A POETICS OF SECOND-LANGUAGE ACQUISITION’

Robert Ochsner2
University of California, b e Angeles

Today there are many approaches to the study of eecond-language

acquisition @LA), but thu variety is supeficial and msleadmg, for it
dmgdises a n underlying research ideal the controlled expenment. We’can,
in other words, scientifically “know”facta about SLA only after a controlled
experiment replicates and croee validates,& facta Thrs ideal method is
part of the nomothetx trahtion of acrenck, tradrtion whch subord~nates
and ultimately excludes a hermeneutic & &e of inquiry But these two
approaches to acience need not conflict. I
For our SLA reaearch I p r o m h a change: that we alternate between
two e q d hnds of research. A poetrca of SLA teaches us that the h n g
observed-languageis also the motive inferred-why we speak Where a
nomothetic experimental method gets a t the language ob~ect,another
equally good method of research, a hermeneutx diary study for example,
reveals the biases d o u r work.We should promote in our SLA research this
”bilingual atbtude.”

PREFACE: A GLOSSARY OF TERMS

It is unusual to begin a study with a definition of terms. Unfor-
tunately, the history and philosophy of science makes this glossary a
fundamental starting point. If’ nothing else, i t does preclude some
misreadings. In what follows then I have tried to use term as they are
technically defined by philosophers of science and social science. Of
course, this wage 8an vary among authors, and where i t does, I shall try
to make clear fie differences.
1) nomothetic science (the prefix ‘homo” means lawful): This
tradition goes back to Plato. As a research attitude it assumes that there
is one ordered, diacoverable reality which c a d l y obeys the Lam of
Nature. Social scientists in h s tradition fkther assume that Laws of
Human Nature exist. Windelband (1894)first used the word nomothetic

1Throughout t&s paper I use the term “eecond-language wuisition” in a general

eenae which s u b - the distinction between acquisition and learning.
21 wlsh to thank thedDllowing persons (some of whom disagree with me) for their
helpful remarks:Rofeseors Evelyn Hatch,Richard Lanham,John Schumann, and Robert
Stockwell; and among my collegues: Kathi Bailey and Michael Long.

53
54 LANGUAGE LEARNING VOL. 29, NO. 1
(which he distinguished from idiographic science). Recently, Jurgen
Habermas (1971) has coined the similar term ‘nomologwal.’
2) hermeneutic science (literally, “the art of interpretation”):This
tradition also goes back to antiquity, notably the W h i s t s . Reality does
not, in this research attitude, have a single form, but instead varies. It
follows that no single method of inquiry applies to all events. In
particular, human events must be interpreted teleologcally; that is,
according to their final ends. Modern hermeneutics is primarily a
German tradition that grew out of the Protestant Reformation.
3) experimental method: I have used “this label to cover a n y
nomothetic research that a ) attempts to manipulate and/or control
variables; b) attempts to observe events “objectively”or neutrally; and
c) generalizes results in terms of laws or causality. In this senseeany of
the following can be described as experimental: case studies, natur-
alistic observations, and perhaps even diary studies. Of course, the
controlled experimept (longitudinalkross-sectional) is the ideal nom-
othetic method.
4) linearity of the experimental method (not a statistical measure
of linearity): In theory a sequence of experimental methods (pilot study,
case study, observation, and 50 on) that progresses towards the
controlled experiment. In practice, especially in a new field like SLA,
this sequence is rarely adhered to. Research texts in the social sciences
(education, sociology, psychology) imply this hierarchy either by
labelling other research as “quasi-experimental” or “less rigorous,” or
by choosing to discuss only “pure” research tCampbell an& Stanley
1963,Kerlinger 1964, Cronbach 1963,Arnoult 1972,Helmstadter 1970,
Borg 1963,Matheson, Bruce, Beauchamp 1970, K. D. Bailey 1976).The
attitude of linearity can be summarized this way: a controlled experi-
ment represents the “best” method for verifying empirical facts.
5 ) non-experimental method: I prefer this ‘negative’ category
becaLwe other labels suggest a standardized research where, by defini-
tion, no standard exists. In hermeneutic science, a non-experimental
method proceeds: a) without controlling the existing realities; b)
without presumirrg-a researcher’s objectivity; and c) without generaliz-
ing towards causal laws. This work evolves i:to a systematic, if
personal, method of analysis which can be discussed, but not expressed,
as general laws or principles. The term “introspective” research may
apply here (though the phenomenologists have adopted a better word,
Husserl’s dyadic term “intersubjective:” however, I do not use thls
term because many phenomenologists, like Husserl, are ambiguously
nomothetic.)
6 ) circularlity of the non-experimental m e t h d : An introspective
science focuses on self-examination and is in this sense circular or
 OCHSNER 55

reflexive. Habermas ( 1971)and Bernstein (1976) see in psyc)manalyti?lb

the ultimate form of self-reflection. As Bernstein notes, i t requires of
patient and analyst “a depth of hermeneutics.,’ Or as Habermas
remarks, the act of understanding grows through self-reflection because
i t “releases the subject from dependence on hypostatized (i.e. static,
objective, nomological) powers.”
7) meta-theory: This term’ m a n s any system of axiomatic prin-
ciples that guides our work. Nomothetic and hermeneutic science
e a c h . . . represents a meta-theory.
8) other definitions of research methods: Besides the experi-
mentalhon-experimental methods described above, several other dis-
tinctions have been proposed. It may be useful to note these below:
nomothetic (Windelband) idiographic
empirical (heme) rational
quantitative Filstead ) qualitative
icienhfic Maslow) humgnistic
experimental Bruyn ) o-a t iohal
scientifs Husserl phenomenologwal
natural sciences Kant. Hegel, b l t h e y ) GeisteswssenschaRen
(sciences o w e spint)
classical sciences
classical sciences
classical sciences
Scheflen)
Garfinkel)
Wolcott 1
grounded/struct
ethnomethodolo r’
ethnography /ethnology
vertical thinking de &no) lateral thinking
I find none of these distinctions adequate, and though a more complete
list would show the overlap with my terms, the effort needed to make
these connections seems, at this point, unwarranted. For the reader who
desires a more cornplebe background, my bibliography and appendiqes
(at the end of this paper) should prove helpful (se*e also M i t r d and
Kilmann 1978).

THE SI’ATE OF THE ART

There is a basic research question for second-language acquisition

(SLA)studies. It reads like this: All normal children learn a first
languags; not all adults leann a second language-Why? This question,
or some variant, directs our-SLA work. But we are headed in a single
direction, for we ask this question 0s if AN ANSWER will be found. If
we state this attitude as a research goal, it reads: We can ultimately
explain SLA; to explain it only requires, after much preliminary work
has shown the way, some final ingenious experiments.
There exists ie
SLA research another type of single-mindedness,
For the last twenty years the SLA work of Gardner, Lambert,and
other psycho1oQ;lstshas created, by precedent if nothing else, a pecking
order for our work. From diary reports to case studies to controlled
 LANGUAGE LEARNING VOL. 29, NO. 1
experiments, all SLA research subserves the experimental ideal.
Apart fkom h s a n s k y (1976) and a few other researchers, no one asks
why case studies deserve second or third place. The order of things is
obvious. First lay the groundwork, then perform the ‘real’ experi-
ments. Neither c@ we question the need for a typology of SLA factors,
some SLA ‘buiaing blocks.’ Our research proceeds step by step, so too
the methodology of our research.
It is time to ask ourselves some different questions. How success-
ful is SLA research? Have the results of our controlled experiments
been, if wc can agree on this word, conclusive? If not, are the
researchers at fault? Or haven’t they had time enough? SLA re-
searchers have had, at most, only thirty years to develop the field
(Hatch 1978). Perhaps soon we shall catch u p with the other social
sciences. Wite soon I think: compared to physics and chemistry the
social sciences, in general, have made little progress (Cohen 1931,
Kuhn 1970).
In a recent SLA textbook, Ritchie (1978:2) eulogizes medical
science and asks: Who would challenge its successes? But he asks the
wrong question. Rather, why should biology and medicine.Brovide, in
Ritchie’s terms, “an excellent model for. . . secmd-language
teaching . . . and basic research?” Is language acquisition (or learning)
an entity like the human body; can we find anywhere the anatomy, or
taxonomy, of SLA?
Schumann (1978a) lists nine factorsthat influence SLA. But he no
where defines his method of selecting them. Why not five factors, or
fifty? Any taxonomy (or to be exact, a typology) will make a nice poetic
handlist (Znaniecki 1963),but it does not, and here Schumann surely
agrees, today go beyond heuristics. Twenty years ago, when tkaching
methods dominated the profession, SLA factors were different. Do we
really know more now than in 1960; or has only our perspecuve
changed, revealing new and important insights, but nothing that
another twenty years of SLA work will not redefine?
These questions have no end because the wrong meta-theory has
framed us. No matter how we define SLA, one category will fade into
another. Does a cognitive style (Naiman, Frohlich, Stern 1975) differ
from a biological trait (Lenneberg 1967)? Is a social context real, or
does the &server invent one (Schumann 1976a, 1976b, 1978a, 1978b)?
And w h a t separates language learning from acquisition-Krashen’s
monitor or his admirably rigorous attempt to define one (Krashen
1976, 1977a, 1977b, 1978)? Our experimental ideal forces us to
operationalize these terms, b u t it p v e s us no practical system for
u&rstanding them. I use the word “understanding” in a special sense,
to be defined later (Section 11). For m w it is sufficient to claim that
 OCHSNER 57
hypostatized terms (or more politely said, operationalizgd definitions)
are never fully understood.
How we classify SLA exhibits ouibiases and insights. But i t does
not define SLA; rather, it defines our research attitudes. Or more
precisely, we describe what we do by our terminology and methods.
There are now many SLA viewpoints, each represented by a different
method of research. Most SLA researcb divides into case studies and
controlled expetimeq&(Hatch 1978).But other approaches are used,
and some of *ese imply new a t t i t u d e q u a l i t y research (Clarke
1976); psychd&lytic/transpositional research (Guiora 1972),diary
studies (F. Schumanne4nd J. Schumann 1977, Walsleben 1976,
Telatnik 1977,Bailey 1939).This list could go on.
It is by these research approaches that our attitudes towards SLA
can be defined. But our meta-theory, our Sequence from lesser research
to the controlled experiment, demands oE us ultimately a single
point-of-view. In other words, given the research and time, one fact or
set of facts will someday causally gxplain SLA. But only by controlled
experiments will we isolate and then explain these proto-facts, so other
research exists merely to trailblke new research paths (Abel 1948).
If’ we strive for h s research monism, for one experimental set of
fasts and terms and methods, do we not also trivialize our results?
Consider this analogy. To item analyze a test we note how respondents
agree and dwer. Then we improve the test by throwing out certain
items. We discriminate. But unlike that test procedure, in SLA
research we aim for the middle range of agreement. In other words, we
make ourselves o b j e c t i v m r try to assume this neutral attitude-by
deliberately cutting out personal differences until SLA fits an opera-
tional definition . What remains of SLA, it seems to me, no one finds to
bury-
What things do people usually agree on? Simple physical (or in
behavior, physiological) fads we normally recognize without much
argument. People breathe, sleep, eat. Important as these functions
ar-and worthy of experimental research-they remain, neverthe-
less, trivial as measures of our attitudes, our viewpoints. For these
facts tell us nothing revealing about ourselves and how we each think.
In SLA the “good” experimental studies are, in this sense, fact-
ually trivial. Some facts of language we can isolate; the morpheme-
acquisition s t u d - (both longitudinal and cross-sectional)show that a
narrow research focus may produce important factual results (Bailey,
Madden, Krashen 1974,Dulay and Burt 1974,Larsen-Freeman 1975,
1976,Rusansky 1976).But even at this simple, if important, level, the
complex variables of SLA make conclusions tenuous:
58 LANGUAGE LEARNING VOL. 29, NO. 1

Thus, we conceive the order of acquisition of English grammatical morphemes as

resulting from an interplay of at least two factors. One factor. consisting of
variables such 08 frequency and salience. seems to direct the order of acquisition
toward a universal order. But a second factor, transfer from the native l a w a g e ,
modulates the order so as to produce differences between learners of different
‘languagebackgrounds. (Hakuta and Cancino 1977:308-309, my emphasis)
Are we to have as many (universal) orders of English SLA as there are
languages? Or even categories of language learners? Hakuta and
Cancino do not make a strong claim here.
But they do have a tesdble hgypothesia Bigger studies, in effect,
do not. As we manipulate larger and more complex areas of SLA, the
experimental studies become, to borrow Pirandello’s phrase, char-
acters in search of qn author. As Gardner and Lambert (1972) have
found, Louisiang is not Conneticut, and neither SLA context adds up to
Maine. Thesreal issue here concerns basic units. How accutately can
we define these terms: attitude, motivation, learning, acquisition,
language, and so on ?
If chemists juggled their basic units like we do, their laboratories
would blow up. But more damagmg, and less debatable, is the obvious
fact *thatresearch design and the ‘real’world only sometimes co-vary.
We trade off internal for external validity, or vice versa; either way,
we obtain in our experiments important results only from those small,
and trivial, bits of human reality that allow a reductive analysis.
I have avoided, so far, defining what I mean by “important.” A
study is important if we agree that i t is. Since every author reviews the
literature he/she likes, or dislikes, the important research shapes our
attitudes; it is our individual and collective SLA signpost. Although we
cannot measure tbis influence, we can recognize and discuss it. In this
sense my motive, my theme in this paper involves a shift from w h a t we
normally consider important (i.e., experimental research) to other
forms of research (i.e., non-experimental). To superpose these values,
in other words, for you to consider egperimental research as i f it w e r e
“good” and “bad,” I must attack the prevailing bias.
By this “attack’ I do not mean to berate experimental science or
those who practice it. Rather, I find our attitude towards i t s use worthy
of attack. As I shall argue in the last section of this paper, we must use
the experimental approach to science, but not as a meta-theory for all
of SLA research.
However, this attack does raise two questions. (1) Is there a
prevailing bias for experimental research; and (2) what, if anything,
can replace it? To answer the first question I could spend a week
reviewing old journals. Chart the SLA trends: 45% cross-sectional,
15% longitudinal, a shopping bag of leftovers-IO% contrastive anal-
ysis, 10% other linguistic analysis, 05% tests, 15% odds and ends.
 OCHSNER 59

That’s’excluding all the teacher-talk articles. Byt counting the studigs

begs its o w n scientific dilemma: how tm quantify a n attitude? S p a -
lation doesn’t end with a firm measurement. A survey instrument,
after all, is no better than the questions asked. And controlled
experiments, if the research is done properly, must be narrowly
constrained. In any case, what do the experimental s t u d e s tell us?
How many of them offer conclusive, evidence? Try measuring that.
Counting back-issues equals a wet finger held to the wind. Maybe
the forecasts coming in from other areas will influence us (e.g., see
Mitroff and Kilmann 1978). But I think not. Maybe o y own dissatis-
faction will lead us to change research directions. I hope so.To this end
I shall overstate my review of our profession. Why, despite its often
tentative and discouragein) applications, its questionable designs, i t s
lack of replicated studies, i t s confusing terms-why do SLA researchers
persist in their experimental work? I can, of course, produce evidence for
each of these charges. But I cannot here review all the defects of
experimental studies in SLA. Let me summarize my position this way.
Undoubtedly, some good work has been done; but how many SLA
experiments have sound research designs, correct stati ical measures,
and convincing results? I have asked this question se era1 times. I
would still like to know the answer.
%
I should note that rese’pch is a process, and even.bad or inade-
quate’studies may teach. us something. And if a select few are first-rate
experimenters now, future SLA researchers will, of course, do even
more things right. U we believe this ethodolog?cal positivism, what
f
evidence is there in other, more tra itional disciplines for a purer,
more rigorous social science? For example, Cohen (1931:250),one of
the earlier apologists for social scieTe, claims that i t cannot “readily
be put on par with,the physical h e n c e s with regard to definiteness
and universal demonstrability.”’Hd cites the “greater complexity of
social facts” as making social science research less successful. Half a
century later, Rivers (1975) asserts that no Laws of Human Behavior
have yet been found.
From Laws we can retreat in our SLA research to human ten-
dencies. But this position raises extreme problems. While not arguing
statistical measures of correlation, the psycholopst David Lykken
(1967:158) concludes:
the finding of statistical significance is perhaps the least important attribute of a
good experiment; it is neuer a sufficient conhtion for conduding that a theory hao
been corroborated, that a useful empirical fact has‘been established with
reasonable confident-r that an experimental report ought to be published. The
value of any research can bq determind;not from the statistical results, but only
by skilled, s u b m v e evalyation...
 LANGUAGE LEARNING VOL. 29, NO. 1
Lykken, and another psychologmt, Smith ( 1970) also remark that
studies in their field are rarely corroborated. Researcherenot the
research m e t h o d i a r e to blame for these “once-only” experiments. But
in a devastating review, the psychologist Deese (1968)dismisses as
trivial the findings of even the cross-validated and replicated studies.
He concludes: ‘‘ In the study of the higher mental processes, despite
nearly a century of investigation, the results have been little more
than a long history of doubt, frustration, and trivial generalities
(517).”
If we regard psychology as the most “scientific” of the social
sciences, then w h a t success can the SLA profession expect to have?
And if historical trends show us the future, then our mother disci-
p l i n e d u c a t i o n , linguistics, sociology, and psychology-appear to-
day less and less inclined t~ be experimental. The newer trends, apart
from linguistics, are: educational ethnography (Wolcott 1975), ethno-
methodology (Garfinkel 1967), and psycholoscal humanism (Severin
1965).
In SLA we have done very little work using these alternative
research methods. Moreover, our single-minded experimental ap-
proach leads us, by analogy I think, to a one-model theory of SLA.
Vying for a comprehensive SLA model, in recent years there have been
Winker’s interlanguage, Schumann’s acculturation, Krashen’s mon-
itor, Lamendella’s neurofunctional variables; perhaps as many com-
prehensive or basic models as there +re SLA researchers.
I applaud the variety. But not the implicit attitude that one model
explains more than others. In any event, I expect no one will squeeze
SLA all into a single model,’ not one that works causally. Our
experimental method does not help us “squeeze,” for by ultimately
recognizing one method only, we exclude other means of drsprourng
models. We could, for example, use what Burke (1967)calls a “per-
spective by incongmity;” that is, to obtain and test our finaings we
could translate them into “strange” method$ that are outside our
normal view of language. This shift could further involve a transfer of
models. Why not, for SLA, a game theory (Caillois 1961, Huizinga
1950);or Burke’s dramaturgical theory of motives (Burke 1965)?
TOmove easily within our discipline, oscillating between methods
and models, gives to SLA a new meta-theory, not unlike a metaphorical
theory. In this $LA metaphor, each researcher designs hidher ,own
poem or statement of reality; and we read this poem with a critical sense
for the text, the author,and our selfanacious response to both. Earlier
in this paper I asked what, if anything, could replace the experimental
ideal. I shall now offer a n answer.
What we should have for SLA research is the means to alternate
 OCHSNER 61
between two kinds of equal research; one for objective, physical data
and one for subjective, unobsemable facts. The thing observed, lang-
uage, is also the motive inferred, why we speak. A peetics of SLA
teaches us to oscillate between this form and content. Where an
experimental method gets at the language object (e.g., Scovel 1977),
another equally good method of research, a diary study for example
(see Bailey 1979),reveals the biases of our work. (The above distinc-
tiom-objective/subjetive, thing/motive, form/content-are not en-
tirely accurate. I will discuss this problem below, and especially in
Section 11).
Man is a complex symbol-user, and language is the basic tool he
use8 to symbolize everything. Language is a very big topic. SLA
comprises obviously as big a field of study. It should not surprise us
then to find at least two possible ways to investigate a first or second
language. If we study it as a simple entity, a “body of facts,” we can
develop by controlled experimentation a good sense of the thing that
language is. But if,we only look at its physical use, we miss under-
standing the talker’s symbols, hisher metaphorical use of language. It
is possible to study eithet thd language thing o r the language meta-
phor; our research attitude determines this view or way of looking.
Therefore, at another research level, we must look at our looking, and
then through our point-of-view. And then at and through, perspective
and reality in constant oscillation (Lanham 1977).For the researcher
shapes his re&h findings,a fact that has been experimentally, and
ironically, shown (Rosenthal 1977).
This athhrough perspective is itself, remember, a metaphor-a
kind of applied poetics. Paula Kurman (1977),a speech therapist, gives
the nicest statement of what-I mean, in practice, by this “bilingual”
attitude. h sum, I can only steal her remarks:
It is useful b read of developments in fields other than one’s own
I t 18 useful to ‘ b r r o w ” question structures and research techniques from
other fields.
Is m uaeful to seek out perspectrves of t h e whoee work interests are Merent
from one’s own.
It LB uaeful not to restnct such eolicita~oneto the scientdic &aciplines b s t s .
business people, and alert children have i n k r e s b q perceptual frameworks, too
Remember that it was a child, aa yet not thoroughly enmeshed in the eoclal
construct of reality of lus elders, who saw and called the Emperor naked

How do we begin this poetics? To answer that question I must first

discuss what a poetics is not. It is not the prolegomenon of collected
SLA verse, or some other avant-garde art. There is no body or “theory
of literature” here; nothing like Chatman (1972) includes in the
Apprqeches to Poetics. Neither does it pose yet another humanist
versu4 scientist debate. I chose the term poetics because i t jam askew
62 LANGUAGE LEARNING VOL. 29;NO. 1

our ordinary view of language research. I t also makes us just a little

self-conscious. For no matter how objective we intend to be, there
remains in our SLA work an underlying bia-he research act itself.
We must choose to do researchm a c e m i n way, and by this choice we
act out a motive (Burke 1965, Campbell 1975). Call it the advance of
human knowledge. Or call it getting tenure. Whatever ourpurpose for
doing research, it is also a prejudice. One meaning of poetics then is for
us to recognize in SLA research, and then to discuss, our inchoate
biases.
I also find this te& handy because i t s common literary usage can
vary so. Literary poetics has not solved the poem/prose issue, and it
never will show empirically where a poem ends and ordinary language
begins. The distinction lies in our attitudes (Nemerov 1978).Thus, the
term “poetics” suggests a hidden metaphor, an’elastic label, one that
joins things together and blds, especially in this brief paper, whatever
I cannot join easrly by force of outline.
As I shf31 &gue hereafter, there exist two research traditions, the
nomothetic and the hermeneutic. These latter terms, barbarous
enough for any jargon, I shall ulbimately and happily conflate into a
poeticgof SLA.Of course, what I define as poetics we can , and should,
debate. So another broader meaning inheres to this term,for poetics is
the meta-metaphorical theory of SLA, the study of how and what we
study.

TWO RESEARCH TRADITIONS

AS language researchers we have a choice, one tjmt all social
scientists make, one that guides our work and our lives fundamentally:
we can choose either to explcrin or understand reality (Palmer 1969,
von Wright 1971,Winch 1958).Of course, not everyone consciously
chooses to explain or understand things. In SLA I expect most
researchers have unquestioningly done the former. In any case, the
first option--to explain-follows a single method of inquiry, the
nomothetic (Windelband 1894,von Wright 1971, but also see Ryan
1970).In nomothetic science we systematically describe SLA and its
general features in order to someday predict how it will occur. bality’s
lawfblness, its order, allows us to deduce whatever ‘causes’ SLA. And
to describe this order, to predwt i t depends on our ability to control
reality, to ‘stop’ i t for our experiments.
The second choice-to understand SLA-humbles us a little. If
h w s of Reality (a point of faith, not fact) exist, then these laws apply
only to the physical world and not to our behavior IBurke 1969, Harre
 OCHSNER 63

1976).To paraphrase Descartes, people think reflexively. Although his

maxim “Cogito. ergo sum” is not logically satisfactory, it does
emphasize a n importaqt point-man’s ability to think df-consciously .
In a related sense, what we know of SLA is circular (depending, as it
does, on what we know of SLA).Indeed, our r e s e w describes the
researcher as much as the SLA ‘text’ (Crowle 1971, 1976). For
example, in discussing experimental research Crowle (1976:172)notes:
Individual human subjects are psychologmdly heterogeneous A group of sub-
may respond to a treatment in a uniform way, yet each subject’s response may
h a v e resulted From the interaction of quite different combinations of independent
and intervening vanables. When the value of the vanables are unknown, they
must be assumed, and different assumptions lead to different interpretations of
the experimental results because a n expenmenter designs h e treatmenta in order
to test the effects of a specific variable, and hence praserats a brased tnterprehfwn
oQ~isexperiment If, as seem likely, the experimenter’s description is based on a
fake belief that his research workers M o w e d their instructions exactly, his report
is inaccurate If, as is inevitable in complex experiments, the experimenter omits
details which would support a l t e r n a t i v e interpretations, his report is seriously
incomplete. In short, reports of social peychologwal laboratory expenmenta are
very deceptive (My emphasis )
These shortcomings apply to much SLA research; that is, to under-
stand the “SLA text” we must know i t s authors, the stage they enact,
and how these dramaturgical events come to be (Burke 1965). In
non-literary jargon, we must do case studies or ethnbgraphies, and
write diaries.
Our research should therefore interpret what SLA means to us,
and what we think it means to others. W e may each understand SLA
in the same way that a critic understands a poem-no single reading is
right; the SLA text has no immanent Truth. What we can do, what
we do already, is to debate the SLA text, argue why one interpretation
outdoes @her. As the anthropologist Geertz (1973:29)puts it:
“Anthropology, or at least interpretive anthropology, is a science
whose progress is marked less by a perfection of consensus than by a
refinement of debate. What gets better is the precision with which we
vex each other.” In our profession this debate enriches our under-
standing of SLA and ouneelves more so if we look variously, if we apply
many methods of inquiry and not a single nomothetic view. Four
hundred years hence we shall not. find the alchemical Cause that
transmutes us, or the bilingual Eff& that obviates SLA research. But
we m a y learn to ask better questions-the key to understanding (and
explanation).
We have, then, these options: Either we generalize results from an
ideal research method in order to explain SLA;or we limit ourselves to
particular events and try,usually by reconstructing these events, to
understand SLA. In other words, we can do experiments or write
 LANGUAGE LEARNING VOL. 29. NO. 1

histories. I have, of corn, ignored a third possibility-to combine

these methods. But even if a longitudinal experiment can be a history,
and a history ‘scientific,’ the work performed is, nevertheless, per-
fotmed differently. The experiment follows a rigid procedure; thus, i t
cap be repeated. But a history mirrors the writer’s attitudes and
biases; it also reflects the author’s mind, and not the historical eventts)
per se. Histories are not repeated experimentally, they are recon-
structed gubjectively’.
A poetics of SLA cannot fuse opposites. To force these methods
together would only bastardize them; i t would also be wrong to forego
systematic research a n d do whatever amuses our “eclectric” whims.
Instead, I propose that we divide SLA in half, though not necessarily
equal halves. In one sphere we shall try to explain SLA as a ‘physical
entity,’ in the other we shall try to understand it*as symbolical
motives. But this dual research approach requires of us some mental
a w s t m e n t s . As Ctowle (19761163)remarks: “Because [experimental
and ethnographic research] are so different, it is impossible to work or
think in both styles at once, but it is possible with careful gear-
changing to alternate between them.” I t is our research attitude that
directs us to either, approach. By learning to see these attitudes, we can
then Rpwer the creative tension of alternating viewpoints.
To do an experiment or a history we should hgve some research
guidelines; I cannot, however, state precisely which SLA topics should
be explained and which others understood. Like demarking prose from
poetry, the dividing line can v a r y . But i
h SLA research we can agree to
separate language acquisition from language use. This bifurcation
serves to emphasize our viewpoints: either we see language as an
object or as symbolical motives. Obviously these two viewpoints
overlap, and we should never fully separate them (especially in a
single method meta-theory like we have in our profession today). To
repeat my basic point, we can and should analyze the SLA entity as
something different from the symbolical motive. And vice Vera. To
explain what I mean by these two approaches, to define them (albeit
not exactly), I will discuss below their traditional meanings.

NOMOT.HETIC SCIENCE
Nomothetic?means the scientific attitude of lawfblness, of “causal”
relations- (Windelband 1894, von Wright 1971).It has the-additional
meaning of applying only to physical nature, but for reasons I make
clear in the final section of this paper, I will eventually apply the term
to humans also. The hstory of this nomothetic attitude goes back a t
least to Plato, 4th century B.C. (for a historical review of nomothetic
science see Appendix A). In his Republic and other dialogues Plato
 OCHSNER 65

eloquently stated the nomathetic principles: There exists one Lawful

(or pure) reality and a single way of knowing it. The flux of our daily
lives obscures this lawfulness, so to see the Real World a n observer (or
philosopher-kin@ must be specially trained: Mathematics te.g., geo-
metry) and abstract reasoning (e.g., logic) are the cornerstones of this
training. As he develops his intellect, the trainee-observer must also
learn to control his ‘lesser’ faculties: appezite, emotion, subjectivity.
This control extends to ordinary language and experience, which he
should always distrust. Plato was himself a talented writer, and the
eloquence of his dialogues raises some difficult problem fqr classifying
his attitude towbds literary/ordmary language. However, in his
earlier works (e. g., The Gorgias: A Study of False Rhebric) Plato
obviously rejected rhetoric as a vehicle for conveying truth. However,
he did modify his position in a later work; Phedrus: A Study in True
Rhetoric.
As SLA “experimenters” we have tacitly approved, b m h e applied
work we do, these nomothetic principes:
One world: There is one “lawful” reality
One order: To explain t h e reality we deduce c a w s from the Laws of Nature,
includmg Laws of Human Nature
One me&& There is one best research method. thecontrolled experiment
This methodolopcal monism represente the unity of the scientific
approach in contrast to the flux of reality, or in keeping wth the
stahc world to he investigated (von Wright 1971)

I doubt that many SLA researchers, if asked, would offer these

principles. But if asked to rationalize their experimental work, these
logical axioms are necessary (though perhaps not sufficient).
These principles -mean that someday we can explain SLA. It
follows from this premise that we can predict events, and we can
predict events because there is cause and effect. So to explain SLA we
must control its causal variables. The controlled experiment remains
the one ‘best’ way to do this. In other words, the linearity of the
experimental approach to science (see Preface) culminates as we
cross-sectionalize, or stop, the SLA world in order to explain it.
To do his experiments properly the SLA researcher, like Plato’s
philosopher-lung, undergoes special training. He (or she) almost al-
ways recieves a n M.A. or Ph.D. as the‘ badge of his research compe-
tence, and in his work he must procedurally display this badge. He
remains personally objective and standardizes this attitude for others
to imitate. In effect, he follows the rules of ‘pure’ re-ch (Ackoff
1962).
Plato’s utopia has a modern, scientific reading. There exist Laws
of Physical Nature that only ‘pure’ research can explain. These L a w s
are predictable because they have causes. But causality deceives the
66 LANGUAGE LEARNING VOL. 29, NO. 1

noh-scientific layman; for him, a knockout punch and the ten count
add up to a causal chain. The researcher, though, looks beyond a single
punch; as a nomothetic scientist he learns to control experience-and
himself.. He designs, tests, and replicates his work; he uses computers,
statistics, and hard data. Above all, he does not care who wins the
(ontological) fight. Objectively then, he infers the boxer’s probable
talent: to every action there is an equal and opposite raaction. Of
course, he doubts his finding.
The SLA researcher can do all this-all that Plato imapned-
because a single reality exists. His findings, especially replicated ones,
build his confidence ‘in this single reality. But he always doubts his
findings. No experiment can be truly proven because no one can repeat
a n experiment ad infinitum (Popper 1963, 1965).He also distrusts
words and their uncontrollable nuances. Hard (statistical)facts should
not flirt with meaning; instead,.they should fix semantically into
place, like building blocks. When he must use words the SLA re-
searcher expresses himself “non-personae;” or a t least attempts to
(Campbell 1975).But he prefers statistics, letting the numbers speak ‘

for themselves.
This nomothetic attitude involves a specific-pattern of thought, or
reasoning process, that is behaviorally acted out, most typically, as
follows:
Model 1: Nomothetic method ( a priori ad a posteriori)

1
Describe ,Causal

I event A,B

Predict
event C
Linear sequence
( hypothetico-deductive)
theory tA,B

Explain
&event C

Here, based on the chart above (Model l ) ,is a simple explanation, but
one that refers to a n ordinary-non-linguistic-event; (at this point I
- Cl

do not wish to compare the research domains of SLA):

You have parked your car outside. It is a winter night and the
temperature drops well below Zero. You have no anti-freeze in the car’s
cooling system which, instead, is filled with water. The next morning
you can predict that your radiator is damaged and your engme block
cracked. Such a prediction fits the laws of physical reality, in this case,
the expansion of water when i t freezes.
Now back to the chart (Model 1). Event “A” is water ( H 2 0 ) in one
of three stable forms-gas, liquid, ice. Event “B” is a change in
terfiperature. M y radiator example (actually I have borrowed it from
von Wright who took it fi=w Hempel) is event “C,” the result of A,B.
We explain the broken radiator by noting that water becomes ice at 32”
F.The whole explanation begins with causal theory, for we expect to
find a predictable and linear chain of events.
 OCHSNER 67

opposite direction. I think this change is wrong, but unimportant

(Popper 1963).From fact to theory or vice versa, the causal sequence
remains linear. It is this linearity that makes cause and effect
plausible, but not provable. Indeed, the tapic of c a u s a t i h has no
philosophical end. (For a thorough discussion of causality see von
Wright 1971, especially Chapter Two; see also Ryan 1970. Of course,
the whole modern issue begins with Hume’s An Enquiry Concerning
H u m a n Understanding, 1777.) Causality does, however, resolve the
problem of order. We can say “this happened, because that had
occurred” (von Wright 1971).A cause, even one not directly observed,
must exist; we cannot, otherwise, explain “how” something is or w a s
possible .

HERMENEUTICS
Hermeneutics (the s is option$) means the “art of interpretation”
(Doty 1972, Funk 1966,Palmer 1969).Its first formal practitioners
were Biblical scholars, notably the post-&formation Protestants ( c .
1650: for a historical review of hermeneutic research see Appendik B).
But the principles of this a r t - o r s c i e n c r x t e n d well beyond, and
before, Scriptural exegesis, for the practice of hermeneutics surely
p r e d a t e s recorded history (Palmer 1969). Saints, holy men, Old
Testament leaders have all, at some time, interpreted an other-worldly
Truth; and to translate from one context to another, from God’s truth
to man’s ignorance, requires of these men (or ultimately their readers)
a hermeneutic performance (Funk 1966). Reader and translator per-
form together because each must interpret the bias of a translator’s
words. Furthermore, to understand the text, these performers must
estimate how well it accommodates a n audience.
This hermeneutic translation continues 0s i f reader, translator,
and text were a circle:
For the interpreter to “perform” the text, he must “understand” it, he must
preunderstand the subyxt and the situation before he can epter the horlzon of i t s
meaning. Only when he can step into the mag~ccircle of i t s horizon can the
interpreter understand i t s meaning This is that mysterious “hermeneutical
c m l e ” without which the meanmg of the text cannot emerge. (Palmer 1969:%)
This rather mystical idea has a more mundane reading. In discussing
ways to interpret Middle Eastern culture, the anthropolopst Clifford
Geertz puts it thus: we must somehow find our footing (1973:13).
This ontological1 problem of interpreting reality goes back to
antiquity. The Sophists, arch enemies of Plato, believed that many
‘artifical’ realities exist; no single world underlies them, so nothing
absolute can be idealized. The Physical world, with its N a t u r a l Laws
(physis), differs froin social artifiue (the polis). We act by following
68 LANGUAGE LEARNING VOL. 29, NO. 1
social conventions or rules, and it behooves us to see these rules as
practical means to practical ends. If our purpose is not fuifilled, then
we should change the rules (Barker 1964, Morris 1967).
What we know of the Sophists, of course, comes largely from
Plato’s dialogues. Hardly a n “objective” source, buk on this issue of
objectivity, the Sophists would not be much surprised or alarmed by
distortions. They earned their livelihood because people see things
differently. As counselors to the Greek aristocrats, they trained their
customers to interpret city-state laws. They also taught the subleties of
persuasion, of legal salesmanship. This training evolved as the art of
rhetoric, a code of rules-not immutable Lawsbf Nature-that prepare
u s to shift realities, to superpose our world view onto another (Lanham
1969, 1976,Golden, Berquist, Coleman 1976).
If‘reality can vary, for humans if not for objects, then explanations
have for us no absolute meaning. W e cause the caused. In other words,
people can act, and they often do, with a purpose in mind (Burke 1969).
To ask what causes this purpose throws us into a hermeneutical circle
(Palmer 1969).The key idea here is purpoge. Remember that in
nomothetric science we can explain something because we can predict
it. But explanation means nothing if we cannot predict a cause and
effect sequence for human acts. Our behavior fouls the nomothetic
chain of events. Every act defines us because it has a meaning we s v e
;it. Not only does this meaning change from place to placqand time to
time, it also varies between persons and within the same individual
(Goffman 1959).
Physical nature obeys nomothetic laws. Animate nature, because
i t elrolves, obeys more than just physical laws. For example, if we
stretch the idea of purpose to include evolution, then all life has
purpose: it exists to go on living (Barash 1977).Of wurse, we do not

P
find a m o e b a reflecting on their meaningfulness; on Hamlet and
perhaps the higher primates can ask why and why n t. But purpose
and meaging, however slight distinguish all life.
Aristbtle, a pupil of Plato, taught that all life functions teleolog-
ically. Rather than look for i t s causes, he argued,we should note final
(teleogical) ends. By attributing purpose to an event, we can therefore
hope to understand why something happened. A causal explanation, on
the other hand, shows how an event occurred. In other words, we don’t
say a radiator fieezes because it wants to-crack and die. It. has no
Purpose. ‘

W e can, however, ask why animate behavior is meaningful.

Barash (1977)gives the example of,a squirrel and a dog. The nomo-
thetic researcher predicts that a dog, because it has greater intelli-
gence (operationalized as brain mass), will outperform the lowly
squirrel. Their’task is to reach some food just beyond them. Each
 OCHSNER 69

animal has been secured to a long leash, half of which the experi-
menter winds around a swke. By going backwards--away from the
food-the animals can eventually unwind themselves and reach the
food. The squirrel wins.
Without upraveling evolution, the nomothetic researcher cannot
explainsthis result. However, had he taken the “squirrel’s point of
view” the result makes sense. Unlike the dkg, the squirrel lives in
trees where the branches do not connect one to another. So the
squirrel, a tree dweller, acquires the meaningful talent of retracing its
steps; wanting, I suppose, to get a c b s t n u t elsewhere.
By noting the teleological end, a hermeneutic researcher can
interpret this event: And the nomothetic researcher, with a little more
ingenuity, can redesign this experiment for three dimensions. N h o -
thetic science goes on. But consider human behavior as an endless
range of dimensions, of motives. Our behavior overlaps and surely
trends exist, but there is a wild card that makes even our com$exities
insignificant. People are self-conscious. Dogs and squirrels go for the
food; people can say why bother.
Human behavior differs in kind from other phenomena (Palmer
1969, Wihch 1958,but see a l q Ryan 1970, Abel 1948). There zke many
worlds in behavior, not one; anid thme are unlimited motives, not a
single causal order. So to understand a human event, a linear (norno-
thetic) method will “bend” as the interpreter recycles it. I have
diagrammed this circular process below (cf. Model 1):
Model 2: Hermeneutic science (a posteriori ad idiniturn)
Recall, ,Teleologrcal
event A,B theory A.B/ /C
Circular method
1
Understand +
1
,Interpret
event A,B/ /C event C
Let “A” represent Los Angeles International Airport and “ B ” a man
running through a terminal there. We can impute a teleological end to
his running-a plane to catch; and we can understand this motive, for
most of us have had analogous experience. .Because it is a finctwn of
running to catch planes, we can interpret event “C”, his boarding the
plane, as related to A and B. But A and B are related to C only if our
teleological premise holds. of course, there are many other possible
motives for his running. He could, for example, be running away from
someone. But no matter how many motives we consider, no Human
Law of Plane’Catching guides this interpretation. Neither does our
work culminate in such a Law.
A nomothetic scientist would chart this running behavior and
produce a statistical map. It would tell us where people go, but not
70 LANGUAGE LEARNING VOL. 29, NO. 1
why; and after a short while the map would become a solid blot of
inter-secting lines (Burke 196!5), not a very useful measure. To chart
our motives is an infinitely more difficult, and futile, task. The be$
nomothetic research, when applied to humans, only predicts a world
limited to time and place. Not only is this prediction bound to
history-note that Nty year0 ago people rode trai-ut it also
changes when people change their fin$.

A POETICS OF SLA
Below I have charted the basic points of the previous sections
(nomothetic science and hermeneutic research). I have also included
Oarnes of the @'!Eipl thinkers, and certain trends, associated with
each tradition. Please note that this chart is a very incomplete, and
tentative, outline: Moreover, it should be read as a historical con-
tinuum, one where trends converge as well as polarize.
T w o Research Traditions

+
Nomothetic
+
Hermeneutic

+
Explanation

Prediction:
Understanding
.)
Interpretation:

Plato Sophists
Copernicus Aristotle
Galileo Rhetoricians
Kepler Augustine
Newton Protestant Reformation
18th century science Kant Schleiermacher
& technology Hegel
Cornte Dilthey
Social sciences Bergson
&. Poeitivism Darvvln Weber
Popper Husserl
Wataon Chomsky Phenomenology Heidegger
Behaviorism
Skinner Habermas
(Anglo-French tradition) (Claesical/Cermanic tradition)

Basic &nothetic tenets: Basic hermeneutic tenets:

1) Methodological monism 1) Methodologwal pluralism
(controlled experiments) (interpretive histories)
2) Reality is lawful (Knowledge of 2) Only physical reality is lawful
reality is reducible.to causation (Knowl- of human realitres
becauae we can explain the world differ because thpy must be
by controlled experiments) understad bv personal
interpretation)
3) There is only one reality 3) There are marib realitres
 OCHSNBR 71
These traditions form opposjte parts of a long philosophical
debate. Do people differ from things (vitalism/materialism)? Does the
mind exist separately from the body (dualism/monism)? Is there one
reality or many worlds (realism/nominalism)? For each of these
questions the nomothetic reseracher would answer one way and the
hermeneutic researcher another.
A poetics of SLA makes possible our use of both traditions. W e can
(and should) adopt a physicalist attitude towards SDA, especially when
we consider how language is acquired. Moreover, we can guide this
attitude by noting, with Ebel (1967:82), that the nomothetic tradition
works best when i t is applied to “simple” facts:
Nomathetic science has been highly successful in some areas, such as mechanics,
electromagnetic radiations, and atomic physics I t has been distinctly less
successful in other areas, such as weather phenomena, acoustics, and human
behavior The difference seems to lie in the extent to which there is simple
structure in the phenomena involved-few variables and simple (often linear)
relations among them If the structure is complex, if the variables are numerous,
if they interact complexly, if the lines of relationship have critical repons of
nonlinearity or discontmuity, then the nomothetic net is likely to be sketchy or
nonexietent.
A hermeneutic study gives us something different. We cannot
replicate such a study or generalize from i t s findhgs to some Law of
SLA. But we can note its story-value; that is, how well the author
describes an SLA event. Furthermore, the lopc of a hermeneutic study
can be observed in what von Wright (1971) calls a “practical syl-
lbgism.” If we accept certain “common sense’: premises, then con-
clusions may, or may not, logicalry follow. Deese ( 1968)offers a similar
guideline-we should use our‘everyday common sense in order to
understand human behavior (seealso Antilla 1933, Itkonen 1974).
This everyday m e t h 4 , in fact, requires Cxtraordinary skill. The1
art of translation provides an apt analogy. Ia a simple text, for
example, we can probably substitute one word for another. This
(nomothetic) approach will not, I think, much affect the meaning. But
a difficult passage, a poem for instance, does not equal X number of
words in another language. Indeed, the poetic form ofZen conveys
p e a n i n g in addition to the words. Thus, no word for word translatidn
will express the poem’s same, or even its approximate, meaning.
A poetics of SLA lets us shift our perspectives; we can, in other
words, view SLA as a simple or a complex event. This viewpoint is our
research attitude, for i t determines how we‘“trans1ate” the SLA text.
As a meta-theory, poetics teaches us to look ut SLA; however, we must
also learn to see through it. L k e the translator, the SLA researcher
must consider two realities: the hermeneutic and the nobothetic. We
need to develop, therefore, a perceptual bilingualism; that is, the
ability to study SLA from at least two points of view.
72 LANGUAGE LEARNING VOL. 29,NO. 1

But if attitudes have words, then SLA researchers speak but one
languagethe experimental paradigm. I assume we exist to promote
bilinguals. How strange then, knowing that two languages Wer more
than one, yet today SLA reseachers continue to use but one pattern of
thinking, one attitude, the nomethetic.
Too few researchers in SLA h a v e tried other, non-experimental,
methods. Clarke (1976)and Schumann and Schumann ( 1977) signal, I
hope, the future. Clarke’s study blends hermeneutic insight with
phenomenology, Eastern religion, schizophrenia theory and “modern
consciousness.” In other words, he metaphorically frames an attitude
and then unveils SLA as a “clash. of consciousness.” More recently,
diary studies have been used to probe what is psycholopcallj! hidden
(J,ones 1977,Telatnik 1977,Walseleben 1976,Schumann 1979, Bailey
1979).So f a r I think this work is. as a research attitude, ambiguously
nomothetic. But Schumann and Schumann do provide a good her-
meneutic rationale for diary studies: the method allows us to find
individual factors in SLA (as opposed to the statistical averages of a
group).
Here is a starting point for future SLA research. Diaries con-
tribute little, probably nothing, to a stop-frame picture of SLA. On the
one hand, cross-sectional/longitudirwlexpe&ments work for us’if we
stop the world, provided it can stop; ow the other hand, diaries work as
a remembrance of things past. We should regard both methods as
documents of our SLA perceptions. Or we can define our workin other
ways, as a vocabulary of attitudes or a research bilingualism. What-
ever metaphor it is that a poetics defines, simply put-we need two
research traditions, or more.

APPENDIX A

Nomothetic Science

Where Plato gave us a single, lawful reality and the philosopher-

king to explain it,*Galileo, in the early seventeenth century, proyided
nomothetic scientists with a research ideal: the controlled experiment.
In other words, we can infer physical laws in the same way that Galileo
predicted the heavenly-bodies’ orbits (Znaniecki 1963, Boring 1950).
Galileo demonstrated that Plato’s Ideal world, in a physical sense,
exists. This Platonic/Galilean world view is fundamentally static
(Cassirer 1946, Popper 1965, 1963).The first modem researcher, Sir
Isaac Newton (Iate seventeenth century), s a w the world as if i t were a
perfect machine, a huge clock. In this mechanized universe chance
 73
events do not occur. Or they Uexist”only because we mistake certain
orderly events as randomness. This order provides the groundwork for
a ‘building block’ notion of reality. A static world must have its
individual units or atoms, and joined together these atoms become
molecules that then form larger and more complex chunks of existence
( K q l i n g e r 1964).
Newton gave to science ‘a mechanistic theory; he also made
paramount the method of controlled experiments and applied mathe-
matics (Boring 1950,Znaniech 1963).Science now had its formula for
success. Thermodynamics and steam engines (1769)made clear the
future order: by controlled experiments the world was ours to study
and know, and once known, we could engineer any blueprint or social
goal, even our freedonl (e.g., Comte 1957,Skinner 1971).Science thus
liberated man. And no areas of study did this better than physics,
Chemistry-the ‘exact’natural sciences (Boring 1960).
Success guarantees a copy. If we can successfully know the
physical world, we can also, scientifically, look at ourselves. The
supreme model for ~ u r“human” research should be mathematical
physics; therefore, we should also idealize the experimental method.
August Comte in 1830 founded sociology on the “positivist’’ zeal for
nomothetic science (Simpson 1969).The human (Barthe 1967),cul-
tural (Znaniecki 1963), and social scie:ices (many authors) have in
general followed.
Positivism shifts our world view as radically a s Galileo’s experi-
ments. It redefines us. In nomothetic science there is one physical
world, and people in this world do not differ, ultimately, from things.
By defining people as physical objects (albeit with certain peculiari-
ties), we can resolve the m i n d m y duality (Boring 1950,Feigl 1967).
From a materialistic viewpoint, the mind is a physical entity, an
organism t h a t responds by “conditioned reflex” (Watson 1913,Boring
1950).Because it exists physically, we can measure it. Irr sum, the
external world is us.
Watson’s “behaviorism” is a subgeneric model of Newtonism
physics (Boring 1950). Whether it descends as the “purest” form of
pure (social) science, not everyone agrees (Jenkins 1968).Nor should
they. In a nomothetic heritage, the social kiences will always be, when
compared to the ‘exact’ sciences, merely distant cousins. This poor
relation directly bears, of course, on the status of a social sciencg.
Experimental psychology was, and probably still is, the most nomo-
thetic social science, and the most imitated (Jenkins 1968).Even the
positivist spirit stays alive in this discipline (e.g., Skinner 1971 ).
Linguists have turned away, in recent years, from the psychol-
ogical model. But Chomsky’s revolut,ion does not place language
74 LANGUAGE LEARNING VOL. 29, NO. 1

theory totally outside the nomothetic tradition. Instead, the Platonic

world-view triumphs; so thoroughly, in fact, as to make purely formal
theory the main concern of generative lipguists (Dougherty 1976, cf.
articles in Wirth 1916, also see Antilla 1973, Itkonen 1974).
The other social sciences remain overwhelmingly nomothetic, al-
though in recent years a “paradigm” shift may have begun ( I use the
term “paradigm” loosely, for as Kuhn ( 1970) notes, the social sciences
may be, as a whole, pre-paradigm). In SLA research, a variety of
methods exists-but in few studies has the issue of a meta-theory been
reoognized or discussed (seeClarke 1976, Schumann and Schumann
1977). One could argue that we have no meta-theory, that no SLA
researcher adheres to a nomothetic tradition. But then how does one
explain the kind of SLA work commonly done? We control variables,
look for general trends, try to be neutrally objective. That’s nomothetic
science (seeTarone et al. 1976. for an implicit statement of nomothetic
values). In any case, i t is naive to think that SLA resear%hers do
whatever amuses them; as if we have a smorgasbord of methods ta
choose from. If we don’t set the table, then someone else does.

APPENDIX B:

Hermeneutic SCience

It would be wrong to juxtapose causality and teleology, Plato and

Aristotle, as if these men and their ideas were wholly a t odds. They
weren’t (Morris 1967, Russell 1945).Neither was Aristotle a champion
of the Sophists’ ideas, or a disparager of nomothetic science. His own
work supported both the hermeneutic and nomothetic ideals. But his
Rhetoric and particularly the treatise “On Interpretation” place
Aristotle more clearly in the hermeneutic tradition. He also empha-
sized language study, a point of obvious contrast with Plato, who
suspected language as a vehicle for ideas.
Akistotle’s respect for language as an instrument of discovery-a
tool that compares favorably with m a t h e m a t i c e w a s continued by the
great Ronian rhetoricians: Quintilian, Cicero, Seneca (100 B.C.-100
A.D.). But i t was Saint Augustine (350430)who gave the art of
rhetoric, and hermeneutics study in general, a modern heritage
(Murphy 1960). Augustine s a w rhetoric as a way to mediate between
spiritual and human worlds, to move from the religious to the secular
planes of reality. For this reason he sanctioned priestly sermons.
Rhetoric, I should emphasize, can be either a means of persuasion,
a high-level adman’s game; or i t can, as in nomothetic science,
 75

represent but one virtue: clarity. Because material facts “clearly”

pertain to a physical world, research reports can be standardized. In
other words, a single format will serve as a window through which an
empirical reality is described. But in hermeneutics studies the data
has meaning imbued by its form: the window fogs up. And because
hermeneutic facts often are not directly o k r v e d , we must teleo-
logically deduce them. This distinction is at the heart of Augustine’s
rhetoric. The Word of God does not burst-with clarity-through
human ignorance; not, at any rate, in Augustine’s Catholicism, so he
encouraged priests to interpret the City of God for the city of man.
Two worlds, God’s and man’s, required hermeneutic performances
of the Church and layman. By the Reformation (early sixteenth
century), the precedent of Augustine’s “priestly rhetoricians” led tp a
more radical ideal: free interpretaion of the Bible. Some lively theo-
logical disputes soon followed. Luther and Zwingli, a t the begmning of
the Reformation, argued with each other then broke. away; and as
Protestantism grew, the Puritans, Calvinists, Anabaptists, Quakers
and other sects interpreted their own kinds of relison (Chadwick
1964).
While science became more powerful in the seventeenth and
eighteenth centuries, relimons splintered into compet’ 8‘ dogmas.
’$
Schleiermacher ( 1768-1834)’a German theologan and hilosopher,
tried to reconcile his German Protes&tism with modern science.
Earlier in the eighteenth century, Biblical scholars had greatly refined
t h e art of trdlnslation; by Schleiermacher’s lifetime, hermeneutics had
been further applied to non-religious texts, especially in a philolopcal
spirit (Palmer 1969).If i t worked for language studies, Schleiermacher
argued, then hel’mencutics could more generally serve us. He sub-
sequently proposed that hermeneutics be the method for all “human ”

areas of scientific inquiry (Palmer 1969).

The successor to this scientific hermeneutics is Wilhelm Dilthey
(1833-1911).Although not the first to develop this idea, his Geistes-
wissenschaften (or science of the spirit) holds the study of human
events separate from inanimate nature (Abel 1948). He also main-
tained that we should understand (Verstehen)human events, but that
we should explain nature by its laws. Historiography, Dilthey claimed,
provides the best model for this hermeneutics of understanding. In
other words, to understand a person’s “inner life” the historical
make-up of this person’s subjectivity must be explored. Each herm-
eneutic study is therefore a unique account of human acts, utterances,
products institutions. Although the method is particular, a study cun
show aggregate wholes or “types” which other studies m a y then share
(Hamilton 1974).
76 LANGUAGE LEARNING VOL. 29, NO. 1

Many other Germans have contrjbuted to hermeneutic theory.

During the early 1900’9, h c k e r t , Cohen, Natorp and Windelband
emphasized a “neo-Kantian” form of hermenteutics (Hamilton 1974).
It was Windelband, of course, who coined the terms nomothetic and
idiographic. But as Weber ( 1864-1920)noted, aggregates of individuals
can form a trend which, if not causally based, nevertheless involves
more than a single (idiographic) subject.
Like Weber, Husserl ( 1859-1938) has developed the Hegelian
tradition in German thought, one that does not directly fG into
hermeneutic tradition (Hamilton 1974). But both men have introduced
to this country-Weber by his anti-positivism and Husserl by his
phenomenology-some of the general ideas shared by a Kantian/
Hegelian philosophy.
Hermeneutics has been more directly channeled into our modern
ibeas, not surprisingly, by Protestant theolog.lans. Barth and
Bultmann have, I think, popularized this German tradition most
among current divinists; the atheist/existentialist Heidegger has also
been influential (Funk 1966, Doty 1972, Braaten 1966).
Presently, the German social philosopher Jurgen Habermas is
making hermeneutics a lively topic among philosophers of social
science (Hamilton 1974). Entire issues ofjournals have been devoted to
his ideas (see the reference listed under Habermas in the biblio-
graphy); and at least one American writer (Bernstein 1976) has
developed his neo-Marxist point of view. However, hermeneutics need
not fit a Marxist paradigm (Pinar 1978). For example, Palmer (1969)
and Hirsch (1967, 1976) have applied i t to literary research as an
a-political, but fully developed, system.
As a final remark I shoulfi add this disclaimer. To interpret an
event is fundamentally a rsonal act. For this reason a hermeneutic
8“
method cannot be standar ized, and neither can it be easily described.
Each hermeneutiec method fits a particular context; there are as many
methods as there are people to apply it. This non-standardized re-
search should not, hawever, be interpreted as a license to do anything.
As noted earlier, we can rely on our common-sense judgements when
doing this work and when evaluating it.
But we must refine our sensibilities or risk being utterly foolish.
Historians and literary critics know this dilemma well, for no hand-
boo)ucan express an individual approach to learning. It comes with
practice. That now disdained phrase “good taste” is what makes a
hermeneutic study worthwhile. ofcourse, one does not learn taste; like
a second language i t is largely acquired.
 77
REFERENCES
Abel, T. 1948. The operation called Verstehen. American Journal of Sociology 54.211-
218.
Ackoff, R. 1962. Scientific Method. New York: John Wiley.
Antilla, R. 1973. Child language, abduction, and the acquisition of linguistic theory by
linguists. In von H e r - E n g e l (ed.1,Roceedinga of the International Symposium
on First Language Acquisition, 24-36. Brueeels: University of Bruseels Press.
Antilla, R. 1975. Generalization, abduction, evolution, and bnguage. In Koerner (ed.1,
The Tradormational-Generative ParaQgm and Modern Lnguistic Theory
Amsterdam: John Beqpmins B.V.
Apel, K. 0.1967. Analyt~cPhiloeophy of Language and the C e i e t e s w s e e n e c h n .
Dordrecht-Holland: D. Reidel.
b o u l t , M. 1972. Fundamentals of Scientific Method in Psychology. Dubuque, Iowa:
Wm. C. Brown.
Bailey, K. 1979. An intrqpechve analysis of an individual's language learning ex-
perience. In Krashen and Scarcella (&.), Proceedings of the Second Language
Reeearch Forum. Rowley, MA: Newbury House (in press).
Bailey, K. D. 1976. Methods of Social Reeearch. New York: Free Press.
Bailey, N., Madden, C. and Krashen,S. D. 1974. Is there a,"natural sequence" in adult
second language learning? Language U i n g 24.235-243.
Baken, D. 1972. Psychology can h& the science habit. Psychology Today 5.26-28,86-88.
Barash, D. 1977. Sociobiology and Behavior. New York: Elesevier.
Barker, E. 1964. Greek Political Theory. London: University Paperback.
Barthes, R. 1968. Elementa of Semiology. New York: Hill and Wang.
Bernstein, R. J . 1976. The Restructuring of Social and Political Theory. New York:
Harcourt. Brace, Jovanovich.
Borg, W. 1!363. Educational Research. New York: D. McKay
Boring, E. G. 1950. A Hhtory of Experimental Psychology. New York: Appleton-
Century-CroRs.
Braaten, C. E. 1966. New h e c h o n in Theology Today (Vol 2): History and Her-
meneuhcs. Philadelphia: Westminster Presa.
Buchler, J . (ed.) 1955. Philoeophcal Writings of Plerce. New York: Dover Publications.
Burke, K. 1965. Permanence and Change. Indianapolis: Bobbs-Memll.
Burke, K.1967. Piuloeophy of Literary Form. Berkeley: University of California Press.
Burke, K. 1968. CounterShtement. Berkeley: University of California Press.
Burke, K. 1969. A Grammar of Motives. Berkeley: University of California Press.
Caillois, R. 1961. Man, Play, and Games.New Ywk: Free Press.
Campbell, D.and Stanley, J. 1963. Experimental and Quaai-Experimental Designs for
Research. Chicago: Rand McNally College Publicabons.
Campbell, P. N. 1975. The pemnue of scientific discourse. Quarterly Journal of m h
6 1.391405.
Caplan, N. and Nelson, S. 1973. On being useful: the nature and consequences of
psychological reeecurch on social problems. American Psychologist 28.199-211.
Chadwick. 0. 1964. The Reformation. London: Penguin Boob.
Chatman, S. (ed)1978. Approaches t:, Poetics. New York: Columbia University RF.
Clarke, M. 1976. k n d language acquisition as a clash of 0onac;ousnesa. Language
Learning. 26.377-590.
Cohen, M.R. 1931. Reamn and Nature. New York: Harcourt, Brace.
Comb, A. 1957. General View of ?oeitwiam. New York: R. Speller and Sons.
Cornford, (ed.)1954. The Republic of Phto. London: Oxford University Press.
Cmnbach, L. 1963. Educational Psychological. New York: Harcourt. Brace and World.
Crowle, A. J. 1971. Poet experimental interviews:an experiment and a emciolingustic
analysis. Unpublished Ph. D. diseertauon. The University of California a t Santa
Barbara.
Crowle, A. J . 1976. The d-+ve language of the laboratory. In Ham6 (ed.),
Life
Sentences. London: J. Wiley and Sons.
 LANGUAGE LEARNING VOL. 29, NO. 1
Deese, J. 1969. Behavior and, fact. American Psychologmt 24.515-522.
Doty, W.G. 1972. Contemporary New Testament Interpretation. Englewood Cliffs, N J :
Pren tice-Hall .
Dougherty,R.1976. Argument invention: the linguist’s “feel” for science. In W i r t h t e d . ) ,
Asreesing Linguistic Arguments. New York: J. Wiley and Sons.
Dulay, H. and Burt, M. 1974. Natural sequences in child second language acquisition.
Language Learning 24.37-53.
Ebel, R. 1967. Some limitetions of basic research in education. Ph Delta Kappan
4g-L?
Egner, R. Denoon, L. (eds.)
Simon and Schuster.
1961. The Basic Writinga of Bertrand Ruseell. New York:

Feigl. l& 1967. The “Mental” and “Physical.” Minneapolis: University of Minn-ta
press.
Feigl, H. and Brodback, M. 1953. B_sa&ngs in the Philoephy of Science. New York:
Appleton-Cen tury-Cr’ofts.
Filstead, W.ted.). 1970. Qualitative Methodology. Chicago: Markham.
Funk, R. W.€966.Lanpage, Hermeneutic and the Word of God.New York: Harper and
ROW.
Gardner, R. and Lambert, W. 1972. Attitudes and Motivation in Second Language
Learning. Rowley, MA: Newbury House.
Garfinkel, H. 1967. Studies in Ethnomethodology. Englewood Cliffs. NJ: Prentice-Hall.
Ceerth, H.and Mills. C. W. teda). 1975. From Max Weber: Essays in Sociology. New
Yo& Oxford University Press.
Ceertz, C. 1973. The Interpretation of Cultures. New York: Basic Books.
Goffman, E. 1959. The Reaentation of Self in Everyday Lde. Garden City, NY:
Doubleday.
Golden, J., Goodwin, F. B., and Coleman, W. 1976. The Rhetoric of Western Thought.
Du buque, Iowa: KendaWHunt .
Guiora, A. 2.1972. C o n s h c t validity and transpositional research: toward a n empirical
study of psychoanalytic coricepte. comprehensive Psychiatry 13.139-150.
Habermas,J. 1971. Knowledge and H u m ~ nIntbrests. Boston: Beacon.
Habermas, J. 1972. (Review symposium on Habermas. Philosophy of Social Sciences 2 ) .
Hakuta, K. and Cancino, H. 1977. ”’rends in second language acquisition. Harvard
Education Review 47.294-316.
Hamiltoh, P. 1974. Knowledgb and Social Structure. London: Routledge and Kegan
Paul.
k e , R (-pi.).1976. Life Sentences. b n d o n : J. Wiley and Sona
Hatch, E. (ed.).1978. Second Language Acquisition: A Book of Readings. Rowley, MA:
Newbury H o w .
Heidegger, M. 1961. Introduction to metaphysics. New York: Doubleday-Anohor.
He+&uh.er, G. C. 197Q Research Concepte in Human Behavior. New York: Appleton-
Centuy-Crofb.
Hirach, E. D. 1967. Validity in Interpretation. New Haven: Yale University Press.
H i m h , E.D.1976. The Aims of Interpretation. Chcago: University of Chcago Press.
Huizinga, J. 1950. Homo Ludens. Boeton: Beacon.
hkonen, E. 1974. Linguistics dnd Metaecience. Koke&: Risteen K j a p n o .
Jenkins, J. J. 1971: The challenge to psychological theoriste. In Johnson (ed.), Learning: I
Theory and Ractice. New York: Thomas Y.Crowell.
Jones, R. 1977. Psycholo ‘4,social, and personal factors in second language acquisi-
tion. Unpublished % l
.A. thesis. University of Callfornia at Los Angeles.

Kerli?r ,F. 1964. Foudataons of Behavioral €beearch. New York: Holt, Rinehart and
inston.
Krashen, S. 1976. Formal and informal linguistic environmenta in language learning
and language acquisition. TESOL Quarterly 10.157-168
&ashen, S . 1977a. The monitor model for adult second language performance. In M.
Burt, H. Dulay, and M. F i n n o c c b (eds.),Viewpoints on English as a Second
Languad, 152-161. New York: Regents.
 OCHSNER 79
Krashen, S. 1977b. Some issues relating to the monitor model. Paper presented at the
Eleventh Annual TESOL Convention, Miami Beach, Florida, April 26-May 1.
Krashen, S. 1978. Individual variation in the uae of the monitor. In W . Iltchie (ed.),
second Laqpage Acquisition Reseakch: Issues and Implications. New York:
Academic Prese.
Kuhn, T. 1970. The Structure of Scientific Revolutions. Chicago: University of Chicago
Pr088.
Kurman, R. 1977. Research: a n aerial view. E X .34.265-276.
canham,R..A. 1 w 9 . A Handlist of Rhetohcal Terme. Berkeley: University of Callfornia
ReSS.
Lanham, R.A. 1976. The Motives of Eloquence,New Haven, CT:Yale University Preee.
Lanham, R. A. 1977. Switching on and off. Mimeographed paper.
Larsen-Freeman, D. 1975. The acquisition of grammatical morphemes by adult ESL
students. TESOL Quarterly 9.409419.
Larsen-Freeman, D. 1976. An explanation for the *owheme acquisit-lon order of
secdpd language learners. Language Learmng 26.125-134.
Lenneberg, E. 1967 Biological Fouf&htions of Language. New York: J . Wiley and Sons.
L y k k m , D. l!W. Statistical signTicance in psycholofical research. Fsychologmd
Bulletin 70.151-159. -
Matheson, D., Bmce, R. and Beauchamp, K. 1970. Introduction to Experimental
Psychology. New York: Holt, Ilnehart and Winston.
Meehl, P. 1967. Theory-testing in psychology and physics: a methodologxal paradox.
Philosophy of Science 34.103-115.
Mitroff, I. and Kilmann, R. 1978. Methodologkal Approaches to Social Science. San
Francisco: J ossey - Bass.
Morris, C. 1967. Western Political Thought tVol 1). New York: Basic Books.
Murphy, J. 1960. Saint Augustine and the debate about a Christian rhetoric. Quarterly
Journal of Speech. 46.400411.
Naiman, N. Frolich, M. and Stem, H. H. 1975. The Good Language Learner. Toroqto,
Canada: Modem Language Center, Ontario Institute for Studies in Education.
Nemerov, H. 1978. Figures of Thought. Boston: M i n e Press.
Palmer, R. E. 1969. Hermeneutics. Evanston, IL: Northwestern University Press.
F'inar, W.1978. Notes on the curriculum field 1978. Educational Reshamher, September.
7.5- 12.
Popper, K. 1963. Coqjectures and Refutations. New York: Harper and Qw.
Popper, K. 1965. Logic of Scientific Discovety. New York: Basic Books.
Ritchie, W. C. (ed.)1978. Second Language Acquisition Research. New York: Academic
Press.
Rivers, W. 1975. Finding Facts. Englewood C l f i s . N.J. Prent-lce-Hall.
Rosaneky, E. H76. Methods and morphemes in SLA research. Language LearNng.
26.409425.
Rosenthal. R. 1976. Experimenter Effects in Bebvioral Research: Enlarged Edition.
New York: Lrvington.
RosenthAl, R. 1977. Brasing effects of experimenters. ETC 34.253-264.
'Russell, B. 1945. A History of Western Phloeophy. New York: Simon.
Ryan, A. 1970. The Philosophy of the Social Sciences. London: Maximillan Press.
Scheflen. A. 1973. Communicational Structure. Bloomington, IN: lndiana University
PreSS.
Scheflep, A. 1977. Classical biacles and the structureal approach to research. ETC
34.290-300.
Schumann, F. 1979. Unpublished M.A. thesis. University of California a t LQS Angeles
(in progress).
Schumann, F. and Schumann, J. 1977. Diary of a language 1earner:'an introegective
study of aecond language learning. In H: D. Brown, C. A. Yoria. and R. H. Crymea
(eds.), Teaching and Learning: Trends in Research and Practice. Selected Papere
from the 1977 TESOL Convention. Wadungton, D.C.: TESOL Georgetown Unie
v e r s ity .
 LANGUAGE LEARNING VOL. 29, NO. 1
Schumann, J. 1976a.Second language acquisition research: getting a more global look
at the learner. In H. D. Brown (ed.),Papers in Second Language Acquisitio?,
Language Learning Special Issue 4.13-27.
Schvmann, J. 1976b. Social distance as a factor in second language acquisitio&
Language Learning i5.135-143. I
Schumann, J. 1978a.Social and peychologxal factors in second language acquisition. In
J. C . Richards (ed.),Understanding Second Language Learning. Fbwley, M A :
Newbury House..
Schumann, J. 1978b.Acculturation: the causal variable in second language acquisitio;.
Prepublication manuscript.
Scovel, T. 1977.The qntogeny of the ability to recognize foreign accents. In C. Henning
(ed.),Roceedl'nga of the Lim Angel- Second Language Research Forum.
Severin, F.T.(ed.)1965. Humanistic Viewpoints in Psychology. New York:.McGraw-
Hill.
Simpeon, G. (ed.11!369.+AugusteComb: Sire of Sociology.New York:Thomas Y.Crowell.
Skinner, B. F. 1971. Beyond Freedom and Dignity. New York: Alfred A. Knopf.
Smith, N. EL 1970. Replication studies: a neglected aspect of psychologma1 r e s e a r d .
American Psychologist 25.970-975.
Stake, R. 1978. The case study method in social Inquiry. Educational Hesearcher,
February. 7.5-8.
Taronq, E., Swain,M.and Farthman, A. Some limitations to the classroom applications
of' current second Language acquisition research. TESOL Quarterly 10.19-32.
Telatnik, M.A. 1977.The intensive puma1 as a self-evaluative instrument for the ESL
teacher. Unpublished M.A. thesis. University of California a t Los Angeles.
Tuckman, B. 1972. Conducting Educational Research. New York: Harcourt, Bra&,
Jovanovich.
von Wright, G. H. 1971. Explanation and Understanding. Ithaca, NY: Cornell U d -
versity Press.
Walsekben, M.b976. Cognitive and affective factors influencing a learner of Persign
(Farsi). Unpublished term paper.
Watson, J. B. 1913. Psychology as a behaviorist views it. Psycholoscal Review
20.158-177.
Wirth, 3 . (ed.)1976.Assessing Lnguistic Arguments. Washington: Hemi-Publishing.
Winch, P. 1958. The Idea of a Social Science and I t s RelatioPcto Philosophy. London:
Routkedge and Kegan.
Windelband, W. 1894.Geachichte und Naturwissenschaft. Reprinted in J. C. B. Mohr
(ed.),Praludien, Tubingen: 1907.
Wolcott, H.1975.Criteria for a n ethnographic approach to research in schools. Human
Organization 34.11 1-127.
Znaniecki. F. 1963.Cultural Sciences. Urbana. IL: University of Illinois Aess.