setting as [some people are my people]. The omens (the pause, I
mean, and now) and the setting do all the rest of the work.
All of these examples of substitution make it clearer that this same behavior is
displayed in some cases of derivation (by conventional set):
233
[4-38] there was an anecdote about a group of English and
Americans visiting Germany, more than a hundred years ago. In
the railway station at Berlin, a uniformed attendant was
chanting, Foreigners this way! Foreigners this way! One
woman went up to him and said, But you are the
foreigners.
R06 0410
In this case, you is understood implicitly in the chant (You) Foreigners this way!
This example works the same way as a substitution, in that the prominence emphasizes
that there has been a change in a variables value without actually specifying that new
value, but in this case the counterparts value (a pronoun/pointer) needs to be derived
since it is only implicitly understood in context: [you (ref. x) vs. you (ref. y)].
The point is that disambiguation of reference is not a subfunction of revelation,
but rather that revelation can be used to help disambiguate the reference in a scenario
like this one, as can other linguistic devices.
3.4 Reference and Parallel Structures
This is a special application of substitution, and it is prone to being used in
some fairly complex, convoluted utterances. The problem is that volitional
prominence added to a pronouns form can have an effect on the meaning of a referent
located somewhere else. I am going to start out by using some simple, generated
examples to explain this behavior and then lead into a description of the complex,
actual instances from the corpora. What will end up happening is that the generated
examples involving single substitutions will be explained in this section, and they will
be used to introduce the description of the parallel reference behavior, but then the
234
remainder of this part of the analysis will be deferred to a section in the chapter on
multiple instances of volitional prominence, which is where the actual examples from
the corpora deserve to be categorized and treated. I will do what I can to keep the
natural convolution of the data from infecting the prose used to describe it.
To help keep things straight, I am introducing one addition to the notation.
Until now, a single level of underlining has been sufficient to mark the single
occurrence of primary prominence, but the examples in this section need to have the
locations of both their primary and secondary prominence marked. In this section
alone, then, 1) a single underline will be used to mark the word which takes secondary
prominence, 2) a double underline will mark the word which gets primary
prominence, and 3) CAPS will be used for any word which has intense prominence. I
realize that a potential for confusion lies in the intuition that primary = single and
secondary = double, but Im appealing to a different intuition, namely that more
lines = more prominence, which is an intuition I prefer because its notation is
consistent with that used in the rest of this analysis where the most lines mean the most
prominence (e.g. secondary vs. primary stress).
I credit getting this material clear in my mind with having had the opportunity
to compare research methods and results with those of Jennifer Balogh, who is
performing some interesting psychology experiments on abnormal prominence and
the disambiguation of reference in parallel structures. The following generated
example is from one of her experiments, with the prominence left unmarked:
[4-39] The butcher hit the baker and the waiter hit him.
235
Subjects are presented with a routine prominence pattern, then with a version which
has volitional prominence on the pronoun, after which they take tests designed to
ferret out in each case which nominal the subject identifies as the pronouns referent.
Here is the routine prominence pattern for the first part when it is by itself:
[4-40] The butcher hit the baker.
Here is a typical alternative to this pattern in terms of volitionally placed prominence:
[4-41] The butcher hit the baker.
This is often treated as no different than:
[4-42] The BUTCHer hit the baker.
In other words, the butcher (and not someone else) hit the baker. This expression
is assumed to take place in the midst of a discourse context, and not out of the blue,
where in the latter case it would have an emotional meaning something along the lines
of [Shock + Surprise] The BUTCHer (of all people) hit the baker.
The routine pattern for the whole parallel structure is:
[4-43] The butcher hit the baker, then the waiter hit m.
In this sense, first the butcher and then the waiter pound on the baker. The
prominence follows a routine pattern, and so the alignment of the reference in the
parallel structures gets treated in a similarly simple, routine manner, that is to say:
236
predicate with predicate; first argument with first argument; and second argument with
second argument. Used in this routine manner, test subjects normally identify him
with baker (80% of the time as measured by Balogh). Because this is the routine
structure and pattern, this identification should be more regular from subject to subject
than if they were dealing with patterns which were more novel, simply because the
subjects experience with the conventional pattern is greater, and their responses
should more closely approximate that of a reflex.
Now let's compare the routine pattern in [4-39] to a number of its alternatives,
starting with volitional placement on the first half, and no change on the second:
[4-44] The butcher hit the baker, then the waiter hit m.
The difference in the meaning is that the butcher is now a mild change, that is to say,
a correction from an even earlier stage of the discourse. This meaning effect would be
stronger, but most of the attention is on the second half, where the prominence pattern
and the identification of the pronoun with the butcher remains the same.
This is how the expression looks when the second half has a volitional
prominence pattern, and the first half is either routine or volitional:
[4-45] The butcher hit the baker, then the waiter hit HIM.
[4-46] The butcher hit the baker, then the waiter hit HIM.
The prominence on the pronoun can be primary or intense, and the meaning will be
interpreted as if it were intense due to the great difference in level when compared to
237
the cliticized non-prominent pronoun. In each case, HIM is identified with the
butcher (60% as measured by Balogh), but the volitional prominence placement in
the first half of [4-46] again implies that the butcher is in itself a correction of an
earlier discourse stage. Again, it makes sense that the consistency of the behavior with
the novel constructions (60%) is less than with the conventional patterns (80%).
A routine prominence pattern would not disturb the expectation that a routine
argument order would align the structures for reference, but volitional prominence
suggests that there is a different alignment of some kind. The first thing that strikes L
as different, however, is not the prominence on HIM, but the lack of the prominence
on waiter. L would normally expect to come across primary prominence on the
entity in the position occupied by waiter, which would be the newish information or
rheme. The lack of prominence indicates that waiter is not the rheme, and so there is
a signal for L right there that the point of the expression is not going to be about what
happens when the waiter is treated as just another hitter who has been substituted for
the butcher. L knows that the waiter is going to be involved in something different.
A subsequent lack of prominence on hit tells L that the same action as before
is going to be performed, so the change must be to the second argument, which means
that what has changed from the first half of the parallel structure to the second half is
the baker. It is really important to notice that while this draws L to the point of not
identifying HIM with the baker, this is crucially not the same thing as having L
identify HIM with the butcher in specific. It is only the artificial narrowness of the
context which suggests that butcher would be the obvious alternative to baker.
238
If we pretend that the context of these examples is one step closer to real life,
as if they were used in a discourse context, it is easy to understand that L would
absolutely not be surprised to hear or see any of the following alternative endings:
[4-47] (a) The butcher hit the baker, then the waiter hit the BARtender.
(b) The butcher hit the baker, then the waiter hit them BOTH.
(c) The butcher hit the baker, then the waiter hit the CEILing.
(d) ? The butcher hit the baker, then the waiter hit the BAKer.
All is fine, just so long as the waiter doesnt hit the BAKer (with the exception of
some contexts which I will defer for the moment). So, while L knows that HIM does
not identify the baker, there is no necessary reason to choose butcher as the only
alternative. People should only identify the butcher 60% of the time if the pronoun is
prominent because of the novelty of the construction, especially considering the range
of alternatives available to butcher, a range which was nonexistent for the referent
baker when the pronoun was not prominent.
In this particular kind of substitution, then, the first half of the parallel structure
sets up the expectation of a routine prominence pattern, and so any deviation during
the second half signals to L a detachment of referents, rather than just a deviance in the
placement or level of one prominence. The range of potential discourse counterparts
determines what the alternative referents are for the newly prominent pronoun. People
should be expected to behave similarly to one another when it comes to the assignment
of reference when the pattern is routine and their responses more closely approximate
a reflex (80%), and there should be more deviance in their behavior when the pattern is
novel, and when the range of alternative referents grows (60%).
239
Okay, now lets get back to this example:
[4-48] ? The butcher hit the baker, then the waiter hit the BAKer.
In this case, elaboration is used to distinguish an individual BAKer in the second
half of the structure from some other baker in what must be a very close context
indeed, probably the very baker whom the butcher has just clobbered. This
revelation produces an elaboration as a side-effect, because the BAKer is implied to
be identifiable as the real baker due to its greater power or paragonal status, or it can
even be set apart specifically as the so-called baker through its periphery. It should
be kept firmly in mind that we are getting into linguistically rarefied air here, which is
one of the dangers associated with an analysis of generated examples, and so it should
be clearly understood that there are no examples in either of the corpora displaying
this detailed a level of explicit contextual differentiation.
Things have to get one step worse before they get better, so here is the
obnoxious rendering that occurs when both instances of baker are equally
volitionally prominent:
[4-49] * The butcher hit the BAKer, then the waiter hit the BAKer.
There is only one way that I can think of which makes an expression like this work,
and that would be in a context where person after person hits the same baker until it
is pure monotony, in which case the additional prominence would be emphasizing the
regularity of the rhythm in a sing-song pattern reflecting the beating itself:
240
[4-50] The butcher hit the BAKer, then the waiter hit the BAKer, then
the poodle hit the BAKer, then my mother hit the BAKer
(and so forth)
But just in its dual version, where two identical words are equal in their volitional
prominence, the sentence is rotten. If [4-47] and its alternatives are going to work, the
two objects of hitting must be unequal in some way which does not tie them to the
same referent, whether segmentally, suprasegmentally or otherwise.
This brings us to the linked prominence variation of parallel reference, which
is a very common form of [4-45], namely:
[4-51] The butcher hit the BAKer, then the waiter hit HIM.
This has the same sing-song type of pattern as the longer [4-50]. The baker gets just
that much more prominence in anticipation of maintaining a parallel rhythm, which is
the result of TIMING, as demonstrated in the next chapter (cf. 3). Suprasegmentally,
BAKer and HIM are equally intense, but even though they are segmentally
different, they still are not allowed to be taken to refer to the same thing any more than
in those rotten cases where baker was used twice instead of the pronoun.
Further analysis of these examples will be deferred until chapter 5, where they
will be compared against the examples of linked volitional prominence gathered from
the corpora. The most important point to be made until then is that while volitional
prominence is used to shake referents loose, it is the context which reattaches them,
and so it is not the volitional prominence itself which actually functions to
disambiguate reference.
241
3.5 Summary of Substitution
When S identifies a discrepancy between the portrayal of R and R
S
during
discourse, it draws Ls attention to a correction with a prominent phonological form
that can be substituted for the perceived error. L is aided in making the appropriate
exchange through apparent parallels in the older and newer contextual settings, which
cradle the word representative of Ss desired change. The change will be substituted
for its discourse counterpart, both of which are housed in similar settings. In addition,
S tends to warn L about an impending substitution with an omen. In noting a mismatch
between R and R
S
, substitution can upset the routine assignment of reference in a
context, but it does not actually reassign reference. Derivation was also shown to
exhibit similar uprooting behavior, leading to the conclusion that revelation in general
detaches reference, but does not disambiguate reference.
4 Addition
These are four examples (one with two instances) in the Brown corpus where a
null counterpart or a gap in the setting is filled by a volitionally prominent word:
[4-52] Yet adequate compensationand particularly merely adequate
compensation.
A35 1450
[4-53] Indeed, the set of endings can be replaced by the name of a set
of endings.
J32 1650
[4-54] I guess she was between affairs or something, but anyway, she
had set her sights on Johnnie, my Johnnie.
P22 0400
[4-55] I admired their easy way of doing things but I couldn't escape
an uneasiness at their way of always doing the right things.
R02
0700, R02 0700
242
There are also two cases in which part of a word fills a gap:
[4-56] If it is an honest feeling, then why should she not yield to it?
Most often, she says, its the monogamous relationship that
is dishonest.
G13 1140, G13 1140, G13 1150
[4-57] Since then, and since the pure grain had gotten him divorced
from every decent and even indecent group from Greenwich
Village to the Embarcadero, he had become a sucker-rolling
freight-jumper.
N29 0040
One of the examples that appeared earlier with [4-56] (unfunny) is a derivation by
conventional pairing, discussed below, and another (pre-attack, post-attack) is a
matter of linked prominence, and so it is discussed in chapter 5.
There is a difference between: 1) the notion of counterpart being inapplicable
in elaboration; 2) there being no explicit counterpart in a derivation; and 3) there being
an explicit, null counterpart or obviously filled gap in a repeated setting. Take [4-52]
for example. If merely were an elaboration, it would emphasize the paltriness of
merely; however, that does not account for the meaning conveyed by the repetition
of the setting _ adequate compensation. It could be argued that an elaborate form of
merely was dropped into the gap, but that gap is still evident enough that the
addition itself cannot be ignored, even if the change itself is also a case of elaboration.
Taking [4-52] once again, addition can be told from derivation because S does
not use merely to evoke a conventional or contextual counterpart like sufficiently.
The repeated setting assures that L will equate merely adequate compensation with
_ adequate compensation, portraying merely as taking the place of a previously
unspecified modifier, not an antonymic one. The rest of the examples behave similarly.
243
The consistency of the behavior across these few examples, plus the support
rendered by dozens more in the Other corpus, promotes addition as a subfunction of
revelation, as opposed to suggesting that it is merely a special variation of derivation
or substitution.
5 Derivation
Substitution provides an entirely explicit counterpart in a setting encapsulated
as a unitary word, and addition suggests that there is a null counterpart which is being
filled, but derivation provides a counterpart which is either 1) not entirely explicit in
the given context, or 2) not unitary, or 3) some overlap of 1 and 2. Derivation is most
easily thought of as having its counterparts more tenuously or broadly DISTRIBUTED
than they are in substitution. For example, the counterpart might be given explicitly in
the context, but could consist of anything from two words (which differs minimally
from substitution) to a much longer narrative. Alternately, the setting might have to be
derived with a little more effort by L based upon what it knows about the entities with
which the change normally associates. The counterpart that L derives might be able to
be summed up in one easily encapsulated notion, or it might just be too difficult to put
into one word. Cases of derivation range within these boundaries.
CONVENTIONAL material in an utterance allows for implicit counterparts to be
more easily accessed (with a range of encapsulation), and CONTEXTUAL or novel
material allows for the explicit specification of new or alternative counterparts to a
change. These counterparts can be derived through an appeal to conventional pairs
244
(5.1; [word vs. picture], or [mind vs. body]), conventional sets (5.2; numbers,
modals, or prepositions), basic conventional domains (5.3; [BODY], [TIME]), and
complex conventional domains (5.4; the guest/host or doctor/patient relationship);
likewise, there are also contextual pairs (5.5), contextual sets (5.6), basic contextual
domains (5.7), and complex contextual domains (5.8).
5.1 Conventional Pairs
Some derivational domains are conventional pairs ([me vs. you], [up vs.
down], [left foot vs. right foot]), many of which are antonymic modifiers ([big vs.
small], [smart vs. stupid]). While these pairs can be members of larger sets, such as the
pronouns and prepositions, that does not preclude their being treated as an
autonomous pair for the purposes of derivation. Their mutual exchange in cases of
explicit substitution supports their being appealed to as a pair for derivation. For
example, the familiar pair [now vs. then], which has already been shown to appear in
substitutions, is also used in derivation:
[4-58] Oh, [the cavalry charge] would be butchery all right, the
European said. We would see it that way, but it was glorious
then.
G75 0730 (cf. D16 1630)
change = then
setting We would see it that way when
counterpart = now
omen = would, but
There is no explicit counterpart in this context for then, but now comes to mind
naturally. For any L which notices it, the gap after We would see it that way _ can
245
act as an implicit, encapsulated counterpart for then, helping to support the
derivation of now. In comparison, the following example displays two explicit,
distributed counterparts:
[4-59] I'll get around to it a little later, he mumbled desperately.
Just as soon as I go to the bank, and Huh-uh. Now, Mis-ter
McBride, said Lord, and he laid a firmly restraining hand on
the field boss's arm.
N09 1360
change = Now
setting go to the bank when
counterpart = a little later + as soon as > then
omen = , Huh-uh
In other words, just because then belongs to a conventional pair with now doesnt
mean that no other counterpart will be offered, although L might still derive some
sense of now from the explicit counterparts.
Other conventional pairs include: [economy vs. profligacy]
F48 1310
; [rough
vs. fine]
E17 0530
; [exclusive(ly) vs. inclusively]
J27 0130, J43 1380
; [specifically vs.
approximately]
E27 0010
; [natural vs. unnatural]
J57 1170, L10 1280
; [prepared vs. not
prepared]
D09 0210
; [using vs. studying]
F12 1610
; [responsible vs. careless]
F15 0480
;
[allowing vs. denying]
D11 0760, D11 0770
; [because vs. that]
G02 1600
; [industrialized
vs. agrarian]
G08 0730
; [given vs. proven]
G16 1800
; [spirit vs. body]
G17 0230
;
[monogamous vs. nonmonogamous]
G13 1150
; [non-partisan vs. partisan]
G21 1580
;
[explain vs. describe]
G30 0900
; [unconscious vs. conscious]
G21 1600
; [one vs.
many]
G22 0760
; [special vs. normal]
G30 0220
; [primary vs. ancillary]
G30 1770
;
[person vs. animal]
G35 1160
; [this vs. that]
D11 0260
; [direct vs. indirect]
J50 0690
;
246
[implement vs. maintain]
F15 1000
; [force vs. yield]
F15 1010
; [formally vs.
informally]
G50 1090
; [prove vs. believe]
G29 1230
; [defend vs. attack < take the
initiative]
B23 1410
; [mind vs. body]
G70 0360
; [chance vs. guarantee]
L08 1360
;
[qualitative vs. quantitative]
J17 1180
; [linear vs. nonlinear]
J50 0730
; [regardless vs.
with regard to]
M03 0590
; [dark vs. light]
N19 0960
; [acting vs. actual]
P27 1270
;
[elected vs. appointed]
P27 1320
; [subconscious vs. conscious]
F03 1470
; [before vs.
after]
D09 0460
; [independent vs. dependent]
J59 1460
; [level vs. rate]
J41 1250
;
[transversally vs. lengthwise]
E26 1710
; [slowly vs. quickly]
E24 0920
; [head vs.
underling]
L20 0080
; [conduct vs. support]
F48 1070
; [readers vs. writers]
P10 0240
;
[ordered vs. chaotic]
G27 0500
; [immediate vs.delayed]
J24 1600
; [generalize vs.
specify]
J59 1080
; [most vs. least]
G13 1140
; [if vs. when]
L20 1380
; [choices vs. whims
< not arbitrary and whimsical]
G57 0370
; and [real vs. unreal]
D02 1530
.
Many instances of partial word prominence belong here:
[4-60] He made many tasteless, irreverent and unfunny remarks, not
only about me in the title role, but about religion in general.
R03
1550
When this sort of prominence appears in a text, the whole word tends to be italicized
or underlined, and so its spoken (and perhaps signed) frequency is probably
misrepresented. There are a number of instances in the list given above (exclusive,
non-partisan, independent, unconscious, subconscious) which effectively
display this same sort of behavior, but each of then just happens to be marked as a
whole word.
247
If a pairing is semantically conventional, but phonologically novel as a pairing
because the counterpart is lacking or distributed, then that counterpart tends to be
treated as a phonologically null form or gap, which superficially resembles addition:
[4-61] In the latter research program, information is available for
2,758 Cornell students surveyed in 1950 and for 1,571 students
surveyed in 1952. Of the latter sample, 944 persons had been
studied two years earlier; hence changes in attitudes and values
can be analyzed for identical individuals at two points in time.
G57 0100
If this were addition, then there would be an explicit mention of the _ attitudes and
values earlier in the context. There is no readily available unitary counterpart for
changes, just some not-strictly-conventional or distributed alternatives like
nonchanges, unchanged things, or even the things that didnt change. The easy
solution to the problem is just to point out that pairings dont need to be unitary,
particularly because one of the features which distinguishes derivation from
substitution is the need to accommodate distributed counterparts. L then derives novel
alternatives to changes.
The prepositions as a whole provide a common conventional set, but they are
categorized as pairings because they typically act as conventional pairs, rather than as
if any given prominent preposition were chosen in opposition to all of the other
pronouns in the set:
[4-62] A BTU is a unit of heat, and the BTU rating of a conditioner
refers to how much heat your machine can pump out of your
house in an hour.
E20 1120
248
In other words, [4-62] pairs [out vs. in(to)], and not [out vs. {on, around,
through}]. Further examples of prominent prepositional pairings are: [within vs.
between]
F37 0330, J54 0500
; [in vs. out]
J27 0720, N12 0760
; [(choose) between vs.
(choose) both]
J59 1210
; [top vs. bottom]
E24 1110
; [up vs. down]
E24 0940
;
[out(side) vs. in(side)]
F33 1520, G36 1230
; [or vs. and]
H28 0400, H28 0410
; and [without
vs. within]
F15 1530
. Some conventional pairs like this are infused by the determining
characteristics of conventional basic domains (5.3), such as when an antonymic pair
like [off vs. on] has an inherent link to the positive/negative opposition, or when the
[left vs. right] pair is understood to be just one division of dimensional space or
bilateral symmetry, but this is to be expected when meaning is distributed and
encyclopedic in nature.
One such positive/negative appeal is when and evokes and not:
[4-63] In the absence of direct evidence to the contrary, decomposition
of solvent alcohol and coordination of its fragments to the metal
were not considered, following the above heretofore-accepted
assumption in preparative coordination chemistry.
J72 1230
[4-64] I'm talking about the grand manner of the Liberal North and
South who is not affected personally.
G17 0910
[4-65] A signal cannot be cleared until all the related turnouts are
properly thrown and locked.
E07 0300
In [4-63], only the decomposition of solvent alcohol was considered, and not the
coordination of its fragments to the metal, and so deriving or as a counterpart to
and, despite their otherwise popular pairing, does not convey the right meaning.
The same thing happens in the next two examples, which balance North and South
249
against North and not South rather than North or South, likewise thrown and
locked is compared to thrown and not locked rather than thrown or locked. The
counterpart and not is two words, but the other half of a conventional pair need not
be unitary. This is discussed in more detail in the section on contextual pairs (5.5),
which often appeal to multiword counterparts.
The prominent possessive pronouns tend to provide interesting cases. The
following examples would be substitutions, but the change is possessive, and so it
really requires a possessive counterpart for pairing, which is derived in an essentially
trivial fashion from the context:
[4-66] Having (through my unflagging effort and devotion) achieved
stardom, a fortune and a world-renowned wife at an age when
most young men are casting their first vote, Letch proceeded to
neglect them all.
R03 0740
[4-67] But didn't [the New Englanders] get off too easy? The slaves
never shared in their profits, while they did share, in a very real
sense, in the profits of the slave-owners: they were fed, clothed,
doctored, and so forth; they were the beneficiaries of
responsible, paternalistic care.
G17 1270
In [4-66], S is Letchs wife, and the possessiveness of my keeps this from being a
straightforward substitution like [me vs. Letch]; instead, what happens is more
like: [my vs. Letchs < Letch]. In [4-67], you have: [their vs. slave-owners <
slave-owners]. Other examples are: [its vs. mankinds < mankind]
G22 0190
;
[my vs. Jacks < Jack]
K28 1380
; [her vs. my < I]
L24 1630
; [his vs. Phils <
Phil]
P24 1520
; [their vs. her < she]
R07 0590
; and [his vs. the Souths < the
South]
G17 0180
.
250
Finally, here is an example which could either be derivation by conventional
pair or elaboration for precision, depending upon Ss intent (or Ls interpretation):
[4-68] The obvious natural fact to ancient thinkers was the diurnal
rotation of the heavens.
G30 0050
If L takes this example to be a case of [the vs. a], then it is a derivation, and the
intended meaning is something along the lines of There were a lot of obvious natural
facts, but this was the most important one of the bunch; however, if L treats this
utterance as if it were [the vs. the], then the word-internal meaning change is an
elaboration for precision, and the meaning becomes, At the time, there was more than
one obvious natural fact that was identified by its great importance, but this one was
the most important obvious natural fact of them all. Examples of the are what got
me started in this line of research, which soon extended to cover prominent nominal
grounding predications, then the verbal grounding predications, and then all the rest.
5.2 Conventional Sets
As you might expect, conventional sets are similar to conventional pairs
(implicit counterparts), but they are less encapsulated. Here is a good set of related
examples all taken from within the same context:
[4-69] (a) [Jesus Christ] is in your hands now.
D16 1510
(b) God is in your hands, now.
D16 1610
(c) What He does with you then depends on what you do with Him
now.
D16 1620, D16 1630
(d) Then it will be a fearful thing to fall into the hands of the living
God if you have abused Him in your hands.
D16 1630
251
There are a lot of nows and thens in this example, but they are not all used the
same way. In (a), when now is used for the first time, its counterpart is not then,
but rather it is any time in the conventional set of temporal adverbials other than now
(e.g. Dont worry about what went on last year, and dont worry about whats going to
happen tomorrow, and in fact, dont worry about whats going on at any other time,
because [Jesus Christ] is in your hands now).
The use of now in (b) could be called recapitulation, but for the reasons
explained earlier (3.2), it makes more sense just to identify it as the second of two
close derivations. Example (c) is a case of DISSOCIATION, which is also analyzed in
more detail in the next chapter (1). In brief, dissociation emphasizes what the two
volitionally prominent words do not mean in common, which strengthens the borders
which divide their meanings, after which they can be more easily interpreted as
taxonomic labels which divide up the context between them, which in this case is
present and future time.
By the time that then comes along in (d), its counterpart is not the whole
conventional set of temporal adverbials (note that in this context, then is not a
substitute for at any other time), neither is it the other half of the conventional pair to
which it belongs with now, but rather this then is an emotionally charged repetition
of the earlier then. In addition to this emotion, the earlier then is also volitionally
prominent, and so rather than finally identifying a case of recapitulation, this is an
example of COORDINATION (chapter 5, 2).
252
Here are three more examples of prominent adverbials of time:
[4-70] Noting such evidence is the first step; and almost the only
cure is early detection and removal.
G22 0070
[4-71] And if he surrendered after raving at her.
P28 0660
[vs. without
having raved at all]
[4-72] For example, the marksman gets 5 shots, but we take his score
to be the number of shots before his first bull's-eye, that is, 0, 1,
2, 3, 4 (or 5, if he gets no bull's-eye).
J19 1410
Each of these prominent adverbs has a counterpart of no time at all, which suggests
that no time is a member of this conventional set, just as null is a member of every
set. As shown in the previous chapter, the negative adverbial of time never always
behaves as if it were elaborated for precision.
The cardinal and ordinal numbers are special conventional sets, namely
continua:
[4-73] Some experiments are composed of repetitions of independent
trials, each with two possible outcomes.
J19 0010
[4-74] Even granted that the Congo should be unified, you don't
protect Western security by first removing the pro-Western
weight from the power equilibrium.
B23 1480
[4-75] The binomial probability distribution may describe the
variation that occurs from one set of trials of such a binomial
experiment to another.
J19 0040
There are some numbers which often act as pairs ([one vs. two], [none/zero vs. one],
[third vs. second]), but the data suggest that they are usually substituted for one
another in the less-than-ten range ([one vs. four], [six vs. nine]), or substituted by
factors of ten ([eighteen vs. eighteen-hundred], [five vs. five-thousand]).
253
When quantifiers (and their related indefinite pronouns) are not being
elaborated for precision, they are used for derivation by conventional set. To begin
with, there is [some vs. none]:
[4-76] But the practice is likely to be misleading, since it may seem to
support a conclusion that, as long as the revenues from any
class of service cover the imputed operating expenses plus
some return on capital investment, however low, the rates of
charge for this service are compensatory.
J50 1260
[4-77] One thing we haven't discussed, expense money. We'll need
some at least, if only bus fare to the scene of the crime.
L24 1490
[4-78] You can get something, Nadine would snap.
P18 1630

(blackballed)
Others examples of [some vs. none] are: [some new homes]
E20 0080
; [some of
us]
M01 0930
; and [somebody vs. nobody]
B21 0670
.
Here are some examples of [all vs. some]:
[4-79] Eichmann himself is a model of how the myth of the enemy-
Jew can be used to transform the ordinary man of present-day
society into a menace to all his neighbors.
F14 1470
[4-80] The aim was to state the results of all available determinations
of atomic positions in crystals.
J73 1120
and [all vs. most]:
[4-81] Molotov, in particular, is being charged with all kinds of sins
especially with wanting to cut down free public services, to
increase rents and fares; in fact, with having been against all the
more popular features of the Khrushchev welfare state.
B25
0640
[4-82] contraception was condemned by all Christian churches as
immoral, unnatural and contrary to divine law.
F15 1220
254
Others are: [all cultures vs. most]
G22 0330
; and [all nations vs. most]
G72 1160
. Some
of these are similar to the cases of quantifiers elaborated for precision, and can be
interpreted that way by L, but when compared to the behavior of never and none,
these examples stand out more clearly as derivation. Again, it is a matter of whether
the meaning change is taken to be word-internal, or contextual.
Here are some examples of intensely prominent, individual relative quantifiers,
which were first introduced in chapter 3, 2. As you recall, the proportional relative
quantifiers were all prone to prominence for precision, but each of the individual
relative quantifiers was described as having a particular functional aspect emphasized
by prominence which differentiated it from the others in its set:
[4-83] Moreover, the cost of operations is always high in any new
store, as the conservative bankers who act as controllers for
retail giants are beginning to discover.
J60 1210
[4-84] The DRDW statement may also be used to generate an RDW
defining any area specified by the programmer.
J69 1370
[4-85] It is interesting that it is not the getting of any sort of knowledge
that God has forbidden, but, specifically, the knowledge of the
difference between good and evilthat is, abstract and moral
judgments, which, if they reside anywhere, reside in the
neocortex.
17.98
[4-86] The performer just gets the assistant to lie and say that he saw
the performer seal the prediction days before. There are
illusions as infuriatingly misleading as this. Every illusion is
misleading, somehow.
11.192
The related indefinite pronouns work the same way:
[4-87] Out cold, if not dead; and he'd never known what hit him- he'd
never known that anything had hit him.
L24 0610
255
[4-88] You know it and Ill tell everybody exactly how it happened.
P03
0240
[4-89] Have a party an leave em out, hon, he suggested. A swell
party, send an invite to everbody but them.
P03 1540
[4-90] Ekstrohm nudged it with a boot. Hey, this is pretty close to a
wart-hog. Uh-huh, Ryan admitted. One of the best
matches I've ever found. Well, it has to happen. Statistical
average and all. Still, it sometimes gives you a creepy feeling to
find a rabbit or a snapping turtle on some strange world. It
makes you wonder if this exploration business isn't all some big
joke, and somebody has been everywhere before you even
started.
M04 0650
The two examples of any from the Brown corpus, namely [4-83] and [4-84], both
emphasize the random nature of an instances selection: no matter which new store or
area gets chosen, it will have some amount of the desired quality, which would be a
high cost of operation or susceptibility to being specified, respectively. The any
example in [4-85] works the same way, posing the getting of a randomly selected sort
of knowledge against the getting of one deliberately chosen, specific sort of
knowledge.
The only example of every is also from the Other corpus, where in [4-86] Ls
attention is drawn to the fact that it is not just the specific illusion currently under
scrutiny which is misleading, but rather that no matter which trick had been chosen as
a representative, it would also have been misleading in some way. Unfortunately, there
were no examples in either corpus of each, and so for now I am only speculating (on
the basis of imagined examples) that the difference between each and every would
be the emphasis on the type of scanning involved.
256
There is another set which I am treating as conventional simply because a lot
of instances in the data appeal to it, and that set is [want vs. need vs. like vs. tolerate
(vs. their negative oppositions)]:
[4-91] [Some men] like to be dominated.
F08 1020
[4-92] Obviously, such a Northern tourists purpose (in ignoring
Southern industrial reality and hallucinating Southern romantic
fantasy) is somewhat akin to a childs experience with
Disneyland: he wants to see a world of make-believe.
G08 1510
[4-93] I sometimes feel that God, in His infinite wisdom, wants us to
have these inexplicable little lapses of memory.
R03 1130
This conventional set, just like any other, encompasses a range of appropriate
alternatives to the change, which is itself a set member.
Similarly, there is a set which is made up of the union of two other sets, namely
the [smells vs. feels vs. looks (and so forth)] set with [looks vs. seems vs. is]:
[4-94] The aborigine lives on the cruelest land I have ever seen. Which
does not mean that it is ugly. Part of it is, of course. There are
thousands of square miles of salt pan which are hideous. But
much of the land which the aborigine wanders looks as if it
should be hospitable. It is softened by the saltbush and the
bluebush, has a peaceful quality, the hills roll softly.
G04 0450
[4-95] Spatiality becomes part of the tactual sensation only by way of
visual representations; that is, there is, in the true sense, only a
visual space.
J53 0210
[4-96] Charlie grinned. She didn't sound like a pale girl.
P23 1360
The values for the senses and appearances can all act as appropriate counterparts for
one another, such as [smells vs. is], or [feels vs. seems].
257
Finally, there is this last set which deals with likelihood (epistemics):
[4-97] By contrast, a good deal of nuclear pacifism begins with the
contingencies and the probabilities, and not with the moral
nature of the action to be done; and by deriving legitimate
decision backward from whatever may conceivably or possibly
or probably result, whether by anyones doing or by accident, it
finds itself driven to inaction.
D11 0580
[4-98] surely anything is better than a policy which allows for the
possibility of nuclear war.
D11 0830
These examples both came from the same tract on nuclear war, but as with the earlier
conventional sets, this set is backed up by a significant data from the Other corpus.
5.3 Basic Conventional Domains
Examples from two basic conventional domains are categorized here, where
the domain of SIZE demonstrates that basic can be simple, and where the POSITIVE/
NEGATIVE OPPOSITION shows that basic can be complex, albeit systematic. Examples
of another domain (WHOLE/PART) which resemble PROMINENT ADMISSION (a type of
positive/negative opposition) are included at the end of this section not just because
they illustrate a common type of construction, but precisely because some of the
positive/negative opposition examples behave similarly to the whole/part instances,
and the two types of revelation need to be differentiated.
To begin with, there is a natural gradation from conventional sets to basic
domains which involves not just an increase in the number of members, but an
increase in the density of the interconnections among members. There is no strict
boundary between a large set and a small domain. For example, one common domain
258
is composed of concepts related to size, and while some of the members also take part
in conventional pairings ([big vs. small]) and sets ([smallest vs. smaller vs. small vs.
big vs. bigger vs. biggest]), size as a basic conventional domain encompasses a greater
number of members than does a typical conventional set like the prepositions, and the
interconnection between its members has a greater degree of freedom ([ungainly vs.
minuscule] could work) than is found in some sets ([five vs. 7,987,754] rare).
Here are examples evoking the size domain from the Brown corpus:
[4-99] I think that readers generally hate minute polemics and
recriminations.
C05 0540
[4-100] From maturity one looks back at the succession of years, counts
them and makes them many, yet cannot feel length in the
number, however large.
G14 0310

The derivations are [minute vs. regular-sized] and [length vs. quantity <
succession]. The vast scope of the domain allows for a great deal of freedom in
choosing a counterpart, but conventionality places enough limits on that freedom to
make the task possible.
This is where the conventional positive/negative opposition comes in. I am
including some examples of substitution and addition here as well because their
behavior is all based on this same domain, and because most of the examples by far are
derivations. There are three main types of behavior common to this domain, namely:
1) PROMINENT NEGATION, in which a positive value of a variable in the setting is refuted
by not; 2) PROMINENT CONFIRMATION, in which both the setting assertion and the
change are positive; and 3) PROMINENT ADMISSION, in which the setting variable has a
259
negative value, but the change is positive. There are no cases in the Brown corpus in
which a negative setting is associated with a negative change, although this behavior
occurs in speech.
Prominent negation is specifically contextual or external to the phrase in which
it resides, refuting an assertion made earlier in the discourse (cf. Langacker, 1991:
132-141). Here are examples in which there is a positive setting and a negative change
(prominent negation):
[4-101] All is not sex, declared Lawrence.
G13 0100
[4-102] Suppose it was not us that killed these aliens.
M04 0900
[4-103] For the only time in the opera, words are not set according to
their natural inflection; to do so would have spoiled the
dramatic point of the scene.
J64 0690
[4-104] And women were not expected to know that the pitcher was
trying not to let the batter hit the ball.
F38 1700
[4-105] the truth of the matter is that most American Catholic
colleges do not owe their existence to general Catholic support
but rather to the initiative, resourcefulness and sacrifices of
individual religious communities.
A35 0340

[4-106] But the main point here is that even if such a restatement were
not possible, the demand to demythologize the kerygma would
still be unavoidable.
D02 0890
[4-107] And, as we know, the Virgin Lands are not producing as much
as Khrushchev had hoped.
B25 0580
Without prominence, Lawrence is simply declaring what might be a completely new
discovery, but with prominent negation, he is specifically denying an earlier contention
to the contrary. That contention might be conventional or contextual. Likewise, the
volitional prominence in [4-102] indicates that the guilt of the astronauts has already
260
been suggested, where without that prominence (Suppose it wasnt us that killed these
liens) the statement could be the first mention of that possibility after the aliens were
found dead. In each of these cases, the setting amounts to a positive assertion which is
denied by the change. In [4-103], words are always (positive) set according to their
natural inflection, except not (negation) at one dramatic point. The rest of these
examples behave similarly.
For not, the setting is always positive in the Brown corpus, but for the
volitionally prominent auxiliary verbs, the setting can be negative or positive. Because
the volitionally prominent forms of these utterances always have a positive change,
sometimes the setting and the change are equal in polarity, and sometimes not. I find it
interesting that the distributions of the examples across equal and unequal polarities
are similar. There are 31 examples in which prominence confirms a positive setting
(equal: 16 BE, 8 DO, 2 HAVE, 5 modals) and 27 where prominence admits a positive
outlook despite a negative setting (unequal: 13 BE, 6 DO, 3 HAVE, 5 modals). I dont
know if the distribution among routine utterances is as balanced.
The prominent confirmation (positive-positive) examples look like this:
[4-108] The pretty little twittering WACS said he had the look of
eagles and Penny, hating the cliche, had to admit that in this
case it applied. Keith was an eagle.
N23 1130
[4-109] THE MOST surprising thing about the Twenty-second
Congress of the Soviet Communist Party is that it was
surprising perhaps quite as much, in its own way, as the
Twentieth Congress of 1956, which ended with that famous
secret report on Stalin.
B25 0020

[4-110] This, he was sure, was the way they would act; laughing at a
dying man, laughing as a man was beaten to death.
N09 1680
261
[4-111] That unused room was large enough for well, say an elephant
could get into it and, as a matter of fact, an elephant did.
G40 0070
[4-112] The commander has failed in his duty if he has not won victory
for that is his duty.
C03 1120, C03 1130
The assertion in the setting is positive, and it is either unitary (had, surprising,
could), or distributed (descriptions of actions or duty in background), in addition to
which the change is also positive (was, would, did, is).
Finally, the examples of prominent admission (negative-positive) look like
this:
[4-113] I'm not saying you're yellow. I am saying you're not a
professional ballplayer.
P24 1270
[4-114] Naturally, the patient does not say, I hate my father, or
Sibling rivalry is what bugs me. What he does do is give
himself away by communicating information over and above
the words involved.
F01 1701
[4-115] Dave has qualities of leadership.
P27 1730
[4-116] You can take it with you.
E14 0700
[4-117] But, as Scripture everywhere reminds us, God does have need
of his creatures, and the church, a fortiori, can ill afford to do
without the talents with which the world, by God's providence,
presents it.
D02 0760
These settings are all negative (not (saying), (does) not, has voiced with doubt
in an earlier sentence, cant, a lack of need suggested earlier in context), but they are
still either unitary (not) or distributed (ICM: You cant take it with you).
Prominence draws Ls attention to the admission of a positive outlook in the face of
these negative settings.
262
The interpretation is straightforward in all of these examples: there is a
contention made at an earlier stage of the discourse, which when positive is either
denied (negation) or supported (confirmation), and which when negative is countered
in a positive light (admission). Once again, ICMs serve to make background assertions
just as well as the discourse context.
The following examples superficially resemble prominent confirmation, but
they actually appeal to a basic conventional domain defining whole/part relationships:
[4-118] With no strong men and no parliament to dispute his will, he
was the government.
B26 1690

[4-119] Helva was unconditionally graduated and installed in her
ship, the XH-834. When she awoke, she was the ship.
M05 1580
[4-120] It implies two misconceptions. One is that whatever is
ecumenical has to do with some over-all organization at the
top and needs only to be understood at the so-called lower
levels. The truth, however, is that the ecumenical church is just
the local church in its own true character as an integral unit of
the whole People of God throughout the world.
F37 0430
In paraphrase, [4-118] is really saying, At first, he was in the government, but then he
was the government. Thats actually quite a bit different than if this were a prominent
admission, which would be more like, At first, he was not the government, but then
he was the government. Similarly, [4-119] works something like [was _ (the ship) =
WHOLE vs. was in (the ship) = PART]. In [4-120], rather than saying that the
ecumenical church is over the local church (and so is a distinct part of the whole
governing structure of the church), it is the local church (and so they are whole),
which can be represented as [is _ = WHOLE vs. is over < (top, lower) = PART].
263
[4-118] could be a derivation by conventional pair if the verb to be were
interpreted as one half of a typical pairing with the verb-particle construction to be
in. An extension of this same judgment would also allow [4-119] to be treated as a
derivation by conventional pair, namely [was vs. was in < installed in]. An
example like [4-120] would simply be [is _ vs. is over < (top, lower)]. A more
generous lumping together of verbs and verb-particle variations as counterparts would
even allow [4-119] to be categorized as a simple substitution, namely [was vs.
installed in].
5.4 Complex Conventional Domains
Because complex conventional domains are composed of extensive, well
familiar details, people notice quickly when something is mismatched, and they treat
that error specifically as if it were odd (more subjectively or emotionally disturbing)
rather than just wrong (more objectively or rationally disrupted). This feeling of
strangeness means that examples of this type are often interpreted as similar in
meaning to peripheral elaborations. Here is an example which wreaks havoc with the
complex conventional domain that defines the formal relationship holding between a
doctor and a new patient, where the patient is speaking:
[4-121] Don't give me a lot of talk, Joe.
P19 1480
The doctors expectations are disturbed by the patients use of a familiar form of
address, on top of his being bewildered by the fact that a total stranger knows his first
name. It turns out that the patient is his long-forgotten, barely recognizable, formerly
264
alcoholic, ex-wife. She is using the doctors first name deliberately to shock him, and
it works. The counterpart is Doctor X as prescribed by convention, and so this is a
form of revelation, but this use of his first name is meant to strike the doctor as weird,
and the meaning of Joe has changed Joe is far more ominous than usual. This is
why cases similar to this one can come to be classified as elaboration for periphery.
This simply shows that revelation and elaboration are not mutually exclusive
behaviors, but the use of volitional prominence for one can have side-effects like the
other, and only context determines which is the real use in any given case, depending
upon Ss intent or Ls interpretation.
5.5 Contextual Pairs
Derivation by contextual pairing comes as close to substitution as you can get
without actually being substitution. For example, had the adjective revolutionary
been used in the following passage rather than the noun revolution, then the
adjective simpler could have substituted for it directly rather than derived from it
less so:
[4-122] Historical records indicate that Copernicus was unaware of the
fundamental aspects of his so-called 'revolution', unaware
perhaps of its historical importance, he rested content with
having produced a simpler scheme for prediction.
G30 1190
As it is, the counterpart for simpler has to be derived from the nominal form of the
counterpart, namely revolution. Similarly, the verb juxtaposed was used below
instead of an adjective like dependent:
265
[4-123] Large planes juxtaposed with other large planes tend to assert
themselves as independent shapes, and to the extent that they
are flat, they also assert themselves as silhouettes; and
independent silhouettes are apt to coincide with the
recognizable contours of the subject from which a picture starts
(if it does start from a subject).
J59 1460
The counterpart has to be derived from an adjectival synonym for juxtaposed that
makes a contextual pair with independent. This is also influenced by the complex
contextual domain described later.
The difference between this type of derivation and substitution is not just
pickiness. There is an important difference between pairs which are conventional
enough to share grammatical class, making them easily accessible by L even when
implicit, and those which are contextual, and which therefore must be made explicit
while L is paying attention. But take a look at this next example, which would be a
substitution were it not for the counterpart being two words:
[4-124] They are two sides of the same coin and the South will not
change cannot change until the North changes.
F42 1190
This [cannot vs. willnot < will not] would be [can
not
vs. will
not
] were it not for
a spelling convention. There is no contextual derivation, so this is really substitution.
But there are many cases in which the other half of the contextual pair is a
multiword counterpart which can only come directly from the context:
[4-125] The assisting musicians from the Vienna Octet are somewhat
lacking in expertise, but their contribution is rustic and
appealing.
E04 0660
266
[4-126] Had Churchill been returned to office in 1945, it is just possible
that Britain, instead of standing fearfully aloof, would have led
Europe toward union.
G72 0420
Neither of the changes elicit the other half of a conventional pair which will lead L to
the right meaning ([led vs. followed] no); instead, they rely upon contextually
defined counterparts. Its not just that the counterparts are distributed (rustic and
appealing, standing fearfully aloof), but longer constructions are more likely to be
novel than shorter ones, and so are less likely to be conventional. Other examples are:
[processes vs. formula]
J27 1320
; [one vs. long and diverse tradition]
D02 1240
;
[language vs. mode of action]
F48 0790
; [(make) suable vs. not to cause
constitutional problems]
J43 1000
; and [defend vs. take the initiative]
B23 1410
.
5.6 Contextual Sets
A contextual set amounts to one prominent word contrasted with at least two
separate (possibly multiword) counterparts:
[4-127] Papa was disappointed that none of the brothers had heard the
Call. Not George, Townley, or Ted, certainly not Ludie. Burt
was at Hackettstown and Will at Albany Law School, where
they surely could not hear it. Someday God would choose him.
K06 1260
[4-128] After all, the money dispensed by the state goes not to the
farmer, the laborer, or the businessman, but to foreigners.
F46 0240
In [4-127], the counterpart set is {George, Townley, Ted, Ludie, Burt,
Will}, and in [4-128] it is {farmer, laborer, businessman}]. The counterparts
are multiword in [causes vs. {causes directly, (causes) indirectly}]
D11 0160
, as
267
well as in [done vs. {deliberately willed or intended, not intended or not directly
intended}]
D11 0180
.
5.7 Basic Contextual Domains
These are contextual sets with some complexity or systematicity, which leaves
them prone to being treated as basic conventional domains, or as complex contextual
domains. For example, a word such as newspaper can provide a satisfactory basis
for such a domain. People know about a newspaper and its parts, but it is arguable
whether it gets used often enough to warrant the status of a conventional set:
[4-129] Since brevity is the soul of ambiguity as well as wit, newspaper
headlines continually provide us with amusing samples.
R05 0700
This provides the derivation [headlines vs. {set of other parts of a newspaper}],
which is categorized as contextual simply because 1) the domain seems to lack the
familiarity of other conventional domains and 2) the passage from which this example
was taken provides an explicit contextual definition for this set, and so conventionality
is either superfluous or superseded.
5.8 Complex Contextual Domains
Longer contexts have the room to define plenty of specific terms, and to
narrow down the range over which L is likely to select the counterparts for changes. In
the Brown corpus, there are two extended contexts which take advantage of this
latitude. In the first, the Ptolemaic description of the universe is contrasted with the
systematic Copernican explanation giving sense to that description:
268
[4-130] Let us re-examine the publicized contrasts between Ptolemaic
and Copernican astronomy. Bluntly, there never was a
Ptolemaic system of astronomy. Copernicus achievement was
to have invented systematic astronomy.
G30 0640, G30 0650
To begin with, system is used as an elaboration for negative position, meaning
system in the strictest use of the word, and then invented is a revelation whose
counterpart is derived to be a type of initiation which is less prestigious, like
developed or borrowed. So, while Ptolemy merely recorded a staggering number of
astronomical observations, Copernicus actually discovered that those observations
could be organized into a system of astronomy. The elaboration of system is then
repeated:
[4-131] But none of this has prevented scientists, philosophers, and
even historians of science, from speaking of the Ptolemaic
system, in contrast to the Copernican.
G30 0860
This determines the interpretation of prominent forms in the rest of the context, such
that Ptolemy is nothing more than an aggrandized clerk, while Copernicus unveils a
system of interrelated explanations. Take the next prominent word as an example:
[4-132] Ptolemy recurrently denies that he could ever explain planetary
motion. This is what necessitates the nonsystematic character of
his astronomy.
G30 0900
Whereas explain might be used in other contexts in a conventional pairing with
confuse or obfuscate, or in some form of elaboration, this contextual domain has it
defined to contrast with a poorly connoted form of describe. Likewise:
269
[4-133] It is the chief merit in Copernicus work that all his planetary
calculations are interdependent.
G30 0970
[4-134] In a systematic astronomy, like that of Copernicus,
retrogradations become part of the conceptual structure of the
system; they are no longer a puzzling aspect of intricately
variable, local planetary motions.
G30 1040
In this context, systematic is not derivatively countered with something like
chaotic or asystematic, but it specifically acts in partial synonymy (PARANYMY)
with the term interdependent, and both are run counter to notions like independent.
The potential counterparts chaotic and asystematic are not evoked because they are
not members of the complex contextual domain that S has been defining.
The other extended context involves a reviewer trying to support a
differentiation between literal and figurative depth, spilling over into the contextual
definition of related concepts:
[4-135] By its greater corporeal presence and its greater extraneousness,
the affixed paper or cloth serves for a seeming moment to push
everything else into a more vivid idea of depth than the
simulated printing or simulated textures had ever done.
J59 0410
The initial difference between real and illusory depth on a two-dimensional field
comes into play as a contextual set [idea vs. actual < texture]. This contextual
definition helps L to avoid appealing to a conventional pair like [idea vs. action], or
from floundering around trying to come up with some other implicit counterpart.
In the next example, paper applied to the canvas is noted to make more of an
impression in terms of its flatness than its additional thickness on top of the canvas,
270
thus this application of the paper has the opposite effect of that which was intended,
because it supports the 2-d rather than the 3-d perspective:
[4-136] Because of the size of the areas it covers, the pasted paper
establishes undepicted flatness bodily, as more than an indication
or sign. Literal flatness now tends to assert itself as the main
event of the picture, and the device boomerangs: the illusion of
depth is rendered even more precarious than before.
J59 0500
Bodily is not used as part of the conventional pair [mind vs. body] as was in earlier
examples, but rather as part of the growing contextual domain [(body = literal =
flatness) vs. (illusion = figurative = thickness)]. Contextual or novel descriptions, then,
can be used to redefine or supersede conventional associations between concepts.
5.9 Derivation Summary
The counterparts for substitution are not only unitary and encapsulated, but
they are cradled in a setting so similar to that which houses the prominent change that
they are just that much more easily identified. Derivation is the process of identifying
counterparts which are more obscure, either because they are not tightly encapsulated,
or because they are not entirely explicit. Derivation provides L with a number of
strategies to ferret out the distributed counterpart that S has in mind, or, alternately,
derivation provides S with a number of strategies to make the identification of an
obscure counterpart clearer for L.
To begin with, a counterpart might be entirely implicit, but conventional
associations of the prominent change with a counterpart can make its identification
significantly easier, and so such relationships are relied upon heavily during
271
conversation. These conventional links run from simple pairings, to small sets, to
simple networks, to complex systematic domains. When no conventional ligature is
available, one or more can be provided by the context, where such contextual
definitions can become quite involved, providing cues to the identification of
counterparts for several prominent changes over the course of the conversation.
6 Conclusion
Rather than reiterating the definitions of the subtypes of revelation one more
time, or mentioning that the data can be exhaustively categorized according to the
explicitness and encapsulation of the counterparts, I would like to discuss one
conclusion that I was not able to support substantially in this chapter, given that I
chose to analyze written rather than audio or video recorded data. I would like to have
been able to show that the more strongly that S asserts the correction to the mismatch,
or the more deeply that the mismatch invades the archive, the stronger the intensity of
the volitional prominence. A fine gradation is hard to support firmly with written data.
Much of the material in chapter 2 suggests that such a gradation exists, as does
other work linking intonation patterns with memory depth, but the written recorded
data analyzed here simply cant be measured to back that up. That is not as much of a
problem when it comes to the examples of elaboration, because it seems to be easier
for native users of the language to internalize the written representation of those
examples into a feeling for how intense the volitional prominence is supposed to be,
and how its form would change if the meaning were made even more intense.
272
CHAPTER 5
Linked Instances of Volitional Prominence
ANTIPHOLUS OF EPHESUS twin brothers and sons to
ANTIPHOLUS OF SYRACUSE Aegion and Aemelia
DROMIO OF EPHESUS twin brothers, and attendants on
DROMIO OF SYRACUSE the two Antipholuses
ANTIPHOLUS OF SYRACUSE
This purse of ducats I receivd from you,
And Dromio my man did bring them me.
I see we still did meet each others man,
And I was taen for him, and he for me,
And thereupon these errors are arose.
William Shakespeare, Comedy of Errors
When instances of volitional prominence appear together in close context, their
proximity is often just coincidence as CLOSE INSTANCES OF VOLITIONAL PROMINENCE, but
it is so natural to interpret shared elevation as meaningful that any relations already
holding between the words will likely be treated as significantly LINKED INSTANCES OF
VOLITIONAL PROMINENCE. Members of conventional pairs or sets, for example, can be
subjected to taxonomic DISSOCIATION (1), and consequential or causal links between
words can be activated by their COORDINATION (2). Both of these effects can be
augmented by rhythmic TIMING (3), the strength of which is proportional to how
parallel the structures are that hold the linked prominent instances. So, volitional
prominence itself only lifts the instances clear of the baseline, and the significance of
this upheaval is drawn from existing links between the words meanings.
273
1 Dissociation
Dissociation occurs in a context which drives linked prominent words apart,
causing them to function as taxonomic labels which identify two (and rarely more)
exclusive sections of a domain. These instances border on being disallowed as data
because they approach the behavior of words being made prominent for definition as
technical terms; however, these cases are saved by the difference between 1) using
prominence to draw out a special or technical meaning of a word, and 2) the familiar
treatment of prominence as if it necessarily implied that the word serves a special
purpose with its normal meaning intact. Obscuring this difference makes dissociation
seem to be shaded by elaboration. Here are some examples of linked volitionally
prominent members of a conventional pair, a conventional set (pronouns), a basic
domain (positive/negative opposition) and a complex domain (Christianity),
respectively:
[5-1] we have insufficient specialists of the kind who can compete
with the Germans or Swiss. we have not enough generalists
who can see the over-all picture.
G20 1600, G20 1660
[5-2] The whole act is tailored to her pleasure, and not to theirs.
F08
0160
[5-3] The word that is not used can be as important as the word that is
used.
F01 1800
[5-4] Notice that this man had a threefold conception of God which is
the secret of his faith. First, the Lord is my light. This is the
faith that moved the psalmist to add his second conception of
God: The Lord is my salvation. And so the psalmist
gives us one more picture of God: The Lord is the strength of
my life.
D07 1220, D07 1390, D07 1600
274
These examples are allowed because they each reflect an analog of isolated volitional
prominence in an American English declarative sentence. Notice the primacy of pairs
in the following list of additional examples: [whether vs. how much]
G02 0600
(left
in because whether is perfectly good data); [what vs. how]
G43 0110, G43 0130
;
[normal vs. excessive]
F11 0470
; [I vs. he]
R03 0810
; [for vs. against]
A36 1370
;
[local vs. federal]
G08 0930, G08 0940
; [down vs. up]
E24 1120, E24 1130; G20 1430, G20 1450
;
[then vs. now]
D16 1620, D16 1630
; [gradual vs. abrupt]
G08 0840, G08 0900
; [means
vs. motive]
F15 1570
; [national vs. local]
G02 0820
; [beyond vs. behind]
D02 1380
;
[aircraft vs. airfield]
E03 1010
; [pre- vs. post-]
E03 1620, E03 1630
; [accept vs.
repel]
F07 0770, F07 0780
; [left vs. upon vs. right]
J13 0230, J13 0230, J13 0240
; and
[necessary vs. (not) sufficient vs. (is) sufficient]
F44 0050, F44 0070, F44 0090
. Those
few examples affected by timing are listed later in this chapter.
This unusual example gives each half of a taxonomic division two labels:
[5-5] the only place left for a three-dimensional illusion is in front
of, upon, the surface. In their very first collages, Braque and
Picasso draw or paint over and on the affixed paper or cloth.
J59 0570, J59 0580
Earlier (cf. 5.8, p. 267), a passage was analyzed in which literal and figurative depth
were opposed to one another, and this example is from that same passage. Both front
and upon are used as labels applicable to the domain of literal depth, while over
and on refer to figurative depth, where surfaces which are drawn or painted on only
give the illusion of depth. This example does not really represent four linked
prominences so much as two instances of two linked prominences each.
275
Now here are some examples of novel divisions being defined in context, ones
which are allowed as data because they also reflect volitional prominence:
[5-6] One such disagreement, which will receive attention in this next
chapter, concerns the question whether rates for different kinds
of service, in order to avoid the attribute of discrimination, must
be made directly proportional to marginal costs, or whether they
should be based instead on differences in marginal costs.
J50 0100,
J50 0110
[5-7] Now good definition is one thing that all of us can acquire with
occasional high-set, high-rep, light-weight workouts. But
contest definition that dramatic muscular separation of every
muscle group that seems as though it must have been carved by
a sculptor's chisel is something quite different.
E01 1170, E01 1190
Notice that these terms do not form pairs conventionally, but rather contextually. This
is natural behavior for definitive sources like Principles of Public Utility Rates and Mr.
America magazine.
This is a similar case which was disallowed as data:
[5-8] * If we look about the world today, we can see clearly that there
are two especially significant factors shaping the future of our
civilization: science and religion. *
D13 0030
This example does appeal to a conventional pair, [science vs. religion], but in this
particular example, the members of that pair are only marked mechanically to identify
them as labels, and not to change the intensity of their prominence, disallowing this
example as an instance of data.
This delicate care taken with the rules brings up the question of whether an
allowance might have been made for some instances of individual volitional
276
prominence which were ruled out due to their mechanical nature. The fact of the
matter is that these same issues did come up during the sifting of the data, and
instances were ruled out only when they seemed entirely mechanical, such as when
they followed identifying phrases that seemed to put the prominent material in quotes,
as in the word death, or the letter F. Such instances were left out because they
did not reliably reflect spoken volitional prominence.
Just as taxonomy can be used to divide up a domain, there are examples where
prominence is used to change or switch just such a previously defined taxonomy:
[5-9] the sound coming through the walls like something on the
other side of the curtain, so you knew they heard you when they
were quiet.
P09 0230
[5-10] The most unbelievable thing

about the chance meeting was that
he seemed interested in me, too.
P22 0370
Such examples are entirely dependent upon the context already having established a
taxonomy to switch, just as the taxonomic cases are entirely dependent upon how
strong the pre-existing relations are between the prominent words. For members of
conventional sets, this link is familiar enough to make it difficult to come up with a
continuous context in which their linked prominence would be entirely coincidental,
but contextual domain members are often only schematically bonded outside of that
specific context, and so their coincidental prominence is less unlikely (I dont have
the strength to find a light, cf. [5-4] outside of Christianity). Linked prominence
makes it easier or possible for the words to be related at all, but neither dissociation
nor taxonomy nor switching are actual functions of linked volitional prominence itself.
277
2 Coordination
COORDINATION activates or registers as significant any causal or consequential
conduit which familiarly links two words. Prominence does not actually function to
create new links, and so such a tie must have existed previously, when the words were
not prominent, but even a tenuous tie can be given strength in context. Using linked
volitional prominence triggers the response that there should be a link to find between
the words, and so one will tend to be found, even if it requires a stretch:
[5-11] And if I ever hear you say Mist Laban again Ill scream.
P03 0650
[5-12] But I have been blest with excellent spirits, and to-day have
been running about the deck, and dancing in our room for
exercise, as well as ever.
G37 0500, G37 0500
[5-13] It is because each side has sought to implement its distinctive
theological belief through legislation and thus indirectly force
its belief upon others.
F15 1000
[5-14] It will be shown that the objectives of the cooperative people in
an organization determine the type of network required.
G20
0010
[5-15] Then it will be a fearful thing to fall into the hands of the living
God if you have abused Him in your hands.
D16 1630
[5-16] One must first detect a fleeting mobile or moving target, decide
that it is worthy of destruction, select the missile to be fired
against the target, compute ballistics for the flight, and prepare
the missile for firing.
E03 0430 - E03 0450
In [5-11], the charge applied to ever is as absolutely extreme as the power of the
scream itself, where this scream is also linked along the entire temporal expanse
of ever. In [5-12] through [5-14], respectively, dancing is for exercise,
implementation causes forcing, and the type of network is a consequence of the
278
objectives. As you no doubt recall, the then in [5-15] is an emotionally charged
repetition of an earlier then (cf. 5.2, p. 250). Most of these are only examples of
dual coordination, but regular conversations sometimes use longer strings, and [5-16]
is the longest example of coordination in the Brown corpus.
The sense of coordination comes through because: 1) the meanings of the
prominent dual entities are not strongly in dissociative opposition; 2) the omens if,
for, because, determine, first, and the alternation of now and then
indicate that there is a tie between the words; and 3) the linked volitional prominence
suggests that it is this very tie which is important in this context. Notice that in [5-16],
the prominent words could be used in strong dissociative opposition if their sequential
linking as steps was ignored, that is to say, if they were treated as discrete members of
the set of things that are done during the firing of a missile. Again, linked volitional
prominence only functions to elevate these words and to prompt a search for the
significance of their being lifted together, and it is the pre-existing links which lend
themselves to being interpreted in dissociating or coordinating terms.
3 Timing
TIMING (3) uses parallels in structure to promote the regularity of the rhythmic
pattern in a set of volitionally prominent beats. The more closely parallel that structure
is, the stronger the regularity of the rhythmic pattern, where the strength of that rhythm
augments the significance or strength attributed to the dissociation or coordination of
the words on which the beats fall.
279
To begin with, here are a couple of examples in which dissociation gets a boost
from timing:
[5-17] while common peril may be the measure of our need, the
existence or absence of a positive community must be the
measure of our capacity.
G72 0160, G72 0180
[5-18] Conscience and religion are concerned with private sin: The
civil law is concerned with public crimes.
F15 1870, F15 1880
Of course, this timing would also have a greater effect if there were more than two
beats with which to establish a rhythm, but there are few examples which display three
or more close prominences; in fact, there are only three examples with three instances
apiece, and then there is [5-16], which has five instances. Rarer still are those among
such examples which actually use timing to any effect, namely just the coordination of
the missile-launching steps in [5-16]. This constitutes linguistically rarefied air, but the
few examples of timing found so far are clear and consistent enough to warrant the
proposal of this filter for the sifting of audio- or video-recorded data in the future.
This analysis of timing harkens back to the end of the discussion about
volitional prominence and parallel reference (chapter 4, 3.4). In that section, I
generated an example which would only work in a context where a series of people hit
the baker until doing so became monotonous, where the additional prominence
emphasized the regularity of the rhythmic sing-song pattern of the beating:
[5-19] The btcher hit the BAKer, then the witer hit the BAKer, then
the podle hit the BAKer, then my mther hit the BAKer
(and so forth)
280
The measured rhythm of the prominence over the parallel structures can now be seen
more easily as linking the high points together, giving the impression of a conjoined
meaning which involves the notions of the prominent words representing respective
stages in a series. The rhythm even suggests that these actions are performed in
regularly timed intervals. This slideshow-like measure of beats is not an irrelevant or
fortuitous aspect of Ss construal, but rather it is a crucial part of the message that S is
trying to convey to L. Without congruence in this regard, their construals will not be in
alignment, and L will not really understand what S means.
4 Conclusion
Neither a conventional nor a contextual basis for linking material with italics
automatically determines that the result will reflect a pattern of volitional prominence;
while this is obviously more likely for conventionally than contextually linked words,
each set of prominent instances must be evaluated individually. Volitional prominence
serves no new function here, essentially making multiple substitutions with absolutely
explicit and thoroughly encapsulated counterparts, lifting the words above the baseline
distinctly enough that their linked prominence is interpreted as necessarily significant.
It is this implication which drives the search for the links between the words which
would have supported their being elevated together, but this search in itself is not a
prominence function. Linked prominent words are either dissociated as taxonomic
labels, or they are coordinated according to pre-existing causal or consequential links,
where both are augmented by the timing provided by close structural parallels.
281
CHAPTER 6
Conclusions and Future Research
You need stress in your life! Does that surprise you? Perhaps so, but it
is quite true. Without stress, life would be dull and unexciting. Stress
adds flavor, challenge, and opportunity to life.
Louis E. Kopolow, plain talk about... HANDLING STRESS
The results of this analysis provide direct support for the organization of the
functions of volitional prominence in terms of elaboration (word-internal meaning
change), revelation (word-external meaning change), and then their application to
instances of linked intense prominence. There is also direct support for the division of
elaboration into power and precision. Contextual effects portion power into inherent
and emotionally adventitious forms, and precision into its uses for intolerance and
position. The division of revelation according to the explicitness and encapsulation of
the changes counterpart is useful, and it reflects the results of other research into the
structure of belief and discourse. These linguistic functions are speculated to have
been adapted for use by language from communication-universal progenitors, whose
origins are in turn tied to research which supports the existence of primitive cognitive
abilities whose functions are analogous to the ones revealed here.
Further results of this study will be discussed in more detail in the following
sections, as they are devoted to describing what will be done in future research to
better support the more speculative conclusions.
282
1 Live Data
While the data analyzed in this study worked perfectly when it came to
revealing a systematic structure underlying the functions of volitional prominence,
there are only a few places where the written data can actually provide direct support
for the direct iconic proportion holding between form and meaning. Those few places
are not enough to supplement empathetic judgments. The questions involved in the
adaptation of sensation into language are the ones that I really want to get at, and so
the first thing that I want to do is incorporate audiotaped and videotaped data into the
analysis. I need to gather data using a format which allows audible and visible gestures
to be measured in terms of placement and level.
2 Increased Complexity
I am gong to increase the complexity of the data gathered, both in terms of the
length of the abnormally prominent sequences within a given utterance, and the types
of utterances within which those gestures reside. This will bring in more examples of
abnormally prominent interrogative pronouns and so forth, as well as expressions of
emotional discharge. As the sequence lengths increase over time, it is going to become
increasingly necessary to come up with a stronger distinction between an emotional
build-up that is discharged on one word, and a mood whose effects are more broadly
distributed over the course of an utterance. While smaller changes in the allowed data
wont change the analysis much, if at all (e.g. prominent one- and two-word sequences
essentially behave identically), there are sure be changes at some point.
283
Increasing the complexity of the examples will facilitate work on puzzles like
the following:
[6-1] It just seemed as if there was nothing else to do.
P18 1440
[6-2] Peter, it wouldnt hurt you to put your journey off for one more
day, Fred had said. It isnt as if youve got anything to go to,
especially. And Greg and Vicky are badly shaken by all this,
you know they are.
21.33
To provoke a preliminary understanding of these examples, just imagine using [6-1] in
answer to, Why didnt you do something? A rough initial analysis posits multiword
phrases of the form to V or to V particle which form conventional sets of the
following sort: {to (go (into)), (to) go (into), ((to) go) into}. This formalism
reflects the notion that the meaning is a purpose-movement-goal triad, and emphasis
on any one of the three is taken in contrast to the remaining two (which might or might
not appear explicitly in the utterance). In other words, when one word in such a to V
particle sequence is made prominent (e.g. to), then the related member of the
conventional set is identified as the change (e.g. to (go (into))), and the remaining
members automatically become counterparts (e.g. (to) go (into) and ((to) go) into).
So, the counterpart of to in [6-1] is (to) do (particle) or perhaps (to) do
(something), perhaps with volitional prominence on do. As things stand, however,
theres not enough data of this sort to support a conclusion of this complexity, so for a
more complete answer: read the sequel.
284
3 Increased Selectivity
I am interested in completing my analysis of elaborate grounding predications,
beginning with the nominal ones (the, a(n), unstressed some), which was shelved
for lack of the stable definition of precision presented herein, which suffered at the
time from a lack of recorded data specific to the problem. The behavior of these
predications as projected from generated data has precision approximating delimiters
equated with those boundaries proposed by Langacker (1991: 276) for his DYNAMIC
EVOLUTIONARY MODEL of reality. For example, the analysis suggested that the verbal
grounding predication may targeted the borders of potential reality which separated
it either from irreality or projected reality. This might be essentially the same behavior
that most displays in approximating either an upper extreme boundary (virtually all
M) or a lower extreme boundary (as little more than half of M as it can manage
without being only some of M). Now that a stable definition of precision has been
developed, the precise modals, for example, can be used to reveal an equally precise
model of reality. But first, a selective set of this rarefied data needs to be recorded.
4 More Modalities of Communication
One natural extension will be to study language forms other than that printed in
texts, including both human and nonhuman communication, and perhaps perception.
Such studies should eventually show that this behavior is communication universal to
the extent that prominence is iconic for power and precision, relative to the modality
of the articulators used in a given language or communication system.
285
This study already nurtures three tentative tendrils along these lines which
reach towards four intriguing statements made recently in Armstrong, Stokoe, and
Wilcox. The first creeper is drawn towards their contention that it becomes evident
that there may have been evolutionary steps from animal to human emotion displays,
to iconic and symbolic visual gesturing, to fully developed gestural language that
involved primarily the visual field (1995: 88). The second one closes in on an
assertion they support with their research, namely that The earliest linguistic units
may have been either visible or vocal gestures or, quite likely, both. We will explore
the notion of visible signing as a basis for subsequent linguistic evolution (p. 19).
What my analysis does at this point is simply to approach this same notion from the
side of audible gesture.
The third thread touches on two quotes from this same source, each of which
describes the abnormally prominent signing of DECIDE: If the answer to the
question is definite or authoritative, the movement will be abrupt, strong, and An
exaggerated lengthening and slowing of the arm might mean that the decision has been
long delayed (p. 87). This portrayal ties in with Bolingers work, where he researches
intonation patterns specifically taking into account the visible gestures which
accompany them (1986: ch. 9). He supports the notion that abstract intonational
meaning is derived from primitive iconic relations holding between the magnitude of
pitch, length, and loudness and the display of emotional arousal, which he takes in
support of a suggestion that at least some universality of meaning should be expected
(pp. 194, 202).
286
Revealing the potential universality of meaning between spoken and signed
language once provided the greatest drive behind this research. Several years ago, I
started to write about the degree of overlap between signed and spoken semantic
structure, showing that it might be extensive despite the disparity in their associated
physical articulations. I wanted to know how this overlap correlated with a continuum
running from iconic to symbolic representation. In effect, I wanted to collapse the
following square into a more triangular diagram by having the spoken and signed
semantics share components. Essentially, I was trying to find out to what degree
connection #4 could be shortened or eliminated into full overlap:
Figure 6-1: Full Symbolic Structure
t
A1
A2
A3
A4
phonological pole
t
OS
semantic pole
A
t
semantic pole
w w w
t
phonological pole
#2
#1
#3
#4
#5
#6
#7
spoken symbolic structure signed symbolic structure
G
287
This dissertation began as ancillary support for this research, addressing the depiction
of iconic meaning in semantic structure and its physical articulation. It got all of my
attention when clearer definitions were needed for prominence, iconic proportion,
sense, and so forth. Now that this dissertation is complete, all of the vocabulary has
been defined which will allow for the original line of research to be pursued, and I feel
far more comfortable suggesting that signed and spoken semantics overlap most
closely in the representation of prominence. I am also far less worried about being
torpedoed by the complaint that prominence is too ill-defined for reference.
5 Developmental Studies
Mehler, Jusczyk, Lambertz, Halsted, Bertonici and Amiel-Tison (1988)
determine that a newborn can show a preference for its mothers native language over
others which are not yet familiar to it. The results of their study indicate a sensitivity to
prosody which develops before birth, much as a newborn can identify their mothers
voice in preference to that of others. This sensitivity to intonation in general is attested
in Walker-Andrews and Grolnick 1983, Fernald 1984, and Fernald and Kuhl 1987.
Trehub, Endman, and Thorpe (1990) show that infants (7 to 8.5 months) can
discriminate timbre (in this case the difference between [a] and [i]) apart from
frequency, intensity, and duration. Children 3 to 4 years old were tested by Baltaxe
(1991) and found to be able to match four emotional intonation patterns well with
drawings of faces (angry, happy, sad, neutral). Studies of this sort on prominence
would be invaluable for supporting the adaptation scenarios proposed here.
288
6 The Big Picture
REASONABLE SPECULATION (6.1) is promoted by the ACTUAL RESULTS (6.2).
6.1 Reasonable Speculation
Meaning used to be just and only form. An environmental stimulus which
dominated a sensory field would naturally evoke an iconically voluminous set of
neural discharges (by magnitude and/or frequency) from an array of sensors that was
directly proportional to that stimulus in breadth, and that massive perceptual load
would be processed and reacted to just as severely as if the stimulus itself were a
proximal event (inside whatever served as the primary neural aggregation) rather than
a distal one (somewhere beyond the sensory extensions of that same mass). Similarly,
stimuli moving swiftly across a sensory field would evoke an iconically quick change
across a sensory array, and so forth. The primitive cognitive ability which allowed for
the immediate evaluation of and reaction to sensory input was power. The initially
derivative but eventually semi-autonomous cognitive ability which allowed for a
comparison between two (or more) such evaluations was precision.
At some point along the way, the equation between meaning and form adapted
into an equivalency, probably first by elaboration (great size now equals both great
size and greater size, plus the immediate consequences of and reactions to great(er)
size), and then by transitivity (hunger evokes the equivalent food, and food
evokes the equivalent green, so sooner or later hunger is cognitively, though not
always environmentally, associated with green). Meaning eventually achieves a
degree of detachment from environmental form which allows for reactions to internal
289
representations of sensory input without the presence of external stimuli (hunger
evokes green beyond the physical presence of any actual green stimulus in the
environment; rapid approach is an external stimulus which evokes fear, which also
comes to elicit cognitive models representing great size).
Phonological forms then get attached to other physical forms in the
environment as mediated by semantic representations, starting with nonsegmental,
nonsequential iconics (big undifferentiated articulations for big generalized stimuli),
and moving toward increasingly arbitrary, increasingly conventionalized units, since
segments are naturally more prone to conventionalization of their sequencing. The
earliest attachment of phonological form to physical form marks the transition from
sensation to communication, where communication runs through and beyond the use
of conventionalized iconic units. Somewhere right after that, these same units are used
to establish novel forms, ones which are not necessarily iconic, and that is where
communication starts to become language. Somewhere just before this process of
establishing new forms becomes conventionalized with its own units, language has
started to really settle in.
Now, power and precision have been tagging along this whole time, but where
once they were solely cognitive abilities, their application to the motion of the physical
articulators provides the venue for their adaptation first for communication, and
eventually for language. At first, they iconically determine the brute force (power) and
finesse (precision) that are used in the shaping of articulations, which are forms that
will come to achieve conventional status. Finally, their functional interaction comes to
290
be conventionalized as linguistic prominence, namely elaboration and revelation,
which has an influence on other linguistically conventionalized primitive functions,
like rhythm.
6.2 Actual Results
My analysis directly supports the characterization of prominence behavior in
terms of two linguistic functions, which are defined here as elaboration and revelation.
In general, prominence is used by S to correct a perceived mismatch between R and
R
S
, where S assumes that L adopts the shared model. S uses prominence for
elaboration on a phonological form as a signal to L that the conventional meaning of
that form is not the one that S has in mind. L can usefully interpret the additional
intensity of that prominence as a directly proportional iconic measure of the change to
be applied to its construal of the words semantic structure, that is to say, the meaning
becomes equally intense.
In those cases where the semantics represent an entity which is prone to having
some inherent quality increased in power, as happens in words like huge, fast, or
even quiet, it is likely that S intends for the additional prominence to be applied
semantically as an increase in power. In other cases of power, however, the additional
energy is simply a discharge of emotion on Ss part, as in surprise. Some words appeal
to spatial components which are prone to being articulated with greater precision, such
as the proportional relative quantifiers which rely upon sets of boundaries which can
be construed intolerantly. The range of variations in a words meaning can essentially
291
be parcelled into instances of a type, where the variations position within that type
becomes important to Ls understanding of the meaning that S has in mind. That
location can be right in the center of the type as a proper instance of the word (where
green means a real green-green), or it can be out on the periphery as an odd
example of the type (where green means a weird green). These subfunctions are
conveyed by context and subtle non-intensity variations in prominence.
Sometimes the perceived discrepancy in reality models is not word-internal,
but word-external, or contextual. In those cases, S draws Ls attention to a word either
by shifting the location of primary prominence (which has the side-effect of raising the
level of prominence on that word, which leads to its being treated as if it were
intensely prominent), or by increasing what would normally be a primary level of
prominence on a word to an intense level. S wants L to exchange that prominent word
for one which was introduced earlier in the discourse, or for one which S expects that
L has preloaded as archival material. L is more easily able to locate this earlier
counterpart to the prominent word because both words supply a value for the same
variable in a similar setting.
When the counterpart is explicit and encapsulated, namely when it appears
earlier as a single word in the context, then L is able to perform a simple substitution.
There is some limited evidence to support a variation in which the counterpart is
actually a gap in the earlier context, and that S is suggesting a prominent word as an
addition. When this counterpart either does not appear as a single word, or appears
only by implication, then both S and L can rely upon a number of strategies which aid
292
in Ls derivation of the counterpart, such as a familiarity with the prominent words
conventional associations, or extra attention paid to new associations described in
context. These subfunctions of revelation are also conveyed by contextual effects and
subtle non-intensity variations in prominence.
Volitional prominence serves no new function in instances of multiple
volitional prominence. When two absolutely explicit and thoroughly encapsulated
counterparts are lifted above the baseline, L interprets their linked prominence as
necessarily significant. This implication drives L to figure out why the words would
have been elevated together, choosing from among their being dissociated as
taxonomic labels, or their coordination according to pre-existing causal or
consequential links. Either of these types of links can be augmented by the timing
provided by close structural parallels.
To the degree that written data can reliably evoke internal representations of
prominence patterns in native users of a language, this analysis provides firm support
for the direct iconic proportion between changes in the phonological intensity of a
form and changes in the semantic intensity of a meaning (i.e. changes in the construal
of a semantic structure commensurate with the imaged or imagined application of
additional energy). This assessment of internal representations of prominence is
significantly more reliable for elaboration than revelation, and so the support for the
direct iconic proportion is not as strong for revelation. Further analysis of tape
recorded data should fortify the fundamental support for revelation and elaboration,
revealing that the change in phonological form for revelation is in fact proportional to
293
the level of disparity between models as perceived by S, who will be shown to use
greater prominence to repair broader mismatches in construal, correcting errors which
have their origin more deeply inside Ls archive or memory.
This division of linguistic prominence into elaboration and revelation falls out
naturally from the behavior of the data, as does the characterization of their
subfunctions. I originally tried to account for the data with a number of other filters
(grammatical, geometrical, directional), but they all failed until I hit upon power
and precision. This heuristic has vigorous lobbyists in other domains, and the lack of
credible competitors increasingly defends the likelihood that prominence has in fact
based elaboration and revelation on these two cognitive abilities. The reliance of both
physical and cognitive articulation upon power and precision, plus their actual
existence as primitive cognitive functions, both seem evident. My work simply
promotes their primitive-to-contemporary adaptation.
^ An End ^
294
Afterword
...the structure was too pretty not to be true.
James D. Watson, The Double Helix
Two years ago, when I opened Alistair Cookes America to sift it for data, a
browned panel from an old Dennis the Menace comic strip fluttered to the floor. It
had been clipped out of the Stockton Record almost twenty years earlier (August 13,
1977). Dennis and Joey are relaxing under a tree, and Dennis simply says, I was just
thinkin... this is what I want to do when I grow up.
Whoever clipped that panel had no idea that I would literally be doing this
when I grew up. I love little ironies like that. Thank you Hank Ketcham.
Tracy C. Mansfield Asheville, North Carolina; 1997.
295
APPENDIX I
The Anatomy of Prominence
The point of this appendix is not just to separate the less controversial material
out of a long research review, but to demonstrate the breadth of the array of
mechanisms available to prominence for activation with variable amounts of energy.
This figure should help in identifying at least some of the anatomical structures
described in the passage that follows:
Figure I-1: Anatomy of the Vocalization Tract
If you reach up and place the tips of your thumb and middle finger close together on
either side of your Adams apple, then the tip of your index finger will naturally come
tongue
nasopharynx
oropharynx
epiglottis
vocal folds
thyroid c.
cricoid c.
larynx
hyoid
{
thyroid
cricoid
trachea
trachea
vocal
fold
vocal
cord
rima glottis
GLOTTIS
epiglottis
arytenoid c.
laryngopharynx
diaphragm
lung
cartilages:
296
down to rest on the v-shaped notch of your THYROID CARTILAGE, which partially
encircles your LARYNX. The larynx is lined with mucous membrane, and it is the
musculocartilaginous opening of the enlarged upper end of the TRACHEA. The HYOID
BONE (an adapted gill bone, like the bones of the middle ear) runs in a semicircle above
the thyroid cartilage, and forms the upper lip of the larynx. The CRICOID CARTILAGE is a
narrow tube sitting just below the thyroid cartilage, and below that is a series of
similar tubes (open in the back) which define the rest of the trachea. The trachea is
lined on the inside with ciliated epithelium, and it runs down for a little over 11cm
before dividing at the CARINA (at the level of the 5th dorsal vertebra) into two BRONCHI,
each of which goes to a LUNG.
All told, the larynx is made up of nine cartilages (three paired and three single)
held together by an elastic membrane and moved by at least a dozen muscles. The
three paired cartilages are the ARYTENOIDS (pitcher- or spout-shaped) at the back of
the larynx, the small yellow CUNEIFORM cartilages which lie just anterior to the
arytenoids in the epiglottic fold (explained below), and the CORNICULATES (having
horn-shaped bits). Two of the single cartilages have already been mentioned, namely
the thyroid (which is actually a pair of broad, vertically curved laminae fused at the
Adams apple or LARYNGEAL PROMINENCE), and the cricoid. The third is the EPIGLOTTIS.
The epiglottis (or EPIGLOTTIC CARTILAGE) is a thin flap of fibrocartilage
anchored just posterior to the root of the tongue. It runs up out of the larynx, and is
pulled down to divert swallowed material away from the GLOTTIS. The glottis consists
of the VOCAL FOLDS and the space between them (RIMA GLOTTIDIS or RIMA VOCALIS),
which form the sound-producing parts of the larynx. (Some researchers identify the
glottis solely as the rima glottidis.) Similarly, the RIMA RESPIRATORIA is the space
behind the arytenoids, the RIMA ORIS is the mouth opening, and the RIMA VESTIBULI (or
VESTIBULE) is the space between the FALSE VOCAL FOLDS. The false vocal folds are also
called the VENTRICULAR FOLDS, and the ventricle of the larynx is the space between the
true and false vocal folds. The third division of the LARYNGEAL CAVITY is the INFERIOR
ENTRANCE TO THE GLOTTIS.
Now for the muscles of the larynx. To begin with, there are exterior muscles
running from the bones and cartilages to the surrounding structures in the neck, such
as the STERNOTHYROID, STERNOHYOID, and OMOHYOID. To end with, there are interior
muscles running in all directions between most of the bones and cartilages of the
larynx, such as the OBLIQUE and transverse arytenoid, the CRICOTHYROID, the
THYROEPIGLOTTIC (or THYROEPIGLOTTIDEUS, the depressor of the epiglottis), and the
INTERNAL and EXTERNAL THYROARYTENOID. Technically, the VOCAL CORDS are
specifically the thin bands of tissue which vibrate in the tracheal airflow, the VOCAL
FOLDS are the edges of these bands, and the VOCAL LIGAMENTS are enclosed in the vocal
folds. Lateral to and touching these ligaments are lengths of the internal
thyroarytenoid, the posterior ends of which are anchored to the ARYTENOID PROCESSES,
which are the upper knobs of the arytenoid cartilages. Muscles which rotate these
processes away from the glottis cause the vocal ligaments and attached tissue to
become taught.
297
Inspiration itself relies upon the contraction of the diaphragm and the external
intercostals to lift the rib cage, enlarge the thoracic cavity, and decrease the mean
subglottal pressure. A normal breath takes in about half a liter of air, and preparation
for speech draws in one-and-a-quarter liters or more. The diaphragm leaves off at the
peak of the inspiration, and air is expelled due to the relaxation pressure of the lungs
(due to the stretched elastic tissues, torque, and gravity) and the contraction of the
external intercostals. Each of the external intercostal muscles attaches to the lower
margin of a rib, and then reaches up and inserts itself in the upper margin of the next
rib, together forming a sheet over the outside of the ribcage. When they contract, they
pull the ribs up and together like a set of horizontal blinds.
When the subglottal mean air pressure is equal to that in the lungs, the external
intercostals leave off and the internal intercostals take over, decreasing the size of the
thoracic cavity, and increasing subglottal pressure. Each of the internal intercostals lies
between two adjacent ribs, underneath the external intercostals, which pulls the ribs
together, but not up. For every half second (or thereabouts) that speech is prolonged
after that, additional muscles act to compress the abdominal contents against the
diaphragm, starting with the external abdominal obliques (which run from the lower
eight ribs down to the iliac crest and linea alba), then the rectus abdominis (up from
the pubis to the cartilage of the 5th to 7th ribs), and finally the latissimus dorsi (up
from the lower spine and tip of the iliac crest to the groove between the two large
processes on the head of the humerus). They all leave off at once when the next
inspiration begins.
My analysis has yet to adequately explore the hands and face as articulators of
prominence, and so I will not include their full anatomical description here, but even
without such a recital it should be evident that they would both be susceptible in their
anatomical sophistication to a wide range of variably powerful and precise activations,
just as are the vocal articulators described above.
298
APPENDIX II
Distribution of Data
Table II-1: Prominence Frequency (Functions in Order of Analysis)
Table II-2: Prominence Frequency (Functions Ranked by Total Instances)
The functions are
listed in the same order
as their appearance in
the analysis.
Grammatical Category
n
o
u
n
p
r
o
n
o
u
n
v
e
r
b
a
u
x
i
l
i
a
r
y
a
d
j
e
c
t
i
v
e
a
d
v
e
r
b
s
i
m
p
l
e

r
e
l
.
w
o
r
d

p
a
r
t
T
o
t
a
l
F
u
n
c
t
i
o
n
power
13 1 12 9 8 2 45
intolerance 19 4 1 13 7 44
position 13 5 4 6 1 29
substitution
11 30 7 11 2 1 62
addition 1 1 1 2 2 7
derivation (conv.) 16 13 19 64 30 32 29 1 204
derivation (cont.)
12 16 4 1 7 3 43
dissociation 15 11 2 3 15 5 9 2 62
coordination 3 9 2 4 18
Total
84 96 57 72 80 68 52 5 514
The functions are
ranked by their total.
p
r
o
n
n
o
u
n
a
d
j
a
u
x
a
d
v
v
e
r
b
s
i
m
p
p
a
r
t
T
o
t
a
l
F
u
n
c
t
i
o
n
derivation (conv.)
13 16 30 64 32 19 29 1 204
substitution 30 11 11 2 7 1 62
dissociation 11 15 15 3 5 2 9 2 62
power
1 13 9 8 12 2 45
intolerance 19 1 4 13 7 44
derivation (cont.) 16 12 7 1 3 4 43
position
5 13 6 1 4 29
coordination 3 2 9 4 18
addition 1 1 1 2 2 7
Total
96 84 80 72 68 57 52 5 514
299
:
Table II-3: Expanded Functional and Grammatical Categories
a. nominal grounding predication (e.g. the, a, relative quantifier)
Whole Word
P
a
r
t

o
f

a

W
o
r
d
T
o
t
a
l
nominal verbal relation
n
o
u
n
pronoun
v
e
r
b
auxiliary complex simple
p
e
r
s
o
n
a
l
i
n
d
e
f
i
n
i
t
e
p
o
s
s
e
s
s
i
v
e
o
t
h
e
r
n
o
n
m
o
d
a
l
m
o
d
a
l
a
d
j
e
c
t
i
v
e
a
d
v
e
r
b
p
r
e
p
o
s
i
t
i
o
n
c
o
n
j
u
n
c
t
i
o
n
n
g
p
a
E
l
a
b
o
r
a
t
i
o
n power 13 1 12 9 8 2 45
intolerance
2 5 4 8 4 1 13 4 3 44
p
o
s

n proper 3 3 4 10
periphery 10 3 2 1 2 1 19
R
e
v
e
l
a
t
i
o
n
substitution 11 24 2 4 7 11 2 1 62
addition
1 1 1 2 2 7
d
e
r
i
v
a
t
i
o
n
c
o
n
v
e
n
t
i
o
n
a
l
pair 14 2 10 1 29 13 11 6 11 1 98
set 8 1 1 1 1 2 1 5 1 21
basic
4 56 5 14 79
complex 2 4 6
c
o
n
t
e
x
t
u
a
l pair 3 1 8 1 1 3 1 18
set 7 1 2 2 1 13
basic
1 1
complex 1 2 4 1 2 1 11
L
i
n
k dissociation 15 6 2 3 2 3 15 5 9 2 62
coordination
3 9 2 4 18
Total 84 47 12 24 13 57 59 13 80 68 25 13 14 5
514
300
Table II-4: Expanded Categories with Subtotals
Whole Word
p
a
r
t

o
f

a

w
o
r
d
T
o
t
a
l
nominal verbal relation
t
o
t
a
l
n
o
u
n
pronoun
t
o
t
a
l
v
e
r
b
auxiliary
t
o
t
a
l
complex simple
t
o
t
a
l
p
e
r
s
i
n
d
p
o
s
s
m
i
s
c
t
o
t
a
l
n
a
v
m
o
d
t
o
t
a
l
a
d
j
a
d
v
t
o
t
a
l
p
r
e
p
c
o
n
j
n
g
p
t
o
t
a
l
E
l
a
b
o
r
a
t
i
o
n
power
13 1 1 14 12 0 12 9 8 17 2 2 19 45
45
intoler. 2 5 4 8 19 19 4 4 4 1 13 14 4 3 7 21 44
44
p
o
s
i
t
i
o
n
prop 3 0 3 3 0 3 4 4 0 4 10
10
per
10 3 2 5 15 1 0 1 2 1 3 0 3 19
19
total 13 3 2 0 0 5 18 4 0 0 0 4 6 1 7 0 0 0 0 7 29 0
29
total 26 5 7 4 9 25 51 16 0 4 4 20 16 22 38 0 6 3 9 47 118 0
118
R
e
v
e
l
a
t
i
o
n
subs
11 24 2 4 30 41 7 0 7 11 2 13 1 1 14 62
62
addition 1 1 1 2 0 0 1 2 3 0 3 5 2
7
d
e
r
i
v
a
t
i
o
n
c
o
n
v
e
n
t
i
o
n
a
l
pr 14 2 2 16 10 1 1 11 29 13 42 11 6 11 28 70 97 1
98
set
8 1 1 1 11 11 1 2 2 3 1 5 6 1 1 7 21
21
bas 0 0 4 56 5 61 65 14 14 0 14 79
79
cx 2 0 2 4 0 4 0 0 0 6
6
tot 16 8 3 1 1 13 29 19 56 8 64 83 30 32 62 11 7 11 29 91 203 1
204
c
o
n
t
e
x
t
u
a
l
pr 3 1 8 9 12 1 1 1 2 3 1 4 0 4 18
18
set 7 1 1 8 2 0 2 2 1 3 0 3 13
13
bas
1 0 1 0 0 0 0 0 1
1
cx 1 2 4 6 7 1 0 1 2 1 3 0 3 11
11
tot 12 4 0 12 0 16 28 4 0 1 1 5 7 3 10 0 0 0 0 10 43 0
43
total 28 12 3 13 1 29 57 23 56 9 65 88 38 36 74 11 7 11 29 103 248 1
249
total
40 36 5 18 1 60 100 30 56 9 65 95 49 39 88 12 7 11 30 118 313 3 316
L
i
n
k
dissoc. 15 6 2 3 11 26 2 3 3 5 15 5 20 9 9 29 60 2
62
coord.
3 0 3 9 0 9 2 2 4 4 6 18
18
total 18 6 0 2 3 11 29 11 3 0 3 14 15 7 22 13 0 0 13 35 78 2
80
Total
84 47 12 24 13 96 180 57 59 13 72 129 80 68 148 25 13 14 52 200 509 5 514
301
APPENDIX III
Other Corpus: Sources for Abnormal Prominence
These sources are listed in the order in which they were searched for data:
[1] Steinbeck, John. 1943. Once There was a War. New York [1960]: Bantam.
[2] Steinbeck, John. 1962. Travels with Charley (in Search of America). New York
[1963]: Bantam.
[3] Lewis, Sinclair. 1922. Babbitt. New York [1961]: Signet.
[4] Lewis, Sinclair. 1920. Main Street. New York [1961]: Signet.
[5] Maule, Harry E., and Melville H. Cane. 1963 [1950]. The Man from Main Street.
New York: Pocket Books.
[6] Steinbeck, John. 1937. Of Mice and Men. Camden: Haddon Craftsmen.
[7] Steinbeck, John. 1939. The Grapes of Wrath. New York: Stratford Press.
[8] Reiser, Paul. 1994. Couplehood. New York: Bantam.
[9] Powers, Joan. 1995. Poohs Little Instruction Book. New York: Duttons.
[10] Poundstone, William. 1983. Big Secrets. New York: Quill.
[11] Poundstone, William. 1986. Bigger Secrets. Boston: Houghton Mifflin.
[12] Sunset Books. 1982. Azaleas Rhododendrons Camellias. P. Edinger (ed.) Menlo
Park: Lane Publishing.
[13] Fosburgh, Lacey. 1991. India Gate. New York: Crown. (only disallowed data)
[14] Thomas, Elizabeth M. 1993. The Hidden Life of Dogs. Houghton Mifflin: Boston.
(only disallowed data)
[15] Lackey, Richard S. 1980. Cite Your Sources. Jackson: University Press of
Mississippi.
[16] Wells, Richard A. 1889. Decorum. Springfield: King, Richardson & Company.
[17] Sagan, Carl. 1977. The Dragons of Eden. New York: Ballantine.
302
[18] Forbes, Malcolm. 1988. They Went That-a-Way. New York: Ballantine.
[19] Brooks, Mel; and Carl Reiner. 1981. The 2000 Year Old Man. New York: Warner.
[20] Sheppard, Muriel E. 1991. Cabins in the Laurel. Chapel Hill: The University of
North Carolina Press.
[21] Francis, Dick. 1991. Comeback. New York; Putnams.
[22] Cooke, Alistair. 1973. Alistair Cookes America. New York: Knopf.
[23] Huff, Theodore. 1951. Charlie Chaplin. New York: Henry Schuman.
[24] Abbey, Lynn. 1987. Unicorn & Dragon. New York: Avon.
[25] Miller, Dorcas. 1986. Berry Finder. Rochester: Nature Study Guild. (no italics)
[26] Wood, Ernest. 1992. The Irish Americans. New York: Mallard Press. (no italics)
[27] Sagan, Carl. 1980. Cosmos. New York: Random House.
[28] Taylor, Robert L. 1967. W. C. Fields: His Follies and Fortunes. New York:
Signet.
[29] White, E. B. 1952. Charlottes Web. New York: HarperTrophy.
[30] Caterpillar Tractor Company. 1975. 50 Years on Tracks. Peoria: Caterpillar
Tractor.
[31] Libby, Bill. 1970. Andretti. New York:Grosset & Dunlap.
[32] Jansz, Natania, and Barbara Davies. 1990. Women Travel. New York: Prentice
Hall.
[33] Helck, Peter. 1961. The Checkered Flag. New York: Castle Books.
[34] Michener, James. 1959. Hawaii. New York: Bantam.
[35] Dick, Phillip K. 1976. Martian Time-Slip. New York: Ballantine.
[36] Brodie, Fawn M. 1974. Thomas Jefferson: An Intimate History. New York:
Norton.
[37] Maclaine, Shirley. 1974. My Lucky Stars: A Hollywood Memory. New York:
Bantam.
[38] Kelley, Don G. 1971. The Edge of a Continent. Palo Alto: American West.
303
APPENDIX IV
Brown Corpus: Sources for Abnormal Prominence
A. PRESS: REPORTAGE
[A35] Maguire, J. 1961. A family affair. Commonweal 11/10: 171-173.
B. PRESS: EDITORIAL
[B25] Werth, A. 1961. Walkabout in Moscow. Nation 11/11: 370-373.
C. PRESS: REVIEWS
[C03] Prescott, O. 1961. Books of the Times. NY Times 11/17: 33.
D. RELIGION
[D02] Ogden, S. 1961. Christ Without Myth. NY: Harper, pp. 128-134.
[D03] Kelly, E. 1961. Christian unity in England. America 6/3: 398-400.
[D05] Miller, P. 1961. Theodore Parker: Apostasy within liberalism. Harvard Theo.
Rev. 10: 280-285.
[D09] Golay, E. 1961. Organizing the local church for effective lay visitation
evangelism. Tidings pp. 54-61.
[D10] Smith, H. 1961. Interfaith communication: The contemporary scene. J. of
Bible and Rel. 10: 308-311.
[D11] Ramsey, P. 1961. War and the Christian Conscience. Durham: Duke Univ.
Press, pp. 200-206.
[D17 0180] Anon. 1961. Guideposts: 15th anniversary issues. Guideposts 1: 1-5.
[D17 1760] Rivero, J. 1961. The night our paper died. Guideposts 2: 1-3.
E. SKILL AND HOBBIES
[E01 0210] Weider, B. 1961. Henri de Courcy: Junior Mr. Canada. Mr. Am. 11: 9-12,
42.
[E01 1170, 1190, 1240] Weider, J. 1961. The mark of the champ. Mr. Am. 11: 22, 25,
52.
[E03] Martin, D. 1961. Will aircraft or missiles win wars? Flying 2: 32-33, 80-83.
[E04] Goldsmith, H. 1961. The Schnabevel/Pro Arte Trout. High Fidelity 10: 75-76.
304
[E07] Larson, P., and G. Odegard. 1961. How to design your interlocking frame.
Model Railroader 2: 57-65.
[E13] Deardorff, R. 1961. Step by step through Istanbul. Travel 12: 51-53.
[E14] Carnahan, A. 1961. Nick Maneros Cook-out Barbeque Book. NY: Fawcett, pp.
13-21.
[E17] Anon. 1961. This Is the vacation cottage you can build. The Family
Handyman 11: 13-19, 58.
[E20] Anon. 1961. What you should know about air conditioning for new homes.
How-to-Do-It-Encyclopedia. Volume 1. NY: Golden Press, pp. 9-14.
[E24] Prudden, B. 1961. The dancer and the gymnast Dance Mag. 4: 20, 52-54.
[E26] Dibner, B. 1961. Oerstad and the discovery of electromagnetism. Elec. Eng.
6: 426-428.
[E27] Bay, M. 1961. What can 'additives' do for ruminants? Successful Farming 12:
51, 68.
F. POPULAR LORE
[F01] Blackmon, R. 1961. How much do you tell when you talk? Vogue 7: 40-41,
100.
[F03] Rapport, N. 1961. I've been here before! Fate 4: 65-70.
[F04] Rosevear, R. 1961. North country school cares for the whole child.
Prevention 9: 82-83, 85, 88.
[F07] Sentnor, M., and S. Hult. 1961. How to have a successful honeymoon.
Sexology 8: 35-38.
[F08] Reaves, P. 1961. Who rules the marriage bed? Pageant 11: 46-51.
[F11] Pompian, L. 1961. Tooth-straightening today. Everywomans Fam. Circ. 3: 8,
107.
[F12] Marian N. 1961. New methods of parapsychology. Tomorrow F: 45-50.
[F13] Scoville, O. 1961. Part-Time Farming. Farmers Bulletin No. 2178, U. S. Dept.
of Agriculture 12: 2-3, 5-7.
[F14] Rosenberg, H. 1961. The trial and Eichmann. Commentary 11: 371-374.
[F15] O'Brien, J. 1961. Let's take birth control out of politics. Look Mag. 10: 67-70.
[F16] Boylan, J. 1961. Mutinity Real 12: 17-18, 59-60.
[F23] Goldwater, B. 1961. A foreign policy for America Nat. Rev. 3: 177-179.
305
[F27] Churchill, C. 1961. A Notebook for the Wines of France. NY: Knopf, pp. 124-
129
[F33] Conant, J. 1961. Slums and Suburbs. NY: McGraw-Hill, pp. 42-47
[F37] Cavert, S. 1961. On the Road to Christian Unity. NY: Harper, pp. 136-141.
[F38] Smith, R. 1961. Baseball in America. NY: Holt, Rhinehart and Winston, pp.
199-200, 202-204.
[F42] Baldwin, J. 1961. Nobody Knows My Name. NY: Dial Press, pp. 66-74.
[F43] Getlein, F., and H. Gardiner. 1961. Movies, Morals, and Art. NY: Sheed and
Ward, pp. 48-54.
[F44] Winter, G. 1961. The Suburban Captivity of the Churches. NY: Doubleday, pp.
70-75.
[F46] Baker, R. 1961. An American in Washington. NY: Knopf, pp. 118-127.
[F48] Ramsey, P. 1961. Christian Ethics and Sit-In. NY: Association Press, pp. 108-
115.
G. BELLES-LETTRES
[G02] Miller, A. 1961. Toward a concept of national responsibility Yale Rev. 12:
186-191.
[G04] Burdick, E. 1961. The invisible aborigine Harpers Mag. 9: 70-72.
[G05] O'Donnell, T. 1961. Evenings at the bridge Horizon 5: 26-30.
[G08] Murphy, F. 1961. New Southern fiction: Urban or agrarian? Carolina Q. Sp:
18-25.
[G13] Glicksberg, C. 1961. Sex in contemporary literature. Colorado Q. W: 278-
282.
[G14] Santmyer, H. 1961. There were fences. The Antioch Rev. Sp: 26-31.
[G15] Nemerov, H. 1961. Themes and methods: The early stories of Thomas Mann.
Carleton Misc. W: 6-11.
[G16] Hayward, J. 1961. Mimesis and symbol in the arts. Chicago Rev. Su: 94-99.
[G17] Stewart, R. 1961. A little history, a little honesty: A Southern viewpoint.
Georgia Rev. Sp: 10-15.
[G20] McLachlan, D. 1961. Communication networks and monitoring. Public
Opinion Q. Su: 196-202.
[G21] Cheney, B. 1961. Christianity and the tragic vision-utopianism USA. The
Sewanee Rev. F: 18-524.
306
[G22] Reiner, K. 1961. Coping with runaway technology. Ethical Outlook 5-6: 91-
95.
[G27] Josephson, M. 1961. Jean Hlion: The return from abstract art. Minnesota
Rev. 4: 346-350.
[G29] Anon. 1961. References for the good society. Manas 4: 1-2.
[G30] Hanson, N. 1961. Copernican and Keplerian astronomy. J. Hist. Ideas 4: 174-
179.
[G31] Fineman, I. 1961. Woman of Valor: The Life of Henrietta Szold. 1860-1945.
NY: Simon and Schuster, pp. 50-56.
[G32] Farr, F. 1961. Frank Lloyd Wright. NY: Scribner's, pp. 182-188.
[G34] Golden, H. 1961. Carl Sandburg. NY: World Pub., pp. 82-87.
[G36] Copp, D., and M. Peck. 1961. Betrayal at the UN: the story of Paul Bang-
Jensen. NY: Devin-Adair, pp. 208-215.
[G37] Hall, G. 1961. Golden Boats from Burma. Philadelphia: Macrae-Smith, pp. 56-
64.
[G38] Goldgar, B. 1961. The Curse of Party: Swifts Relations with Addison and
Steele. Lincoln: Univ. of Nebraska Press, pp. 140-145.
[G40] Fowler, G. 1961. Skyline: A Reporters Reminiscences of the 1920s. NY:
Viking, pp. 170-175.
[G43] Lane, R. 1961. The Liberties of Wit: Humanism, Criticism, and the Civil Mind.
New Haven: Yale Univ. Press, pp. 112-118.
[G47] Davis, G. 1961. The World Is My Country. NY: Putnam's, pp. 92-97.
[G49] van Kuykendall Thomson, P. 1961. Francis Thompson, A Critical Biography.
NY: Nelson, pp. 172-178.
[G50] Davis, C. 1961. The Kings Chevalier: A Biography of Lewis Littlepage.
Indianapolis: Bobbs-Merrill, pp. 192-196.
[G55] Flanders, R. 1961. Senator from Vermont. Boston: Little, Brown, pp. 124-129.
[G56] McKean, K. 1961. The Moral Measure of Literature. Denver: Swallow, pp. 16-
22.
[G57] Williams, R. 1961. Values and modern education in the United States. in D.
Barrett (ed.), Values in America. Notre Dame: Univ. of Notre Dame, pp. 74-80.
[G67] Schorer, M. 1961. Sinclair Lewis: An Am. Life. NY: McGraw-Hill, pp. 468-
472.
307
[G70] Krutch, J. 1961. If you don't mind my saying so. The Am. Scholar W: 120,
122, 124.
[G71] Frank, J. 1961. Andr Malraux: The image of man. The Hudson Rev. Sp: 52-
57.
[G72] Fulbright, J. 1961. For a concert of free nations. Foreign Affairs 10: 12-17.
[G74] McCormick, J. 1961. The confessions of Jean Jacques Krim. The Noble
Savage 10: 8-12.
[G75] Garrett, G. 1961. A wreath for Garibaldi. The Kenyon Rev. Su: 484-489.
H. MISCELLANEOUS: GOVERNMENT & HOUSE ORGANS
[H25] U. S. Secretary General. 1961. Report of the Secretary General in John Simon
Guggenheim Memorial Foundation, Reports of the Secretary General and of the
Treasurer, 1959 and 1960. NY: pp. 16-24.
[H28] Anon. 1961. Carleton College Bulletin. Northfield, MN: 3: 92-99.
J. LEARNED
[J08] Fothergill, L. 1961. Biological warfare, in P. Gray (ed.), The Encyclopedia of
the Biological Science. NY: Reinhold, pp. 145-149.
[J12] McLaughlin, R., et al. 1961. A study of the subgross pulmonary anatomy in
various mammals. The Am. J. of Anat. pp. 154-157.
[J13] Pyle, S., et al. 1961. Onsets, Completions, and Spans of the Osseous Stage of
Development in Representative Bone Growth Centers of the Extremities.
Longitudinal Studies of Child Health and Dev., Harvard School of Public
Health, Series II, No. 12. Monograph of the Soc. for Res. in Child Dev., 26: 1.
Serial No. 79, pp. 20-21, 24-25.
[J17] Gellhorn, E. 1961. Prolegomena to a theory of the emotions. Perspectives in
Bio. and Med., pp. 426-431.
[J19] Mosteller, F., et al. 1961. Probability with Statistical Applications. Reading,
MA: Addison-Wesley.
[J24] Parad, H. 1961. Preventive casework: problems and implications, in The
Social Welfare Forum. NY: Columbia Univ. Press, pp. 186-191.
[J27] Womble, D. 1961. Functional marriage course for the already married.
Marriage and Family Living pp. 280-282.
[J31] Searles, H. 1961. Schizophrenic communication. Psychoanalysis and the
Psychoanalytic Rev. pp. 14-18.
308
[J32] Kelly, H., and T. Ziehe. 1961. Glossary lookup made easy. Proceedings of the
National Symposium on Machine Translation, University of California at Los
Angeles. Englewood Cliffs: Prentice-Hall, pp. 326-331.
[J38] Comm. for Economic Development. 1961. Distressed Areas in a Growing
Economy. A statement on national policy by the Research and Policy Committee
of the Committee for Economic Development. NY: pp. 48-53.
[J41] Braff, A., and R. Miller. 1961. Wage-price policies under public pressure. So.
Econ. J. pp. 163-165.
[J43] Mendelson, W. 1961. Justices Black and Frankfurter: Conflict in the Court.
Chicago: Univ. of Chicago Press, pp. 90-96.
[J46] Perluss, I. 1961. Agricultural labor disputes in California 1960.
Employment Security Rev. 1: 5-7.
[J50] Bonbright, J. 1961. Principles of Public Utility Rates. NY: Columbia Univ.
Press, pp. 342-347.
[J53] Haymond, W. 1961. Is distance an original factor in vision? The Modern
Schoolman pp. 40-45.
[J54] Starr, C. 1961. The Origins of Greek Civilization 1100-650 B. C. NY: Knopf,
pp. 144-150.
[J57] Hexter, J. 1961. Thomas More: On the margins of modernity. J. of Brit.
Studies 11: 28-32.
[J58] Ray, J. 1961. Rhode Island's reactions to John Brown's raid. Rhode Island
Hist. 10: 100-105.
[J59] Greenberg, C. 1961. Collage in his Art and Culture: Critical Essay. Boston:
Beacon Press, pp. 74-79.
[J60] Futterman, R. 1961. The Future of Our Cities. Garden City: Doubleday, pp. 62-
68.
[J64] McDonald, K. 1961. Figures of rebellion. Opera News 1: 21-23.
[J65] Hynes, S. 1961. The Pattern of Hardys Poetry. Chapel Hill: Univ. of North
Carolina Press, pp. 130-137.
[J66] Rexroth, K. 1961. Disengagement: The art of the beat generation, in T.
Parkinson (ed.), A Casebook of the Beat. NY: Crowell, pp. 181-184.
[J69] Anon. 1961. IBM Reference Manual IBM 7070 Series Programming Systems
Autocoder. IBM, pp. 16-22.
[J72] Anon. 1961. Independent research, in Mellon Institute Annual Report 1960.
Pittsburgh: Mellon Institute, pp. 10-13.
309
[J73] Anon. 1961. Directory of Continuing Numeric Data Projects. Washington, DC:
Nat. Acad. of Sciences, pp. 14-21.
[J78] Dolon, P., and W. Niklas. 1961. Gain and resolution of fiber optic intensifier.
Proceedings of the Image Intensifier Symposium October 24-26. Washington,
DC: Office of Scientific and Technical Information, NASA, pp. 93-97.
K: FICTION: GENERAL
[K12] Green, G. 1961. The Heartless Light. NY: Scribner's, pp. 166-170.
[K13] Maxwell, W. 1961. The Chateau. NY: Knopf, pp. 240-245.
[K15] Hebson, A. 1961. The Lattimer Legend. NY: Macmillan, pp. 190-195.
[K20] Bolton, G. 1961. The Olympians. NY: World Pub., pp. 128-134.
[K22] Cheever, J. 1961. The brigadier and the golf widow. New Yorker 11: 53-54.
[K28] Bingham, S. 1961. Moving day. Atlantic Monthly 11: 63-65.
L: FICTION: MYSTERY
[L02] Fair, A. 1961. Bachelors Get Lonely. NY: Morrow, pp. 82-89.
[L03] Dean, A. 1961. Encounter with Evil. Garden City: Doubleday, pp. 48-54.
[L07] Holden, G. 1961. Deadlier Than the Male. Garden City: Doubleday, pp. 96-
104.
[L08] Shannon, D. 1961. The Ace of Spades. NY: Morrow, pp. 184-191.
[L10] Olesker, H. 1961. Impact. NY: Random House, pp. 94-101.
[L13] Hitchens, D. 1961. Footsteps in the Night. Garden City: Doubleday, pp. 156-
161.
[L14] Lockridge, F. 1961. Murder Has Its Points. Philadelphia: Lippincott, pp. 96-
101.
[L16] Gordon, A. 1961. The Cipher. NY: Simon and Schuster, pp. 238-245.
[L20] Lacy, E. 1961. Death by the numbers. Manhunt 8: 8-12.
[L22] Barlow, S. 1961. Monologue of murder. The Saint Mystery Mag. 12: 121-125.
[L23] Rose, J. 1961. Try my sample murders. Trapped Detective Story Mag. 11: 45-
50.
[L24] Brown, F. 1961. The Murderers. NY: Dutton, pp. 116-122.
310
M: FICTION: SCIENCE
[M01] Heinlein, R. 1961. Stranger in a Strange Land. NY: Putnams, pp. 250-255.
[M03] Blish, J. 1961. The Star Dwellers. NY: Avon, pp. 83-88.
[M04] Harmon, J. 1961. The planet with no nightmare. If 7: 7-12.
[M05] McCaffrey, A. 1961. The ship who sang, in J. Merril (ed.), 7th Annual Edition
The Years Best S-F, NY: Mercury Press, pp. 311-317.
N: FICTION: ADVENTURE
[N05] Ferber, R. 1961. Bitter Valley. NY: Dell, pp. 9-17.
[N09] Thompson, J. 1961. The Transgressors. NY: NAL, pp. 9-13.
[N12] Booth, E. 1961. Outlaw Town. NY: Ballantine, pp. 103-108.
[N15] Plantz, D. 1961. Sweeney Squadron. NY: Dell, Inc., pp. 133-138.
[N17] Prather, R. 1961. The bawdy beautiful. Cavalier 4: 64-65.
[N19] Jackson, D. 1961. The English gardens. Partisan Rev. 3-4: 216-221.
[N21] Sommers, C. 1961. The beautiful mankillers of Eromonga. Cavalcade 10: 60,
62.
[N23] Hall, W. 1961. Always shoot to kill. Bluebook for Men 10: 34-35, 52-53.
[N29] Ellison, H. 1961. Riding the dark train out. Rogue 5: 14, 30.
P. FICTION: ROMANCE
[P03] McMeekin, C. 1961. The Fairbrothers. NY: Putnam's, pp. 258-264.
[P09] Ford, J. 1961. Mountains of Gilead. Boston: Little, Brown, pp. 128-133.
[P10] Williams, J. 1961. The Forger. NY: Atheneum, pp. 4-8.
[P12] Callaghan, M. 1961. A Passion in Rome. NY: Coward-McCann, pp. 124-129.
[P14] Biddle, L. 1961. Sam Bentleys Island. NY: Doubleday, pp. 70-75.
[P15] Burrough, L. 1961. The open door. Good Housekeeping 5: 117-118.
[P16] Brown, M. 1961. A secret between friends. Redbook 2: 104-110.
[P17] Hine, A. 1961. The huntress. Saturday Evening Post 2/4: 84-85.
[P18] Anon. 1961. No room in my heart to forgive. Modern Romances 11: 76-78.
[P19] Anon. 1961. This cancer victim may ruin my life. Medical Story 9: 18-21, 36.
[P21] Spencer, E. 1961. The white azalea. Texas Q. W: 112-115.
311
[P22] Anon. 1961. A husband stealer from way back. True Love 10: 6-7, 12.
[P23] Robinson, B. 1961. Something very much in common. McCalls 8: 172-174.
[P24] Elkin, S. 1961. The ball player. Nugget 10: 25-26, 31.
[P25] Butler, W. 1961. The pool at Ryusenji. Harpers Bazaar 1: 146-147.
[P27] McGiffin, L. 1961. Measure of a man. Ladies Home J. 6: 103-104.
[P28] Hoover, C. 1961. The shorts on the bedroom floor. Story 2: 38-42.
[P29] Carson, R. 1961. My Hero. NY: McGraw-Hill, pp. 170-174.
R. HUMOR
[R01] Loos, A. 1961. No Mother to Guide Her. NY: McGraw-Hill, pp. 108-115.
[R02] Mercier, J. 1961. Whatever You Do, Dont Panic. NY: Doubleday, pp. 28-35.
[R03] Dennis, P. 1961. Little Me. NY: Dutton.
[R04] Streeter, E. 1961. The Chairman of the Bored. NY: Harper, pp. 220-226.
[R05] Esar, E. 1961. Humorous English. NY: Horizon Press, pp. 141-148.
[R06] Thurber, J. 1961. The future, if any, of comedy. Harpers Magazine 12: 40-43.
[R07] Wildman, J. 1961. Take it off. Arizona Q. F: 246-252.
[R08] Lemon, L. 1961. Catch up with and Something to talk about. Mademoiselle
7: 8, 15, 17, 47-49.
[R09] Perelman, S. 1961. The Rising Gorge. NY: Simon and Schuster, pp. 201-207.
312
Author Index
A
Abbs: Folkins and Abbs, 74
Amiel-Tison: Mehler, Jusczyk,
Lambertz, Halsted, Bertonici and
Amiel-Tison, 287
Armstrong (D.), Stokoe, and Wilcox, 29,
285
Armstrong (L.) and Ward, 90, 9395, 171
B
Balogh, 234, 237
Baltaxe, 287
Bardovi-Harlig, 137154, 157159, 172,
175
Berman and Szamosi, 110, 117119
Bertonici: Mehler, Jusczyk, Lambertz,
Halsted, Bertonici and Amiel-Tison,
287
Blandford: See Palmer, 95
Bloch and Trager, 101
Bloomfield, 5657, 60, 87, 99102
Bolinger, 20, 77, 107108, 110, 112,
131134, 141, 152, 156, 172, 186,
285
See also Gussenhoven, Bolinger, and
Kjeispur, 134
Bresnan, 110, 116120, 133134, 147,
149
Burnet, 8385
Butler, 8182
C
Carnap, 155
Chafe, 34, 141
Chappalaz, 93
Chomsky, 114, 152
Chomsky and Halle, 110, 114, 133,
162
Chomsky, Halle, and Lukoff,
113114
Coleman, 92, 99, 170171
Cooper: Liberman, Cooper, Harris, and
MacNeilage, 74
Crystal and Quirk, 93
D
Dane, 138139, 141, 163
Diesing, 160
Draper, Ladefoged, and Whitteridge, 66
See also Ladefoged, Draper, and
Whitteridge, 66
E
Endman: Trehub, Endman, and Thorpe,
287
F
Fairbanks: House and Fairbanks, 61
Fauconnier, 31, 216
Fernald, 287
Fernald and Kuhl, 287
Firbas, 141144
Folkins and Abbs, 74
Fowler and Rosenblum, 7576
Francis and Kucera, 5, 10
Fry, 6162, 72
G
Gay, 6972
Giesler, 73
313
Gilbert, 217
Gilbert, Krull, and Malone, 217
Gilbert, Tafrodi, and Malone, 217
Grolnick: Walker-Andrews and
Grolnick, 287
Gunter, 31, 136137, 226
Gussenhoven, Bolinger, and Kjeispur,
134135
H
Halle: See Chomsky, Halle, and Lukoff,
113
Halliday, 152
Halsted: Mehler, Jusczyk, Lambertz,
Halsted, Bertonici and Amiel-Tison,
287
Harris, K.: Liberman, Cooper, Harris, and
MacNeilage, 74
Harris, Z., 101
Hart, 8182, 86
Haugen, 105
Hawkins, 31
Hayes, 128129
Horn, 156
House and Fairbanks, 61
Hultzn, 82
Huntington and Metcalf, 108
J
Jackendoff, 152154, 156157, 215
Jones, 5658, 6566, 82, 9192, 171
Josephs, 73
Jusczyk: Mehler, Jusczyk, Lambertz,
Halsted, Bertonici and Amiel-Tison,
287
K
Kent and Netsell, 7071
Kingdon, 86, 93, 9798
Kjeispur: See Gussenhoven, Bolinger,
and Kjeispur, 134
Klinghardt, 57, 9394, 98
Krifka, 158, 160163, 215
Krull, 31
See also Gilbert, Krull, and Malone,
217
Kucera: See Francis and Kucera, 10
Kuhl: Fernald and Kuhl, 287
Kuno, 140141
Kunze, 66
L
Ladd, 147148, 152153, 186
Ladefoged, 55, 6364, 6667, 71, 73, 76
Ladefoged and McKinney, 66
Ladefoged, Draper, and Whitteridge,
66
See also Draper, Ladefoged, and
Whitteridge, 66
Lakoff, 31, 110, 117119, 149, 168, 195
Lambertz: Mehler, Jusczyk, Lambertz,
Halsted, Bertonici and Amiel-Tison,
287
Lambrecht, 163165, 226
Langacker, 20, 25, 31, 168169, 178,
197, 259
Lasersohn, 167168, 201
Lehiste, 55, 69
Lehiste and Peterson, 72
Liberman, 121
Liberman, A.
Liberman, Cooper, Harris, and
MacNeilage, 74
Liberman and Mattingly, 74
See also Mattingly and Liberman, 74
Liberman, M., 125, 128, 146
Liberman and Prince, 121, 125
Lieberman, 55, 60, 6264, 77, 127
Lifshitz, 61
314
Lindblom, 76
Lukoff: See Chomsky, Halle, and Lukoff,
113
M
MacNeilage
Liberman, Cooper, Harris, and
MacNeilage, 74
See Sussman and MacNeilage, 70
Malone, K., 6566
Malone, P.
Gilbert, Krull, and Malone, 217
Gilbert, Tafrodi, and Malone, 217
Mattingly and Liberman, 74
See also Liberman and Mattingly, 74
McKinney: Ladefoged and McKinney,
66
Mehler, Jusczyk, Lambertz, Halsted,
Bertonici and Amiel-Tison, 287
Metcalf: Huntington and Metcalf, 108
Mitford, 108
Mller, 56
N
Needham, 44
Netsell: See Kent and Netsell, 70
Newman, 105, 110113, 116, 119, 132,
162, 171
O
Odell, 83, 8789
hman, 6869, 71, 128
P
Palmer, 57, 66, 86, 90, 9597, 102
Peterson: Lehiste and Peterson, 72
Pierrehumbert, 121123, 126128
Pike, 62, 101, 103107, 131
Prince, A.: See Liberman and Prince, 121
Prince, E., 44, 141
Q
Quirk: Crystal and Quirk, 93
R
Ripman, 99
Rooth, 156, 158160, 162163
Rosenblum: See Fowler and Rosenblum,
75
S
Schmerling, 139140, 145146, 148, 186
Schubiger, 57
Sheridan, 83, 89
Silvera, 73
Smith: See Trager and Smith, 105
Steele, 83, 85, 8789
Stetson, 56, 5860, 6667
Stokoe: See Armstrong (D.), Stokoe, and
Wilcox, 285
Sussman and MacNeilage, 7071
Sweet, 83, 86, 170171
Szamosi: See Berman and Szamosi, 118
T
Tafrodi: Gilbert, Tafrodi, and Malone,
217
Thelwall, 85
Thorpe: Trehub, Endman, and Thorpe,
287
Trager, 106
Trager and Smith, 62, 101, 105107,
113
See also Bloch and Trager, 101
Trehub, Endman, and Thorpe, 287
Trim, 93, 97
315
V
van Ginneken, 65
von Stechow, 159
W
Walker, 8283, 8587, 89, 154, 191
Walker-Andrews and Grolnick, 287
Ward: See Armstrong (L.) and Ward, 93
Wells, R., 62, 102103, 106
Wells, W., 173, 175
Whitteridge
Draper, Ladefoged, and Whitteridge,
66
Ladefoged, Draper, and Whitteridge,
66
Wilcox: See Armstrong (D.), Stokoe, and
Wilcox, 285
316
Subject Index
A
abnormal prominence: See prominence,
volitional #
accent 81, 87, 133135, 143, 172
bi-tonal pitch 122, 126
boundary tone 122123, 125
dynamic 66
energy 65
expressive 111, 171
focus 134
interest 172
phonemic 113
phrase 122126
pitch 122, 125126, 133134
placement 136
power 172
sentence 160, 162
stress 110, 132
tonic 66
acoustic signal 73
activation cost 34
activity 121
sequence 55
acuteness 85
adaptation 28, 165
adapted resource 55
addition 4344, 136, 291
adjective 182, 185
adverb 180, 182, 201
adverbial of time 200202, 251252
air pressure 58, 64, 66
alchemy 9
alignment 130
boundary 180
conceptual construals 179
intensification 144
meaning 205
non-prosodic and prosodic CD 144
perfect correspondence 144
reality models 51
tune-text 122, 125
all-capitals font: See font, all capitals "$
allosentence 163, 226
alternative semantics 159, 163
amber light 108
amplitude 61
analog curves 91
anatomical configuration 55
anatomy 55
appropriate semantic variable 215
appropriateness 155156, 166
Archive 33, 3536
archive 32, 36, 215, 221, 293
argument 139
articulation 25, 55, 180, 286287, 293
force 110
intensity 111
normal 55
rhythmic 121
articulators 284
ancillary 51
articulatory energy 21, 51
arytenoid cartilage 55
aspect 214
aspiration 111
assertion 154156, 164
atemporal relations 178
attitude of S 103
audiotape 282
B
background 159, 163, 215
317
balloon
esophageal 64, 66
gastric 59
tracheal 64, 66
baseline 122, 127, 176
excursion 76
frequency 77
beat 121, 128130, 137, 153, 278280
major stress 137
placement 137
belief 30, 189, 203, 211, 215, 217218,
281
bellows 54, 59
bold font: See font, bold "%
boundary 194
breath
force 171
group 77, 123
pulse 129
unmarked group 77
Brown 5
corpus 7, 10, 183, 189, 198, 204205,
214, 219, 223, 227, 241, 255,
258260, 267, 278
C
cadaver 56
cardinal number 198, 252
catheter, tracheal 64
CD: See communicative dynamism "%"
CDS: See current discourse space $&
change 43, 216220, 281, 283
chest pulse 5859, 67
child 287
Chinese 84
chordate 25
Christianity 273
cinefluorography 70
circumflex slide 86
coarticulatory interference 76
coefficient 215
cognition 23
cognitive
process 180
routine 50, 176
color blindness 12, 14
comment 119, 139, 141, 153, 156157,
164
communication 9, 20, 165, 284, 289
communication-universal progenitors
281
communicative dynamism 141144, 147,
173
comparison 25
complex conventional domain 263
complex wave 68
conceptual construal 179, 199
conditional 192
consciousness states 34
constituent
focus 134135, 161
metrical 121
syntactic 162
construction, grammatical 132
context 31, 36, 115, 135138, 143,
154155, 163164, 180182,
185, 196, 202205
dependence 141
grammar 136
independent 181
physical 33
sensitive 145
contextual 243
competition 181
meaning difference 2, 43
pair 244, 249, 264265, 275
relevance 137
set 244, 266267, 269
contour, intonation 122
contrast 6, 9, 118119, 138, 171
external 7
318
contrastive
accent 152
stress 151
conventional 243
pair 192193, 224, 228, 242245,
247251, 258, 263, 266,
268270, 272273, 275
set 8, 192193, 201202, 220, 224,
228, 232, 244, 247, 250253,
256258, 267, 272273, 276
conventionalized
meaning 111
unit 289
conversation 35
coordination 47, 228, 251, 272, 277279,
292
Copernicus 268
core 2
corpse 108
correction 211, 236237
counterpart 8, 43, 197, 203, 215220,
222, 224, 227, 229, 238, 241,
243245, 247248, 251252,
256, 258, 263271, 280281, 283,
291292
contextual 242
conventional 242
distributed 196, 220, 222, 243, 247,
270
encapsulated 230, 243
explicit 230, 243, 245
explicit unitary 222
implicit 220, 242243, 250, 269
multiword 249, 265
null 241243
value 233
critical points 80
cue, extradimensional 5
current discourse space 3536, 211, 215
Cycle, transformation 117
cycle, transformation 110, 116118
D
decrescendo 174175
default meaning 2
definiteness 31
deictic 192, 194
delimiter 180
demonstrative 180, 193
demonstrativeness 171
depth of memory 34
derivation 4344, 219, 227229, 232,
241245, 247, 250, 253254,
263265, 267, 270, 292
conventional set 200, 202
Descartes 217
developmental studies 287
diaphragm 57
dimension 1
dimensional
contrast 12
differences 12
limitation 3
direct
evaluation 21
proportion 70, 175, 210, 288
direct iconic proportion 78, 37, 42, 175,
177, 179, 181, 185, 210, 282, 290,
292
directly powerful 21
discourse 8, 20, 26, 30, 135, 137138,
157, 160, 163164, 281
counterpart 216, 219
discrete tone 87
dissociation 4647, 221, 251, 272273,
276, 278280, 292
distal
event 21, 75
signal power 21
distributed 8, 43, 243
dolphins 54
domain 273
319
basic 229, 257, 273
basic contextual 244, 267
basic conventional 244, 248,
257258, 262, 267
complex 221, 273
complex contextual 244, 265, 267,
269
complex conventional 244, 263, 267
contextual 268, 276
conventional 224
depth 274
intermediate focus 134
positive/negative opposition
227229, 248, 257, 273
selection theory 158
size 257258
whole/part 257
drum 84
duration 59, 6162, 6970, 73, 76, 78, 84,
86, 92, 129, 174, 176, 191, 287
dynamic evolutionary model 284
E
ecological event 74
effort 9, 65, 68, 70, 73, 179180, 190
cognitive 210
intense 70
normal 70
physical 65
effortful motion 78
elaboration 3738, 51, 136, 166172,
178179, 181182, 187, 192,
196197, 201205, 208210,
212213, 225, 227230, 239, 242,
250, 264, 268, 271, 273, 281, 288,
290, 292293
elastic recoil of the lungs 64
electromyograph 66
ellipsis 136
emotion 81, 107, 136137, 187188, 191,
194, 235, 281282, 285, 287, 290
predictability 133
emotional charge 86, 181, 188191, 203,
209, 251, 278
emotive coloring 144
emphasis 9, 82, 139, 144, 154, 170171
contrastive 171
negative 170171
positive 170171
proper 170
emphatic
stress 87
encapsulation 8, 43, 216, 218219
energy 65, 176, 183, 292
accent 65
dynamic 66
expenditure 69
intense 70
normal 55, 70
of maintenance 66
of production 65
physiological 69
pulse 69
quanta 51
speech unit 65
static 66
syllable 65
English 10, 16, 1819, 49, 53, 67, 69, 77,
81, 8384, 90, 9294, 99,
102105, 118, 121122, 127, 137,
150, 153, 160, 274
euphemism 205207, 209
evoked 141
excluded data 15
expectation 3031
experience 21
explicitness 8, 43, 216, 218219
expression 34, 137
expressive
accent 111
prosody 111, 113
320
external cue 2
F
falling slide 86
false consensus 33
feeling 172
fight-or-flight 194
finesse 179180, 182, 289
focal
element 166
operators 163
focus 134135, 138, 140, 152158,
160164, 166, 173, 215
argument 164165
associated with 156
broad 147, 152153, 159
complex 163
constituent 153
external 166
information 134
interest 134
internal 166
inward 166167, 169, 172, 177
level 173, 175
multiple 153, 162163
narrow 147, 152153
outward 165, 167, 169, 172175, 177
predicate 164165
scope 154
sentence 154, 164
sentence structure 163
shifted center 151
unmarked 147
word 154
focus-phonology correlation 173
focus-sensitive quantification 160
focussing operator 158159, 162
font
all capitals 1314
bold 14
italic 10, 1415
italics 12, 246
mechanical italics 19, 185
small capitals 15, 136
underlined 10, 12, 1415, 234, 246
force 9, 111
articulatory 110
brute 179, 182, 289
extra 111
nonphysical 179
physical 179
foreground 159
foreign word 16, 18
form 1, 282, 288
form-meaning
ambiguity 1
pairing 1
free choice 157
French 65, 92
frequency 287
baseline 127
formant 62, 76, 78, 176
fundamental 6162, 67, 69, 7172,
7778, 122, 127, 176
functional
sentence perspective 143
value: See communicative dynamism
"%"
G
G: See ground "*(
generated examples 12
German 92, 97
gestural effort 78, 176
gesture 7, 73, 137
audible 282, 285
communication-universal 27, 284
language-universal 27
visible 282, 285
vocal 285
321
glide 103
gradation 37, 271
grammar
category 147
construction 120
gravity 85
Greek 18, 8384, 8788
ground 194, 197198
primary 197198
secondary 197
grounding predication 197, 284
H
head 95, 143
heteromorphic 74
homomorphic 74
homonymic conflict 152
I
ICM: See idealized cognitive model $"
iconic 12, 29, 285287
size 111
iconicity 165
iconics 165, 171, 289
idealized cognitive model 3133
IDS: See immediate discourse space $&
illusion 1
immediacy 194
immediate discourse space 3536
implicit counterpart 8
indefinite pronoun 253
indirect evaluation 21
indirectly powerful 22
infant 287
cries 77
inferrable 141
inferred 44
inflection 80, 84, 86
information 110, 135, 148, 154, 157, 165,
172
focal 154
focus 158
given 141
new 119120, 134135, 140141,
148, 172, 211, 215, 217, 237
new about old 211
old 120, 134135, 170, 211
predictability 135
presuppositional 154
semantic 119
shared 158
structure 165
unshared 157158
informativeness 132135
inherent 281
initial state 35
initiation 216
in-line marking 86
instantiation 169
instructions to leave
Newman 112, 116, 119
intense 173175
effort 70
intensity 7, 59, 61, 64, 69, 164, 170171,
175, 287
absolute 24
articulatory 111
phonological 292
relative 24
semantic 292
interest 135, 172, 175
intolerance 192193, 196, 200202, 209,
281, 290
intonation 80, 97, 120, 137138, 285
as music 99
center 138139
contour 80, 99, 120, 122, 124,
127128, 130, 137, 154
contour skeleton 94
critical point 130
group 94
322
ideophonic origin 121
metaphoric 121
metrical analysis 121
pattern 106, 124, 287
plateau 123
rhythm default 128
sensitivity 287
unit 111
intralinear tonetic dot notation 96
intramuscular probe 66
intratextual transcription 96
Italian 83
italic font: See font, italic "%
J
juncture 113
K
Kennedy 13
kymograph 59, 91
L
L: See looker/listener $&
label 17
taxonomic 273276, 280, 292
taxonomic switching 276
lack of stress 76
lambda operator 155
language 9, 214, 284, 289
gestural 285
signed 286
spoken 286
user 2, 50, 110111, 113, 119120,
122, 125, 131, 133135, 176,
271
user intent 111
user mood 111
user personality 111
laryngeal muscles 55
larynx 5556, 59, 62
Latin 18, 84, 8788
length 282, 285
level
intense 172
phonological 175
semantic 175
stress phoneme 112
tone 82, 87
liminal
stage 99
thing 108
limit 182, 201
linguistic evolution 285
link
causal 272, 277, 280
consequential 225, 228, 272, 277, 280
pre-existing 272, 276, 278, 280
location 180, 192
looker/listener 35, 38, 42, 5152, 131,
135, 137, 154, 163, 170, 172,
178180, 198199, 202,
206207, 209211, 218, 222, 224,
226227, 230, 237238,
241245, 247, 250, 254, 265267,
269270, 280, 290293
belief 43, 45
conceptual construal 179
direct attention 51, 179, 199, 203,
229, 241, 255, 261, 265, 291
interpretation 179181, 250, 264,
291292
mental contact 178, 197
mistake 211
personal space 195
reality model 51
loud 73
louder 84
loudness 57, 61, 64, 73, 8687, 92, 106,
111, 113, 191, 285
maxima 174
323
lungs 54, 58
M
M: See reference mass "*(
magnitude 9, 198, 285
mass
reference 284
matrix 161163, 215
meaning 1, 158, 165, 170, 179, 181, 184,
189, 201202, 205206, 209, 282,
285, 287288, 290
alternative 158
conventionalized 111
postulate 158159
usual 158
Megarians 195
memory 30, 34, 271, 293
mental
contact 178, 197, 224
effort 65
energy 65
spaces 216
mermaids 54
metalinguistics 107
metrical
analysis 121
constituent 121
grid 121, 128
microlinguistics 107
misalignment 209
mismatch 179, 202, 207208, 211, 216,
230, 241, 263, 271
modality 284
Monotone 8687
mood 282
morphological markers 160
Motor Theory of Speech Perception
7376
motor unit 70
discharge rate 70
recruitment 70
recruitment interval 70
muscle
abdominal 58
external intercostal 58
motion 59
rectus abdominis 58
music 80
musical
accent 84
notation 80, 85
representation 85
myth 54, 56, 67
N
native language 287
Necker cube 1
newborn 287
newspaper 267
nonconventional word 16
nonphysical context 32
non-prosodic CD
content 144
distribution 143
normal
focus 147
intonation 120
prominence: See prominence, routine 2
sentence stress 9, 111
sound form 76
stress 100, 106107, 110, 114, 133,
139, 151
utterance 151
noun 183, 185186, 202, 214, 264
classifier 186187
novel 243
Nuclear Stress Rule 116119, 129, 133
metrical 121
nucleus 96, 143
nucleus-tones 95
324
null counterpart 44, 241243
O
occlusion 4
omen 215217, 222, 241, 244245
ordering hypothesis 116120, 134, 147
ordinal number 252
orientation 180
Other corpus 10, 15, 183, 187, 198, 204,
214, 226, 228, 243, 255, 257
Ottomans 195
P
paragon 41, 182, 195, 203204, 239
contender 202203
parallel structure 233
paranymy 269
partially iconic 165
partition 163
passion 86, 154, 191
pattern
stress 132
syntactic 132
peak sound pressure 67, 73
peanut butter 179
pendulum 98
percept 23
perception 23, 72
periodic wave 68
peripheral 41
position 41
periphery 180
perseverence 31
PFSA: See principle of functional stress
assignment "%&
phoneme 99, 113
heavy stress 116
primary 99
secondary 99
sequential suprasegmental 102
simultaneous suprasegmental 101
stress 111
phonemic
analysis 99
intonation 99
juncture 102
pitch 99
principles 80
stress 102
stress accents 111
phonetic
characteristics 175
gesture 75
phonological
form 2, 13, 42, 51, 179, 185, 205207,
209210, 231232, 289
intensity 38, 181, 213
level 175
match 209
phonology
underspecification 2
phonology-focus correlation 173
physical
context 33
effort 65
form 181, 289
power 181
physicality 183184, 192
physiological
energy 69
invariance 76
perception 72
production 56
stress correlates 129
physiology 56, 72
pitch 59, 62, 66, 69, 81, 86, 9293, 111,
122, 127, 129, 154, 174, 188, 191,
285
diacritic 91
extra high 114
height 103, 106, 122
325
profile 172
placement
middle stress 112
non-content 151
non-rightward 151
subordinate stress 112
syntactic predictability 110
plus junctures 106
pointer 231, 233
points of punctuation 80
position 41, 180, 192, 202203, 205, 281,
291
negative proper 41, 203, 268
peripheral 192, 202, 205, 208210,
239, 264, 291
positive proper 41, 203204
proper 41, 180, 202205, 209, 291
positive/negative opposition 258
power 7, 9, 20, 2425, 2829, 3839, 51,
70, 171172, 176, 179187, 189,
191192, 194195, 201, 206, 213,
227228, 239, 277, 281, 284,
288290, 293
adventitious 187
emotion 187
nonphysical 181
physical 181
pragmatic
aboutness 139140
halo 167169
slack 167, 201
pragmatics 164165, 167
precision 7, 9, 20, 2425, 2829, 3840,
51, 70, 171, 176, 179183, 192,
194195, 200201, 225, 229,
231232, 250, 252254, 281, 284,
288290, 293
predicate 139
predictability 110, 186
stress accent placement 112
stress from accent 113
syntactic 110, 112, 118, 131, 133, 138
preference 287
prehead 143
pre-nuclear variation 95
preposition 247, 258
presupposition 154157, 164
presuppositional set 154, 156157
primary 172173, 175
phoneme 99
sentence stress 9
stress 138, 175
priming 44
primitive cognitive abilities 281
principle of functional stress assignment
145146
process 183
prominence 2, 24, 31, 57, 122, 130131,
133134, 137, 144145, 154, 157,
159, 161, 163166, 168173,
175176, 179183, 185188,
192, 194, 196, 199, 202203,
208209, 284, 287, 290
abnormal 109110, 281, 285
baseline 127
baseline units 127
close 225, 272, 279
contextual 131
declination 127
default pattern 121
discourse functional 131
iconic 121
intense 199, 207, 213, 236
intensity 175
inward focus 131
level 173174, 210, 212, 227, 291
linked 38, 46, 4849, 162, 219221,
225, 240, 272274, 276278,
280, 292
normal 165
outward focus 131
pattern 186
326
phonological components 173
phonological level 172
placement 166, 175, 212, 236, 291
pragmatics 131
primary 148, 174, 212213, 234, 236
primary level 213
primary sentence 212
prosodic 143
revelatory 212213
rhematic 181
rhythm 128
routine 19, 29, 4950, 55, 64, 72, 97,
131, 135, 154, 165, 169,
211212, 229, 235, 238
routine placement 212213, 235
secondary 175, 212, 234
semantics 131
shared 47
topic 133
volitional 1012, 19, 29, 3738,
4142, 44, 46, 4851, 53, 55,
72, 89, 109, 128, 130131,
135, 137, 164165, 168169,
175, 209, 211213, 216,
218219, 221, 226, 229237,
239241, 251, 259260, 264,
271272, 274277, 279280,
292
weak 175
prominent
admission 257258
confirmation 229, 258, 260, 262
negation 258259
prominent admission 261262
pronoun 138, 195197, 208209, 219,
223
demonstrative 195, 229
indefinite 209, 224225, 254
interrogative 209
personal 220, 229231
possessive 220, 231
reference 197, 229
proper 41
proportion 187
direct 122
inverse 122
proportional 22
proposition 164
prosodeme 105
prosodemic 105
prosody 165
expressive 111, 113
sensitivity 287
sentence 165
prosody-grammar association 100
proximal
event 21
signal power 21
stimulation 75
p-set 156, 158
Ptolemy 268
punctuation 80
Q
quanta of energy 34, 51, 65
quantification
entities 162
objects/entities 160
situations 160, 162
quantifier 253
absolute (true) 198
individual 200, 254
non-proportional 199
proportional 199, 225, 254
relative 160, 180, 197200, 254
true 198
universal 199
quantifier(restrictor, matrix) 160
quantity 111, 180181, 186187, 199
quantum prominence 68
Question Formation 116
327
R
radial categorization 31, 168
raised voice 18
range 156
reality 31
irreality 284
model 8
potential 284
projected 284
recapitulation 136, 222
reception 21
reduplication 3
reference 31, 136, 223, 233
mass 197200
point 197
relation 179, 183, 192
relational target 121
relations 182
relative 199
phoneme 107
quantifier 199
Relative Clause Formation 116
replacement 136
resolution 4
resonance, 300 Hz subglottal 62
resource 28
respiratory
activity 71
anatomy 55
mechanism 56
pattern 78
structures 55
system 5556, 66
restrictor 161163, 215
revelation 30, 37, 46, 51, 136, 165, 167,
169172, 193, 195197, 203, 208,
211215, 218219, 222223,
228230, 233, 239, 241, 243, 257,
264, 268, 271, 281, 290, 292293
addition: See addition #"*
derivation: See derivation #"*
substitution: See substitution #"*
rhematicity 142
rheme 138139, 141, 143146, 152153,
172176, 237
non-prosodic 144
rhythm 47, 55, 68, 94, 120121, 130, 176,
239240, 272, 278280, 290, 292
default 130131, 159
rhythmic beats 80
rising slide 86
Romans 195
routine 50, 135
cognitive 50, 176
intonation 34
sentence structure 34
tolerance 167
rule
Bardovi-Harligs stress assignment
principle 140
Bresnans topical stress assignment
147
contour dip 126
contour interpolation 126, 128
formal stress assignment 138
global 117
principle of functional stress
assignment 145
Schmerlings stress assignment 139,
145
sentence accent assignment 134
sentence stress assignment 134
straight contour 126
tone spreading 126
tune-text alignment 122, 126
S
S: See speaker/signer $&
SAAR: See Sentence Accent Assignment
Rule "$%
328
scale
appearance/existence 142
attributive 143
scanning 199, 255
scope
focus 159, 161162
phonological 159
semantic 159
word 159
secondary 172173
phoneme 99
stress 138
semantic
content 143, 229
intensity 7, 38, 185
level 175
partition 160
stress assignment 118
structure 205, 207209, 290
variable 155
semantics 165, 287, 289290
predictability 133
representation 154
scale 142
signed 286287
spoken 286287
stress placement 119
sensation 9, 20, 23, 165, 282, 289
sense 7, 50, 8081, 89, 130131, 135,
137, 145, 154, 170, 172, 176177
group 94
sensualizing 180
sentence
news 139140, 145146
normal stress 9, 111
primary stress 9
stress 84
thematic 140
topic-comment 139140, 145146
Sentence Accent Assignment Rule
134135
sequence of activity 55
sequential
segmental symbolism 5
suprasegmental phoneme 102
setting 43, 215, 217, 219222, 226227,
229, 231232, 241245,
258261, 270
sex 220
shouting 14
shrimp 186
signal energy 21
simultaneous suprasegmental phonemes
101
sing-song 239
size 181, 183, 185186
slack
pragmatic 167, 201
regulator 167
slide 86
smarter percussion 84
softer 84
space
figurative 192
literal 180
parcelled 180
space-builder 216
spatial structure 180
speaker/signer 25, 3545, 48, 5052, 131,
135, 137, 139, 154, 163164, 170,
172, 178179, 181, 197198, 202,
204211, 216218, 222, 224,
226227, 230, 241242,
269271, 280, 290291, 293
attitude 103, 165
belief 43, 45, 218
conceptual construal 179
construal 280
emotion 3940, 290
emotional charge 42, 181, 191
experience 181
intent 250, 264
329
mental contact 197
vocal baseline 122
speaking rate 70
speech
event 164, 197
organs 54
power 72
rate 70
speech-act distinctions 165
speed 183
Spinoza 217
states of consciousness 34
strength of conviction 8, 184, 189
stress 57, 59, 6972, 81, 84, 87, 93, 97,
113, 129, 133, 139, 143, 145, 153,
157
abnormal 100, 118
accent 110, 132
context-sensitive 153
contour 118
contrastive 114, 118, 138139,
150151, 154
default 121
deviant 114
emphatic 97, 157
expressive 114
extra heavy 114
function 131
heavy 116
main 154
metrical 146, 148
myth 147, 149, 152
Newmans six levels 112
non-contrastive 118
non-emphatic 118
normal 100, 114, 118
nuclear 112
nuclear heavy 112113
nuclear placement 112
pattern 110
phoneme 111, 113
phonemic level 100
phonetic correlates 129
physiological correlates 129
placement 119, 132133
predictability from accent 113
primary 119, 138, 141, 145146, 148,
153154, 175
primary sentence 116, 125
primary word stress 116
prosodic 118
rheme 146, 148
rhythm 154
secondary 138, 145, 175
semantic 119
semantic assignment 118
sentence 121, 137138, 140
sentence-initial 141
strong/strong pattern 113, 132
strong/weak pattern 112, 132
syllable 81, 84
syntactic predictability 110
theme 146148
topic 119
truth 147, 149, 152
stressed syllables 63
structure
argument-focus 226
deep 117, 119
grammatical 165
information 165
parallel 223, 234235, 238, 272, 278,
280, 292
phrase 160
semantic 290
sentence focus 163
spatial 180
surface 118119, 154155
syntactic 119, 154
subglottal pressure 58, 64, 67, 78, 176
peak 63
subject-verb agreement 214
330
subphonemic features 74
substitution 43, 4647, 51, 136, 219,
222225, 227, 229233,
237238, 241, 243244, 247, 252,
258, 263265, 270, 280, 291
superfix 106
suprasegmental phoneme sequence 103
Swedish 6869, 76
swim bladders 55
syllable 6769, 76, 154
accented 126
metrical representation 125
metrical strength 121
metrically strong 122, 125126
open 58
post-peak 127
stress 57, 81, 129, 143
stressed 58
strong 127
unaccented 126
symbol 1
symbolic 285286
syntactic
category 138
constituency 109, 112
constituent 116
default 135
markers 160
phrasal componentiality 168
predictability 110, 112, 118, 131,
133, 138
stress patterns 147
structure 109, 119, 132
transformation 116
unit 80, 109
syntagmatic combination 168
syntax 147
T
taboo 209
tail 96, 143
target 178, 197, 209
taxonomy 46
technical terms 15
tempo 174
temporal relations 178
tense 214
terminal boundaries 106
theme 138141, 143145, 152153,
172173, 175176
non-prosodic 144
thing 178180, 197198
threshold 29
timbre 287
timing 47, 240, 272, 274, 278280, 292
tolerance 7, 167, 169, 171, 180, 182, 193,
197200
tone
bi-tonal pitch accent 123
boundary 122, 125
elemental: See Pierrehumberts two
tones "##
generator 122, 124, 128
inflection 83
level 82, 87
movement 174175
pattern generator 123
peak 111
Pierrehumberts High 122
Pierrehumberts Low 122
Pierrehumberts rules 122
Pierrehumberts two levels 121
Pierrehumberts two tones 122128
trailing 126
topic 119, 139140, 153, 156157, 164
deep 119
surface 119
tracheal puncture 59
tracheotomy 60
transduction 20
transient data 12
331
transitional element 141142, 173, 175
transitivity 288
truth-theoretic denotation 168
TS: See type specification "()
tune 81
peak 80, 91
type 41
type specification 178179, 186, 202,
229
types of utterances 282
U
underlined font: See font, underlined "%
unit
absolute quantity 65
enumerative 112
iconic 289
intonational 111112, 116
speech 65
utterance
abnormal 150151
normal 150
V
value 195, 197, 215216, 220222,
229233, 258259
variable 195, 197, 215, 220221, 229,
231233, 258
dependent 156157
independent 156157
semantic 162
verb 183184, 202, 228
vermiform 25
videotape 282
visualizing 180
vocal cord frequency 67
vocalization 55
voice vibration 91
voicebox 54
volition 50
volitional prominence 178179,
196198, 205206, 276
volume 186
volume velocity 67
vowel
quality 61, 73, 78, 176
target 76
W
Walkers Monotone 82
wave metaphor 187
weak 175
woodwind 54
word
accent 59
content 147
function 147
order 139, 143, 147
word-external meaning difference 2, 43,
196, 254, 291
word-internal meaning difference 2, 8,
38, 51, 179, 195196, 202,
208209, 250, 254, 291
World 3435
world 33, 156
writing 214
written recorded data 12
Z
zero/weak 172
332
References
Abercrombie, David, D. B. Fry, P. A. D. MacCarthy, N. C. Scott and J. L. M. Trim (eds.)
1964 In Honour of Daniel Jones: Papers contributed on the occasion of his
eightieth birthday 12 September 1961. London, England: Longmans,
Green and Co Ltd.
Armstrong, David F., William C. Stokoe, and Sherman E. Wilcox.
1995 Gesture and the Nature of Language. Cambridge, England: Cambridge
University Press.
Armstrong, Lilias E., and Ida C. Ward
1926 A Handbook of English Intonation. Cambridge, England [1959]: W.
Heffer & Sons Limited.
Balogh, Jennifer
1997 The Effect of Contrastive Stress on Pronoun Referent Assignment.
Masters Thesis. La Jolla, CA: University of California at San Diego.
Bardovi-Harlig, Kathleen
1986 Pragmatic Determinants of English Sentence Stress. Dissertation
revision [1983]: University of Chicago. Originally entitled: A
Functional Approach to English Sentence Stress. Bloomington, IN:
Indiana University Linguistics Club.
Baltaxe, Christiane A. M.
1991 Vocal communication of affect and its perception in three- to four-year-
old children. Perceptual and Motor Skills 72:1187-1202.
Berman, Arlene, and Michael Szamosi
1972 Observations on sentential stress. Language 48:304-325.
Bloch, Bernard, and George L. Trager
1942 Outline of Linguistic Analysis. Baltimore, MD: Special Publications of
the Linguistic Society of America.
Bloomfield, Leonard
1933 Language. New York, NY [1962, rev. 1914]: Holt, Reinhart and
Winston.
333
Bolinger, Dwight Le Merton (Kathleen Hubbard 534-8409)
1965 Forms of English: Accent, Morpheme, Order. I. Abe and T. Kanekiyo
(eds.), Cambridge, MA: Harvard University Press.
Bolinger, Dwight Le Merton
1951 Intonation: Levels versus configurations. Word 7:199-210.
1957 Intonation and grammar. Language Learning 8:31-37.
1958 Stress and information. American Speech 33:5-20.
1961a Contrastive accent and contrastive stress. Language 37:83-96.
1961b Ambiguities in pitch accent. Word 17:309-317.
1964 Around the edge of language: Intonation. Harvard Educational Review
34:282-93.
1968 Aspects of Language. New York, NY [1975, second edition]: Harcourt
Brace Jovanovich.
1972a Intonation. (ed.) Harmondsworth, England: Penguin Books.
1972b Accent is predictable (If youre a mind reader). Language 48:633-
644.
1977 Meaning and Form. English Language Series, Title No. 11. Randolph
Quirk (gen. ed.). London, England and New York, NY: Longman Group
Limited.
1986 Intonation and its Parts: Melody in spoken English. Stanford, CA:
Stanford University Press.
1989 Intonation and its Uses: Melody in grammar and discourse. Stanford:
Stanford University Press.
Bresnan, Joan W.
1971 Sentence stress and syntactic transformations. Language 47:257-281.
1972 Stress and syntax: A reply. Language 48:326-342.
Burnet, James (Lord Monboddo, also Burnett)
1774 Of the Origin and Progress of Language (Vol. II). R. C. Alston (ed.),
English Lingustics 1500-1800 (A Collection of Facsimile Reprints, No.
48). Menston, England [1967]: The Scolar Press Limited.
Butler, Charles
1634 The English Grammar, or the institution of letters, syllables, and words
in the English tongue. A. Eichler (ed.). Halle A. S., Germany [1910]:
Max Niemeyer.
334
Capelli, Carol A., Noreen Nakagawa and Cary M. Madden
1990 How children understand sarcasm: The role of context and
information. Child Development 61:1824-1841.
Carnap, Rudolf
1956 Meaning and Necessity: A study in semantics and modal logic. Chicago,
IL: The University of Chicago Press.
Chafe, Wallace L.
1973 Language and memory. Language 49:261-281.
1974 Language and consciousness. Language 50:111-133.
1994 Discourse, Consciousness, and Time: The flow and displacement of
conscious experience in speaking and writing. Chicago, IL: The
University of Chicago Press.
Chomsky, Noam
1955 Transformational Analysis. Dissertation: University of Pennsylvania.
1971 Deep structure, surface structure, and semantic interpretation. In
Semantics, by D. Steinberg and L. Jokobovits (eds.). Cambridge,
England: Cambridge University Press, pp. 183-216.
Chomsky, Noam, and Morris Halle
1968 The Sound Pattern of English. Cambridge, MA [1991]: The MIT Press.
Chomsky, Noam, Morris Halle and Fred Lukoff
1956 On accent and juncture in English. In For Roman Jakobson: Essays
on the Occasion of his Sixtieth Birthday. M. Halle, H. G. Lunt, H.
McLean, and C. L. van Schoonveld (compilers). The Hague, The
Netherlands: Mouton, pp. 65-80.
Coleman, H. O.
1914 Intonation and emphasis. In P. Passy and D. Jones (eds.). Miscellanea
Phonetica I: to commemorate the 25th year of Le Matre Phontique,
pp. 6-26.
Cruttenden, Alan
1986 Intonation. Cambridge Textbooks in Linguistics Series. Cambridge,
England: Cambridge University Press.
335
Dane, Frantiek
1960 Sentence intonation from a functional point of view. Word 16:34-54.
1966 A three-level approach to syntax. In Travaux Linguistiques de Prague
(Vol. I), F. Dane et. al. (eds.). Tuscaloosa, AL: University of Alabama
Press, pp. 225-240.
Diesing, Molly
1990 The Syntactic Roots of Semantic Partition. Dissertation. Amherst, MA:
University of Massachusetts.
Draper, M. H., Peter Ladefoged and D. Whitteridge
1957 Expiratory muscles involved in speech. Journal of Physiology
138:17f.
1958 Respiratory muscles in speech. Journal of Speech and Hearing
Research 2:16-27.
1960 Expiratory muscles and airflow during speech. British Medical
Journal, 18 June:1837-1843.
Fauconnier, Gilles
1985 Mental Spaces: Aspects of meaning construction in natural language.
Cambridge, England [1994]: Cambridge University Press.
Fernald, Anne
1984 The perceptual and affective salience of mothers speech to infants. In
The Origins and Growth of Communication, by L. Feagans, C. Garvey
and R. Golinkoff (eds.). Norwood, NJ: Ablex, pp 5-29.
1989 Intonation and communicative intent in mothers speech to infants: Is
the melody the message? Child Development 60:1497-1510.
Fernald, Anne, and Patricia Kuhl
1987 Acoustic determinants of infant preference for motherese speech.
Infant Behavior and Development 10:279-293.
Firbas, Jan
1964 On defining the theme in functional sentence analysis. Travaux
Linguistiques de Prague. 1:267-280.
1966 Non-thematic subjects in contemporary English: A contribution to the
problem of central and peripheral phenomena inthe system of functional
sentence perspective. Travaux Linguistiques de Prague 2:239-256.
336
1975 On existence-appearance on the scene in functional sentence
perspective. Prague Studies in English 16: 47-71.
1981 Scene and perspective. Brno Studies in English 14:37-91.
1992 Functional Sentence Perspective in Written and Spoken
Communication. Cambridge, England: Cambridge University Press.
Folkins, John W., and James H. Abbs
1975 Lip and jaw motor control during speech: Responses to resistive
loading of the jaw. The Journal of Speech and Hearing Research
18:207-220.
1976 Additional observations on responses to resistive loading of the jaw.
The Journal of Speech and Hearing Research 19:820-821.
Fowler, Carol A., and Lawrence D. Rosenblum
1991 The perception of phonetic gestures. In Modularity and the Motor
Theory of Speech Perception: Proceedings of a conference to honor
Alvin M. Liberman, by I. Mattingly and M. Studdert-Kennedy. Hillsdale,
NJ: Lawrence Erlbaum Associates, Inc. pp.33-59.
Francis, W. Nelson, and Henry Kucera
1961 A Standard Corpus of Present-Day Edited American English, for Use
with Digital Computers. (Brown Corpus) Providence, RI: Department
of Linguistics, Brown University. In CD-ROM distribution of the
following archive: International Computer Archive of Modern English
(ICAME) Collection of English Language Corpora. K. Hofland and S.
Johansson (eds.). Bergen, Norway [1991]: Norwegian Computing
Centre for the Humanities (NCCH).
1979 Manual of Information to Accompany a Standard Corpus of Present-
Day Edited American English, for Use with Digital Computers.
Providence, RI [revised and amplified edition]: Department of
Linguistics, Brown University.
Fry, D. B.
1955 Duration and Intensity as Physical Correlates of Linguistic Stress. The
Journal of the Acoustical Society of America 27:765-768. Reprinted in
I. Lehiste (cf.). 1967:155-158.
Fudge, Erik
1984 English Word-Stress. London, England: George Allen & Unwin
(Publishers) Ltd.
337
Gay, Thomas
1978 Effect of speaking rate on vowel formant movements. Journal of the
Acoustical Society of America 63:223-226.
Gilbert, Daniel Tiberius
1991 How mental systems believe. American Psychologist 46:107-119.
Gilbert, Daniel Tiberius, D. S. Krull and P. S. Malone
1990 Rejection of false information. Journal of Personality and Social
Psychology 59:601-613.
Gilbert, Daniel Tiberius, R. W. Tafrodi and P. S. Malone
1993 You cant not believe everything you read. Journal of Personality and
Social Psychology 65:221-233.
Gunter, Richard
1966 On the placement of accent in dialogue: A feature of context
grammar. Journal of Linguistics 2:159-179.
1972 Intonation and relevance. In D. Bolinger (cf.), 1972a:194-215.
Gussenhoven, Carlos
1984 On the Grammar and Semantics of Sentence Accents. Publications in
Language Sciences 16. Dordrecht, The Netherlands: Foris Publications
Holland.
Gussenhoven, Carlos, Dwight Le Merton Bolinger, and Cornelia Keijsper
1987 On Accent. Bloomington, IN: Indiana University Linguistics Club.
Halle, Morris, and Jean-Roger Vergnaud
1987 An Essay on Stress. Cambridge, MA: The MIT Press.
Halliday, M. A. K.
1967 Notes on transitivity and theme in English (Part 2). Journal of
Linguistics 3:199-244.
Harris, Zelig S.
1944 Simultaneous components in phonology. Language 20:181-205.
338
Hart, John
1551 The Opening of the Unreasonable Writing of our Inglish Toung: Wherin
is shewid what necessarili is to be left, and what folowed for the perfect
writing thereof. British Museum Royal Manuscript 7.C.VII reproduced
in John Harts Works on English Orthography and Pronunciation [1551
1569 1570]. Part I. B. Danielsson (ed.). Stockholm, Sweden [1955]:
Almqvist & Wiksells.
1569 An Orthographie, Conteyning the Due Order and Reason, Howe to
Write or Paint Thimage of Mannes Voice, Most Like to Life or Nature.
R. C. Alston (ed.), English Lingustics 1500-1800 (A Collection of
Facsimile Reprints, No. 209). Menston, England [1970]: The Scolar
Press Limited.
Hayes, Bruce
1995 Metrical Stress Theory: Principles and case studies. Chicago, IL: The
University of Chicago Press.
Haugen, Einar
1949 Phoneme or prosodeme. Language 25:278-282.
Horn, Laurence R.
1969 A presuppositional analysis of only and even. In Papers from the Fifth
Regional Meeting of the Chicago Linguistics Society. R. I. Binnick, A.
Davidson, G. Green and J. Morgan (eds.). Chicago, IL: The University
of Chicago Press, pp. 98-107.
House, Arthur S., and Grant Fairbanks
1953 The influence of consonant environment upon the secondary acoustical
characteristics of vowels. The Journal of the Acoustical Society of
America. 25:105-113.
Hultzn, Lee Sisson
1939 Seventeenth Century Intonation. American Speech 14:39-45.
Huntington, Richard, and Peter Metcalf
1979 Celebrations of Death. Cambridge, England: Cambridge University
Press.
Jackendoff, Ray S.
1972 Semantic Interpretation in Generative Grammar. Cambridge, MA: The
MIT Press.
339
Johns-Lewis, Catherine M.
1986 Intonation and Discourse. (ed.) London: Croom.
Jones, Daniel
1909a Intonation Curves: A collection of phonetic texts, in which intonation is
marked throughout by means of curved lines on a musical stave. Leipzig
and Berlin, Germany: B. G. Teubner.
1909b The Pronunciation of English. Cambridge, England [1966, fourth
edition, revised and enlarged, reprinted with corrections]: Cambridge
University Press.
1918 An Outline of English Phonetics. Cambridge, England [1960, ninth
edition]: W. Heffer & Sons Limited.
1950 The Phoneme: Its nature and use. Cambridge, England [1967, third
edition, with an Appendix on the history and meaning of the term
phoneme]: W. Heffer & Sons, Ltd.
Josephs, R. A., R. B. Giesler, and D. H. Silvera
1994 Judgment by quantity. Journal of Experimental Psychology: General
123:21-32.
Kent, R. D., and R. Netsell
1971 Effects of stress contrasts on certain articulatory parameters.
Phonetica 24:23-44.
Kingdon, Roger
1939 Tonetic stress-marks in English. Le Matre Phontique III 68:60-64.
1958a The Groundwork of English Stress. London, England: Longmans.
1958b The Groundwork of English Intonation. London, England: Longmans,
Green and Co. Ltd.
Kiparsky, Paul
1966 ber den deutschen Akzent. Studia Grammatica 7:69-98.
Klinghardt, Hermann, and G. Klemm
1920 bungen im englischen Tonfall (Fr Lehrer und Studierende | mit |
Einleitung und Anmerkungen | Von | H. Klinghardt und G. Klemm | B.A.
Lond). Cthen, Germany: Verlag von Otto Schulze.
340
Kopolow, Louis E.
1977 plain talk about... HANDLING STRESS. Plain Talk Series: National
Institute of Mental Health, Division of Communications and Education.
Ruth Kay (ed.), DHHS Publication No. (ADM) 87-502 [1987].
Krifka, Manfred
1992 A Framework for Focus-Sensitive Quantification. In SALT II:
Proceedings of the Second Conference on Semantics and Linguistic
Theory. C. Barker and D. Dowty (eds.), Columbus, OH: Ohio State
Department of Linguistics, pp. 215-236.
Kuno, Susumu
1972 Functional sentence perspective: a case study from Japanese and
English. Linguistic Inquiry 3: 269-320.
1980 Functional Syntax. In Syntax and Semantics (Vol. 13): Current
Approaches to Syntax, by E. Moravcsik and J. Wirth (eds.). New York,
NY: Academic Press, pp. 117-135.
Kuno, Susumu, and Etsuko Kaburaki
1977 Empathy and syntax. Linguistic Inquiry 8: 627-672.
Kunze, Luvern H.
1964 Evaluation of methods of estimating sub-glottal air pressure. Journal
of Speech and Hearing Research 7:151-164.
Ladd, Dwight Robert, Jr.
1979 The Structure of Intonational Meaning. Dissertation: Cornell
University. Bloomington, IN [1980, The Structure of Intonational
Meaning: Evidence from English]: Indiana University Press.
Ladefoged, Peter
1960 The regulation of subglottal pressure. Folia Phoniatrica 12:169-175.
1962 Sub-glottal activity during speech. In Proceedings of the Fourth
International Congress of Phonetic Sciences, by A. Sovijrvi and P.
Aalto (eds.), The Hague, The Netherlands: Mouton & Co., pp. 73-91.
1963 Some physiological parameters in speech. Language and Speech
6:109-119.
1964 Comment on Evaluation methods of estimating sub-glottal air
pressure. Journal of Speech and Hearing Research 7:291f.
1967 Three Areas of Experimental Phonetics. London, England [1975, fourth
impression]: Oxford University Press.
341
Ladefoged, Peter, M. H. Draper and D. Whitteridge
1958 Syllables and Stress. Miscellanea Phonetica 3:1-14.
Ladefoged, Peter, and Norris P. McKinney
1963 Loudness, sound pressure and subglottal pressure in speech. The
Journal of the Acoustical Society of America 35:454-460.
Lakoff, George
1972 The global nature of the Nuclear Stress Rule. Language 48:285-303.
1987 Women, Fire, and Dangerous Things: What categories reveal about the
mind. Chicago: The University of Chicago Press.
Lambrecht, Knud
1994 Information Structure and Sentence Form: Topic, focus and the mental
representations of discourse referents. Cambridge Studies in Linguistics
(Vol. 71). Cambridge, England: Cambridge University Press.
Langacker, Ronald W.
1987 Foundations of Cognitive Grammar (Vol. 1): Theoretical prerequisites.
Stanford, CA: Stanford University Press.
1991a Foundations of Cognitive Grammar (Vol. 2): Descriptive application.
Stanford, CA: Stanford University Press.
1991b Concept, Image, and Symbol: The cognitive basis of grammar. Berlin,
Germany: Mouton de Gruyter.
Lasersohn, Peter
1995 Pragmatic Halos. Univeristy of Illinois: Unpublished manuscript.
Lehiste, Ilse
1967 Readings in Acoustic Phonetics. Cambridge, MA: The MIT Press.
1970 Suprasegmentals. Cambridge, MA: The MIT Press.
Lehiste, Ilse, and Gordon E. Peterson
1959 Vowel amplitude and phonemic stress in American English. The
Journal of the Acoustical Society of America 31:428-435. Reprinted in
I. Lehiste (cf.). 1967:183-190.
342
Leland, Thomas
1757 All the orations of Demosthenes, pronounced to excite the Athenians
against Philip King of Macedon. (trans.) New York, NY [1975]: AMS
Press.
Liberman, Alvin M., Franklin S. Cooper, Katherine. S. Harris, and Peter. F. MacNeilage
1963 A motor theory of speech perception. In Proceedings of the Speech
Communication Seminar (Vol. II, Paper D3). Stockholm, Sweden:
Speech Transmission Laboratory, Royal Institute of Techniology.
Liberman, Alvin M., and Ignatius G. Mattingly
1985 The motor theory of speech perception revised. Cognition 21:1-36.
Liberman, Mark Yoffe
1975 The Intonational System of English. Dissertation: Massachusetts
Institute of Technology. New York and London [1979 revision]: Garland
Publishing, Inc.
Liberman, Mark, and Alan Prince
1977 On stress and linguistic rhythm. Linguistic Inquiry 8:249-336.
Lieberman, Philip
1960 Some acoustic correlates of word stress in American English. The
Journal of the Acoustical Society of America 32:451-454.
1963 Laryngeal activity and the analysis and synthesis of speech. The
Journal of the Acoustical Society of America 35:788 (abstract G2).
1967 Intonation, Perception, and Language. MIT Research Monograph No.
38. Cambridge, MA [1975]: The MIT Press.
1968 Direct comparison of subglottal and esophageal pressure during
speech. The Journal of the Acoustical Society of America 43:1157-
1164.
1986 The acquisition of intonation by infants: Physiology and neural
control. In C. Johns-Lewis (cf.), 1986:239-257.
Lifshitz, Samuel
1933 Two integral laws of sound perception relating loudness and apparent
duration of sound impulses. The Journal of the Acoustical Society of
America 5:31-33.
343
Lindblom, B.
1963 Spectrographic study of vowel reduction. The Journal of the
Acoustical Society of America 35:1773-1781.
Malone, Kemp
1923 The Phonology of Modern Icelandic. Ottendorfer Memorial Series of
Germanic Monographs, No. 15. Menasha, WI: The Collegiate Press
(George Banta Publishing Co.).
1926 Pitch patterns in English. Studies in Philology 23:371-379.
Mattingly, Ignatius, and Alvin Liberman
1988 Specialized perceiving systems for speech and other biologically
significant sounds. In Auditory Function: The neurological basis of
hearing, by G. Edelman, W. Gall, and G. Gowan (eds.). New York, NY:
Wiley, pp. 775-794.
Mehler, Jacques, Peter Jusczyk, Ghislaine Lambertz, Nilofar Halsted, Josiane Bertonici
and Claudine Amiel-Tison
1988 A precursor of language acquisition in young infants. Cognition
29:143-178.
Meredith, Scott Emery
1990 Issues in the Phonology of Prominence. Dissertation. Cambridge, MA:
MIT Working Papers in Linguistics.
Mitford, Jessica
1963 The American Way of Death. New York, NY: Simon & Schuster.
Mller, Johannes
1848 The Physiology of the Senses, Voice, and Muscular Motion, with the
Mental Faculties. W. Baly (translated from the German, with notes).
London, England: Taylor, Walton, and Maberly.
Needham, Mark
1990 Semantic structure, information structure, and intonation in discourse
production. Journal of Memory and Language 29:455-468.
Newman, Stanley S.
1946 On the stress system of English. Word 2:171-187.
344
Odell, James
1806 An Essay on the Elements, Accents, & Prosody, of the English
Language. B. Fabian, E. Mertner, K. Schneider, and M. Spevak (eds.),
Anglistica & Americana (A Series of Reprints, No. 47). Hildesheim,
Germany, and New York, NY [1969]: Georg Olms Verlag.
Ohala, John J.
1983 Cross-language use of pitch: An ethological view. Phonetica 40:1-18.
hman, S.
1967 Word and sentence intonation: A quantitative model. Speech
Transmission Laboratory Quarterly Progress and Status Report 2-3:20-
54.
Palmer, Harold E.
1922a English Intonation: with systematic exercises. Cambridge, England
[1924, second edition]: W. Heffer & Sons Limited.
1922b Everyday Sentences in Spoken English: With phonetic transcription and
intonation marks (for the use of foreign students). Revised by F. G.
Blandford. Cambridge, England [1961; Fifth edition (third revision),
eighth impression]: W. Heffer & Sons Limited.
1924 A Grammar of Spoken English on a Strictly Phonetic Basis. Cambridge,
England [1939, second edition, revised by the Author with the
assistance of F. G. Blandford]: W. Heffer & Sons.
1933 A New Classification of English Tones. Institute for Research in English
Teaching. Tokyo, Japan [1960, ninth printing]: Kaitakusha Pub. Co.,
Ltd.
Pierrehumbert, Janet Breckenridge
1980 The Phonology and Phonetics of English Intonation. Dissertation:
Massachusetts Institute of Technology. Bloomington, IN [1987]:
Indiana University Linguistics Club.
Pike, Kenneth L.
1945 The Intonation of American English. University of Michigan
Publications in Linguistics (Vol. 1). Ann Arbor [1963]: University of
Michigan Press.
1947 Grammatical prerequisites to phonemic analysis. Word 3:155-172.
345
Prince, Ellen
1981 Toward a taxonomy of given-new information. In Radical
Pragmatics, by Peter Cole (ed.). New York, NY: Academic Press, pp.
223-255.
Ripman, Walter
1922 Good Speech: An introduction to English phonetics. New York, NY
[reprinted October 1922]: E. P. Dutton & Co.
Rooth, Mats
1985 Association with Focus. Dissertation: University of Massachusetts.
Amherst, MA: GLSA Publications.
Scherer, Klaus R.
1986 Vocal affect expression: A review and a model for future research.
Psychological Bulletin 99:143-165.
Schmerling, Susan F.
1973 Aspects of English Sentence Stress. Dissertation: University of Illinois
at Urbana-Champaign. Austin, TX [1976]: University of Texas Press.
Schubiger, Maria
1935 The Role of Intonation in Spoken English. Cambridge, England: W.
Heffer & Sons Limited.
1958 English Intonation: Its form and function. Tbingen, Germany: Max
Niemayer Verlag.
Selkirk, Elisabeth O.
1984 Phonology and Syntax: The Relation between Sound and Structure.
Cambridge, Massachusetts: The MIT Press.
Sheridan, Thomas
1762 A Course of Lectures on Elocution: Together with two dissertations on
language; and some other tracts relative to those subjects. R. C. Alston
(ed.), English Lingustics 1500-1800 (A Collection of Facsimile
Reprints, No. 129). Menston, England [1968]: The Scolar Press
Limited.
Steedman, Mark
1991 Structure and intonation. Language 67:260-296.
346
Steele, Joshua
1775 Prosodia Rationalis: or, An essay towards establishing the melody and
measure of speech, to be expressed and perpetuated by peculiar
symbols. Bernhard Fabian, Edgar Mertner, Karl Schneider, and Marvin
Spevak (eds.), Anglistica & Americana (A Series of Reprints, No. 125).
Hildesheim, Germany, and New York, NY: [1971; 1779, the second
edition amended and enlarged]: Georg Olms Verlag.
Stetson, R. H.
1951 Motor Phonetics: A study of speech movements in action. Amsterdam,
The Netherlands: North-Holland Publishing Co.
Sussman, Harvey M., and Peter F. MacNeilage
1978 Motor unit correlates of stress: Preliminary observations. The Journal
of the Acoustical Society of America 64:338-340.
Sweet, Henry
1890 A Primer of Spoken English. Oxford, England: Clarendon Press.
1892 A New English Grammar: Logical and historical. Part I, Introduction,
Phonology, and Accidence. Oxford, England [1960]: Clarendon Press.
Thelwall, John (Esq.)
1812 Illustrations of English Rhythmus. London, England: J. MCreery.
Thompson, Henry Swift, Jr.
1980 Stress and Salience in English: Theory and Practice. Dissertation:
University of California at Berkeley. Ann Arbor, MI: University
Microfilms International.
Trager, George L.
1964 The intonation system of American English. In D. Abercrombie (cf.),
pp. 266-270.
Trager, George L., and H. L. Smith, Jr.
1951 An Outline of English Structure. Studies in Linguistics: Occasional
Papers 3 (G. Trager, editor). Norman, OK: Battenburg Press.
Trehub, Sandra E., Maxine W. Endman, and Leigh A. Thorpe
1990 Infants perception of timbre: Classification of complex tones by
spectral structure. Journal of Experimental Child Psychology 49:300-
313.
347
Trim, J. L. M.
1964 Tonetic stress-marks for German. In D. Abercrombie (cf.), pp. 374-
383.
van Ginneken, Jac.
1907 Principes de Linguistique Psychologique. Bibliothque de Philosophie
Exprimentale IV, E. Peillaube (dir.), M Rivire (ed.). Paris, France.
von Stechow, Arnim
1989 Focusing and backgrounding operators. Arbeitspapier (Universitt
Konstanz, Fachgruppe Sprachwissenschaft) Nr. 6.
Walker, John
1787 The Melody of Speaking Delineated; or Elocution Taught Like Music; by
Visible Signs, Adated to the Tones, Inflexions, and Variations of Voice in
Reading and Speaking; with Directions for Modulation, and expressing
the Passions. R. C. Alston (ed.), English Lingustics 1500-1800 (A
Collection of Facsimile Reprints, No. 218). Menston, England [1970]:
The Scolar Press Limited.
Walker-Andrews, Arlene S., and Wendy Grolnick
1983 Discrimination of vocal expressions by young infants. Infant
Behavior and Development 6:491-498.
Wells, Rulon S.
1945 The pitch phonemes of English. Language 21:27-39.
Wells, W. H. G.
1986 An experimental approach to the interpretation of focus in spoken
English. In C. Johns-Lewis (cf.), 1986:53-75.
Whitbourne, Susan Krauss, Sarah Culgin and Erin Cassidy
1995 Evaluation of infantilizing intonation and content of speech directed at
the aged. International Journal of Aging and Human Development
41:109-116.
The software around which this whole project
revolved was FrameMaker, starting with Version 2 for Unix
circa 1989, and ending with Version 5 for Windows 95. I
cant say enough about the program. It never let me down.
I used Corel Draw 3 (and 3D Design Plus) at certain
intermediate stages of the research when the analysis relied
more heavily upon some fairly elaborate figures to explain
some of Langackers dynamic timeline material, and it was
used for the anatomical sections in Appendix I. Corel Chart
imported frequency tables from Lotus Improv and created
the bases for the graphs in Appendix II.
This document is set in Microsofts version of the
Times New Roman and Symbol fonts (TrueType), generally
in 12pt, and where these fonts did not suffice, I drew upon
characters that I designed for Traces New Roman with
Corel Draw 3.
This dissertation is printed on 24-pound, 100%
white cotton paper.
^