UNIDAD 2

2.1 Meaning and the dictionary

The concept of a word’s meaning is closely linked to the concept of definition.
Definitions have been particularly important for conceptual theories of meaning
which traditionally assumed a close link between concepts and definitions: knowing
the concept horse, for example, is simply the ability to use the word horse in a way
that accords with or fits its definition. E.g.:
A. If X is a horse, X is an animal.
B. If X is a horse, it has a mane.
C. X is a rooster, so X is not a horse.
D. If X is a horse, it is a large four footed mammal with a mane.
As a result, an understanding of definition is necessary for any attempt to develop a
conceptual theory of word meaning.
When people think of a word’s meaning, they are inclined to think of something like
its definition in a dictionary.
It is important to clarify the similarities and differences between the definitions that
might be proposed in theoretical linguistics semantics, and the types that can be
found in dictionaries.
2.1.1 Semantics and lexicography
Dictionary-writing, or lexicography, is, in the words of Landau (1984:121)
“a craft, a way of doing something useful. It is not a theoretical exercise to
increase the sum of human knowledge, but practical work to put together a
book that people can understand.”
Linguistic semantics, by contrast, while also interested in the meaning of words, is
exactly the sort of theoretical exercise with which Landau is drawing a contrast.
Nevertheless, the model of the dictionary or “lexicon” has been decisive in the way
that many linguists conceive the nature of language.
2.1.1 Mental lexicon
Our brains hold a “stock of words in long term memory from which the grammar
constructs phrases and sentences” (Jackendoff 2002:130). This stock of words as
associated meanings is usually referred to as the mental lexicon.
The primary task of linguistic semantics would be the specification of the stored
meaning representation -- the “entry”-associated with each lexeme in the mental
lexicon.
What kind (and amount) of information do entries in the mental dictionary contain?
Much more detailed than ordinary dictionaries  Grammatical properties,
phonological structure, lexical relations...
Even the most comprehensive dictionary entry will be lacking important information
that we speakers seem to know...
Consider, for example, the Concise Oxford Dictionary entry for the verb pour:
v. 1 intr. & tr. (usu. foll. by down, out, over, etc) flow or cause to flow esp. downwards
in a stream or shower 2 tr. dispense (a drink, e.g. tea) by pouring. 3 intr. (of rain, or
prec. by it as subject) fall heavily. 4 intr. (usu. foll. by in, out, etc.) come or go in
profusion or rapid succession (the crowd poured out; letters poured in; poems
poured from her fertile mind). 5 tr. discharge or send freely (poured forth arrows). 6
tr. (often foll. by out) utter at length or in a rush (poured out their story).
At first sight, this entry presents a comprehensive description of the verb. However,
there are a number of aspects of pour’s meaning and use which the definition does
not cover...
The dictionary is silent about the limits on the prepositional and subject combinations
with which pour is acceptable: why are examples (1) and (3) acceptable, but the
others less so?
1. The crowd poured down the hill
2. ?The firemen poured down the pole

3. The tourists poured into the museum

4. ?The surfers poured into the ocean
5. ?The passengers poured into the bus
6. ?Fifty workers poured into the lift
2.1.1 Word-based and meaning-based approaches to definitions
The definitions found in dictionaries are the result of a word-based, or
semasiological approach to meaning. This sort of approach starts with language’s
individual lexemes, and tries to specify the meaning of each one.
Onomasiological definitions have the opposite logic: start with a particular meaning,
and list the various forms available in the language for its expression.
A semasiological analysis would start with a list of verbs like: scare, frighten, terrify,
startle, spook, and panic, and specify a different meaning for each.
An onomasiological analysis would start with a general concept, FRIGHTEN, and list
all these verbs as possible realisations. The difference between the two approaches
corresponds to the difference between a dictionary and a thesaurus.
2.2 The units of meaning
To associate meanings and forms we need to ask what the minimal meaning-bearing
units of language are.
We need to recognize meanings both above and below the word level.
- Below word level: Morphemes (e.g., prefixes) and words
- Above word level: Idioms
2.2.1 Words and morphemes
How can we determine what counts as lexeme (word) in a language? What units
should we be trying to attribute meanings to?
European languages  words are the units surrounded by spaces in standard
orthography. But:
- Languages often have a very fluid practice of word division. A meaning-
bearing unit considered by one speaker as only part of a word will infrequently
be written as a separate word by another speaker:
song-writing/song writing/songwriting

- Writing conventions tend to be quite unstable in the history of a particular

language (i.e. Greek or German)
2.1.1 Criteria for word identification
Words as isolable linguistic units: one is the potential pause, words are units before
and/or after pauses which can be found in spoken language.
For languages like Chinese, which lacks complex morphology, this criterion may be
workable. But for languages which show even a small degree of morphological
complexity, like English, it is clearly unsatisfactory. Dixon and Aikenvald (2002:11)
point out that one may pause at morpheme boundaries within a single word, for
example “it’s very –un (pause) suitable”. Similarly, expletives in English can be
inserted within what we normally consider a single word: abso-bloody-lutely.
Morphemes: in linguistics, a morpheme is the smallest grammatical unit in
language. A morpheme is not identical to a word, and the principal difference
between the two is that a morpheme may or may not stand alone, whereas a word,
by definition is freestanding. Every word comprises one or more morphemes:
Unbreakable: un-bound morpheme; break-the root-free morpheme; able-bound
morpheme.
But... words as “minimum free forms” (Bloomfield) i.e. the minimal unit which may
appear on its own without any additional grammatical material, is insufficient: many
canonical words like the, of or my do not usually appear alone, but must presumably
be considered fully-fledged words.
If words are the clearest type of meaning-bearing unit in a language, they are
certainly not the only ones: the domain of meaningfulness extends both above and
below the threshold of the individual word.
Below the word level, morphemes, by definition, have meanings. Given the definition
of a morpheme as the “minimal meaning-bearing unit” of language, it is clearly
impossible to conceive of a morpheme without a meaning, even if it is often hard to
specify exactly what this meaning is.
Above the level of the individual word: phrasal verbs and compounds are two clear
cases where a single meaning is associated with a combination of lexemes.
Idioms also demonstrate the existence of units of meanings associated with several
words simultaneously.
Thus, although we most often think of meaning as something belonging to individual
words, we must actually recognise that words are only the most obvious of a number
of meaning-bearing units.
2.2.4 Contextual modulation of meaning POTENCIAL PREGUNTA DE EXAMEN
The meaning of words and other morphemes vary according to their collocation, the
immediate linguistic context in which they occur. In English, the meanings of the
verbs seem to vary slightly depending on the noun which they govern.
- ... cut my foot,
- ... cut the grass,
- ... cut a cake,
- ... cut someone’s hair,
- ... cut the wood,
- ... cut a diamond,
- ... cut a deck of cards,
- ... cut a disc
The nature of the event, the means by
which it is accomplished, its typical object, and the extent to which it is deliberated
may all vary in these different uses. Despite this variation, we have the strong sense
that essentially the “same” meaning of cut is involved in all those cases.
COMPOSITIONAL OR NON-COMPOSITIONAL
Cruse (1986:52) refers to this phenomenon as the contextual modulation of
meaning. The degree of semantic “distance” gets even greater if we consider more
“extended” meanings, like cut a deal, cut corners, cut a paragraph, cut prices.
This type of phenomenon poses an interesting descriptive and theoretical problem:
do the differences in meaning of the different collocations arise compositionally or
not?
Are the meanings of the collocations just the results of the combinations of the
meanings of their parts, or are the whole collocations themselves the meaning-
bearing units?
We have two possibilities:
 - One which lists the meanings of cut, foot, grass, cake, hair, etc, and sees the
specific meanings of the collocations cut one’s foot, cut the grass, cut a cake,
as derived compositionally from the meanings of the parts.
- Another one which just lists all the different collocations in which cut appears,
and specifies a different meaning for the entire collocation.
The first possibility is that the meanings of cut one’s foot, cut the grass, cut a cake,
etc. result compositionally from the meaning of the verb cut and the meanings of its
noun objects. The meaning of cut the grass is just the meaning of cut combined with
the meaning of grass.
2.2.4.1 Composionality in meaning POTENCIAL PREGUNTA DE EXAMEN
The general meaning hypothesis: Cut might have the same vague or general
meaning in all its different collocations: it refers to some act of accomplishing a
material breach in a surface, with the particular details of each type of breach being
inferred by the listener, rather than being built into the meaning of the verb itself.
The multiple meaning hypothesis: Cut might have a separate meaning in each
collocation: the cut in cut one’s foot has its own entry in the mental lexicon (“breach
surface of, usually accidentaly”), as does the cut of cut the grass (“sever one part of
surface from another, usually deliberately”).
2.2.4.1 Problems with the general meaning hypothesis

Describing this common core of general meaning supposedly present in all cases of
cut is not necessarily an easy matter: Concise Oxford Dictionary “make an opening,
incision, or wound with a sharp tool or object”
This is not involved when someone cuts butter, for example, nor when a whip cuts
someone’s flesh: the cutting object in these situations need not to be sharp.
In the same way, the definition does not adequately distinguish cut from chop, slit,
stab or unpick: to chop a sausage, slit a letter, stab someone’s side, or unpick a
seam. We have failed to distinguish cut from various non-synonymous verbs in the
same semantic field.
The second option is to propose multiple meanings for cut, a separate one for each
collocation.
- Cut one’s foot, for example, cut could be described as meaning something
like “partially breach a surface with a sharp instrument, typically accidentally”.

- Cut the grass, and cut someone’s hair, on the other hand, the verb conveys
the meaning of more than just a partial breach in the surface of the object: the
meaning of these collocations is that one part is completely detached from the
rest.
These will have highly specific collocational restrictions: “partially breach a surface
with a sharp instrument, typically accidentally” for example, will be a very likely sense
of cut in collocation with foot, but not with cake: cutting a cake is usually an entirely
deliberate action.
The meaning “create by partially breaching a surface with a sharp instrument” is
quasi-obligatory in cut a notch, but excluded in cut wood, which does not, as we
have seen, involve any creation.
2.2.4.1 Problems with the multiple meaning hypothesis
More problems: the sheer number of the different senses to be attributed to cut. The
recognition of a different sense of cut in each of the collocations seems to fail to do
justice to the fact that it is the same verb in all collocations: as a result, we have
some reason to think that it is also the same meaning that is involved in all of them.
The attribution of a separate meaning to cut in each collocation may be inefficient
and inelegant, given the explosion it entails in the number of separate verb entries.
Given this variety of different possible meanings of cut, how does the correct specific
meaning get chosen in a given case? The process of word sense disambiguation is
highly problematic.
2.2.4.1 Problems with compositionality
To sum up, we have been considering the possibility that the meaning of collocations
like cut one’s foot, cut the grass, etc. are derived compositionally from the meanings
of their elements. We looked at two options for the details of this:
- The first is that the meaning of cut is general or vague in each collocation.
This creates the problem of adequately defining this general or vague
meaning in a way which distinguishes cut from other non-synonymous verbs.

- The second option is that cut has a separate meaning in each collocation. But
if we adopt this solution we find that the number of definitions of cut explodes.
Confronted with this array of different meanings, how do speakers know which
one to choose in any given case?
The compositional solution seems quite problematic. This is not to say that we
should reject it, just that it involves us in complex questions. Let us now look at the
non-compositional solution.
2.2.4.2 Second possibility: non-compositionality
A number of the problems of the first solution are avoided if each collocation as a
whole is seen as the relevant definition-bearing unit.
On this approach, the meaning of the collocation is not constructed compositionally.
Rather, we learn one definition for the unit cut the grass, another for cut one’s foot,
a third for cut a CD. The fact that in cutting the grass, a mower is the instrument of
the action, and that in cutting a disc is a CD-burner, is not part of the meaning of cut
itself, but is a property of the collocation as a whole.
So, we do not have to advance a general definition of cut that will work in every
context, as we do in the general-meaning version of the compositional solution.
We do not have the problem of word-sense disambiguation, since each collocation
carries its own definition.
Another consideration in favour of non-compositionality is that it is not just cut whose
meaning is determined by its collocational environment: the collocation also
determines what reading is operative for cut’s object.
Thus, English speakers know that cutting the grass refers to the grown grass blades,
whereas planting the grass refers to grass seeds, and smoking grass refers to the
leaves of a completely different plant.
Because both verb and object have different meanings in different collocations, it
seems reasonable to think that the basic meaning-bearing unit is the collocation as a
whole, not the individual words.
2.2.4.2 Problems with non-compositionality
However, this solution is just as problematic as the compositional one. It seems
precisely to ignore our intuition of the compositionality of the meanings of the
collocations: the reason that cut the grass has the interpretation it does is, surely,
something about the combination of the meanings of cut and grass.
If one takes the analogy of the “mental lexicon” seriously, this option also involves
the threat of an explosion in the number of entries. Analysing each collocation
involving cut as having a separate meaning would lead to an enormous amount of
repetition and redundancy in the mental lexicon.
So it is more reasonable to use the general hypothesis as the most acceptable one,
because having hundreds and thousands of meaning for a single word would be
unconvenient.
2.2.4.2 Should we choose?
In the absence of a clear understanding of how the brain actually processes and
stores language, linguists have assumed that their description of assumed linguistic
competence should reflect the same criteria of economy and non-redundancy that
operate in real paper dictionaries.
Thus, much linguistic research has assumed that the mental lexicon does not
contain a huge number of independently listed entries, but that it extracts the
maximum number of generalizations about the meaning of a verb like cut across all
its collocational contexts, in order to present the most economical, least redundant
entry.
Even though it might seem inelegant to list all the different collocations of cut
separately in the lexicon, this option should not be rejected if somehow it turns out
that this is what speakers unconsciously do (for example, through neuroscientific
experimentation), which would NOT invalidate the idea that speakers at the same
time represent cut as having an independent (or set of) meaning which enters into
composition each time the verb gains a new set of arguments.
2.3 Different ways of defining meanings: real and nominal definitions
The concept of definition goes back to Aristotle, in Posterior Analytics, a treatise
devoted to the explanation of the structure of scientific knowledge. A definition
(horismos) has two quite different interpretations: “in defining,” says Aristotle, “one
exhibits either what the object is or what its name means” (Tredenick 1960:
II.7.92b).
A definition can therefore be considered either as a sort of summation of the essence
or inherent nature of a thing (real definition; Latin res “thing”), or as a description of
the meaning of the word which denotes this thing (nominal definition; Latin nomem
“name, noun”).
2.3.1
Some people have considered that definitions of the underlying nature of objects are
the only types of definitions which can be of interest.
Diderot, for example, stated that “definitions of words differ in no way from definitions
of things” (quoted in Meschonic 1991:102). And since it is scientific research which is
taken to reveal this underlying nature, these definitions will be formulated by
scientific disciplines.
On the other hand, according to Bloomfield, “we have no precise way of defining
words like love and hate”. On this understanding, linguistics should appeal to
technical scientific disciplines in formulating definitions.
The true meaning of natural language word, according to Bloomfield, is to be
identified with the scientific “definition” of its denotation. As a result, whenever a
scientifically established definition of a denotation is missing, there is nothing that
linguists can say with any certainty about the word’s meaning.
Bloomfield’s view is a serious obstacle to a comprehensive account of meaning, for it
is not just “abstract” nouns like love or hate which lack a scientific definition, but the
vast majority of the vocabulary of any natural language:
- Words like unicorn or time machine lack any denotation in the real world but
nevertheless have a meaning.

- Most of the vocabulary of a language has only a small amount of overlap with
terms of the sort which interest empirical science: most of the vocabulary
consists of words for a huge variety of objects, processes, relations and states
 which have no simple analogue in the scientific picture of reality (think of
reportage, postpone, ready).

2.3.1 Nominal definitions

They fulfil two different functions:
- Fixing the meaning of a word so that there can be no ambiguity about its
denotation. (extensional) = para nosotros mismos

- Bringing about an understanding of the meaning of a word in someone who

does not already understand it, typically in order to enable the word to be
correctly used. (Cognitive) = para explicárselo a otros
Thus, the definition “featherless biped” is an extensional definition of the noun
human, since it accurately identifies all and only the members of the class of
humans.
It is not, however, necessarily a very good cognitive definition, since the concept
HUMAN is not typically conceived in terms of bipedality or absence of feathers: when
we reflect on our concept HUMAN, we are likely to think of many other
characteristics (e.g., a certain physical form and a range of behaviours) before
bipedality or absence of feathers.
2.3.2 Definition by ostension
The most obvious way to define many words is, simply, by ostension, or by pointing
out the objects which they denote. In spite of the apparent obviousness of this
method, it is beset by difficulties.
Firstly, verbs, adjectives and prepositions are not open to this definitional method: if
you point at a black cat running along a wall, you are pointing at a cat, not at “black”,
“running” or “along”.
Even in the case of objects, ostensive definition is extremely problematic. Although it
is an appealing idea that meanings can be defined simply by pointing at objects in
the world, in practice this definitional method would seem to give rise to too many
ambiguities to be viable: the glasses example.
The only way to overcome the problems of ostensive definition would seem to be to
use language itself as the medium in which definitions can be phrased: only this way,
apparently, can we get the level of definitional precision we need.
2.3.3 Definition by synonymy
We might try, for example, to define words by providing synonyms, in either the same
language as the word being defined or in a different one.
Thus, one could give mad and furious as English definitions of angry, and kulu as a
Warlpiri one. The problem with this strategy is that it is usually possible to challenge
the identity between the definiens (the metalanguage word proposed as the
definition; Latin “defining”) and the definiendum (the object language word for which
a definition is required; Latin “needing to be defined”).
Thus, one could object that neither mad nor furious is really synonymous with angry,
since mad also means “insane”, which angry does not, and since furious actually
means something like “very angry”.
Similarly, although Warlpiri kulu does often translate English angry, it has a whole
range of other meanings, including “mean” and “fight”, which do not correspond to
those of angry.
2.3.4 Definition by context or typical exemplar
Another way to define a word is to situate it in a system of wider relations through
which the specificity of the definiendum can be seen.
This definitional strategy differs from the synonymy strategy in simply showing the
position of a definiendum with respect to other related notions which are not
themselves identical to it, as alleged synonyms are.
A possible definition of the verb scratch, for example, would be “the type of thing you
do when you are itchy”.
This is an example of definition by context: the definition identifies the event of
scratching by placing it in relation to another event, being itchy, whose meaning is
assumed to be already known, and which is taken as a typical context for the
definiendum.
This definition only works if the definition’s addressee correctly infers the intended
meaning on the basis of the cue given. Thus, if, when itchy, I am in the habit of lightly
striking my head against the wall, and if I believe that others do the same, the
definition will not be effective.
Definition by typical exemplar: The definition is a list of typical examples or instances
of the definiendum. If, given the German definiendum Vogel, I supply a list like
“swans, robins, geese, hens, magpies, etc” and add that bats, butterflies and
aeroplanes are excluded, you could correctly conclude that Vogel means “bird”.
2.3.5 Definition by genus and differentia
The two preceding types of definition are essentially relational, defining a word’s
meaning through its connections with other words. They may often be workable as
cognitive definitional strategies, but are unlikely to be successful as extensional
definitions (they won’t fix the meaning of a word so that there can be no ambiguity
about its denotation).
This is because they leave the essential nature of the definiendum’s meaning to be
worked out by the definition’s addresse, and as a result carry the risk that the wrong
meaning may be inferred.
The only way to convey this essential nature, apparently, is the strategy of definition
by genus and differentia.
Definition by genus and differentia (GD definition), the theory which was developed
by Aristotle in the Posterior Analytics (Tredennick 1960).
According to Aristotle, definition involves specifying the broader class to which the
definiendum belongs (often called the definiendum’s genus), and then showing the
distinguishing features of the definiendum (the differentia) which distinguish it from
the members of this broader class.
A classic example of GD definition is the definition of man (in the sense of “human
being”) as “rational animal”. This definition names the broader class of entities to
which man belongs (animals) and specifies the distinguishing feature which picks
man out from the members of the class of animals (rationality).
2.4 Definition and substitutability
DEFINIENS: The metalanguage word proposed as the definition.
DEFINIENDUM: The object language word for which a definition is required.
How can the accuracy of a definition be checked? For most semantic theories, a
minimum requirement on a term’s definition is the following: substitution of the
definiens for the definiendum should be truth preserving in all contexts.
For example, “keep in equilibrium” can be accepted as the definition of balance if it is
possible to substitute this phrase for balance in all the contexts in which balance
occurs without rendering any of them false. All the sentences below, for example,
remain true if “keep in equilibrium” is substituted:
- I balanced the plank on my head.
- She balanced the ball on the end of the bat.
- Now, children, you have to balance the egg on the spoon.
- I’ve never managed to balance the demands of work and play.
2.6 Problems with definitions
So far, we have been assuming that it is actually possible to formulate successful
definitions for words in a significant number of cases. However, one of the most
frequent criticisms of definitional theories of semantics is that no satisfying definition
of a word has ever actually been formulated.
The skepticism about the existence of definitions is so widespread, in fact, that many
researchers in disciplines closely related to linguistics, such as cognitive science and
artificial intelligence, have completely abandoned the idea that definitions even exist.
The overall rejection of definition outside linguistics as a defensible mode of meaning
analysis has various motivations.
Many of them derive from the problems involved in a psychological interpretation of
definitions as concepts, in which the structure of a definition reflects the structure of
the underlying concept (so that, for example, the concept BACHELOR could be said
to be the combination of the concepts UNMARRIED and MAN in just the same way
as the definition of bachelor might be thought to be “unmarried man”).
Linguistics, however, is not necessarily committed to a conceptualist interpretation of
definitions. As our initial typology of definition suggested, a definition can serve
many purposes, and revealing the apparent underlying conceptual structure of a
unit of language is only one.
The extreme difficulty of phrasing accurate definitions should be an embarrasment to
any theory in which definitions are a privileged mode of semantic analysis.
A classic case of definitional inadequacy is the proposed “definition” of bachelor as
“unmarried man”. This is the type of definition found in many popular dictionaries.
One problem here is that there are many types of unmarried male, such as widowers
or the Pope, whom we would not describe as bachelors. As a result, “unmarried
male” is not substitutable for “bachelor”, and the definition therefore fails.
Even though we typically think of many words as having concise definitions, and that
these definitions accurately convey the word’s meaning, more detailed investigation
reveals that our intuitions on this are in fact mistaken. A possible response here
would be to claim that it is only the extreme brevity of the definition of bachelor which
accounts for its inadequacy.
If a definer tries hard enough, satisfactory definitions can be achieved. This is the
point made by Wierzbicka (1996). According to her, the true definitions of most
ordinary words are significantly longer than the brief statements we are used to
reading in dictionaries.
2.7 Definition, understanding and use
In some domains of human activities, definitions function as the guarantors of the
consistency of language. This is particularly so in the area of Judicial terms: murder,
contract or fraud require clear definitions which fix their denotation by designating
exactly what does and does not count as an example of each.
Conversations and other examples of language proceed without the need for explicit
definition: this is only required to resolve confusions. And when we do ask for
clarification about the correct use of a word, nominal definitions of words’ meanings
are not usually either solicited or provided.
But definitions do take a central role in language use if we take concepts to be
essentially definitional in nature and assume that concepts are or enter into the
meanings of words.
If concepts correspond to word meanings, and word meanings can be captured in
definitions, then it is the definition which is in some sense activated during language
use. To claim that definitions are involved in language use in this way is not to claim
that they are so involved consciously.
We may be able to use a word appropriately, without being able to phrase a
satisfactory definition of it: the knowledge enabling correct use is unconscious. To
say, then, that concepts are or function as definitions is certainly not to say that we
consciously carry around a dictionary-like list in our heads.