Upload
ngodieu
View
214
Download
1
Embed Size (px)
Citation preview
HLW Chap1 rev 071125 circa 8000+ words, 26 pp.Rev after Atkins comments 050607; rev 071125
How Language Works – copyright David Cole 2005, 2007
Chapter 1: The Miracle of Meaning
1. introduction – the miracle of meaning2. Main Traditions in Semantics [bulk of rest of chapter, with critiques]
a. meaning not naturalizingb. hence Platonismc. problems with Platonism
3. Semiotics: man the meaning-maker4. Semiotics as Idealism5. Structuralism 6. Mentalist accounts 7. Critique of mentalism
a. basic considerations against mentalismb. Wittgensteinc. Putnam
8. What a Theory of Meaning should explain (Block, Lycan and Portner desiderata)9. Recap: Chapter summary
1. INTRODUCTION
Minute vibrations barely disturb the still air. And then thunder as a firing squad
terminates a life. The survivors stand quietly. The same pattern of minute vibrations
occurs in the air of a crowded theater, and there none remain quiet: patrons scream and
rush for the exits. Some smudges of carbon black appear on paper, and a reprieve is
granted, a war is ended, or a Continent divided. The miracle of language is the
mysterious power of the physically tiny; the significance of the seemingly insignificant.
The pen can be mightier than the sword – that is the paradox.
The miracle is meaning. The vibrations in the air have meaning. The marks on
paper are meaningful. What then is meaning, that can endow such physically puny
things, as are the physical embodiments of language, with such power?
1
2. Main Traditions in Semiotics
While the noises and marks that bear meaning are physical and clearly parts of
nature, meaning itself is hard to explain in terms of the physical. Meaning doesn't
naturalize well. For example, clouds come and go, but the meaning of "cloud" stays the
same. Meanings can stay fixed as the physical world is in constant flux.
At first blush, we might think the meaning of a word or sentence lies in its
correspondence with reality. Yet skies become cloudless, while “cloud” still has
meaning. A cloudless sky does not deprive the word “cloud” of meaning, even though
the word applies to nothing on that sunny day. A sentence (e.g. “it is two o’clock now”)
can at one moment correspond with reality and at another not, yet the meaning of the
sentence does not change. I can be mistaken in asserting that p, yet the meaning of what
I say is the same whether I am mistaken or not. “Unicorn” is meaningful even if there are
no unicorns and never were. Thus meaning does not appear to be correspondence with
reality after all.
So perhaps the meaning is in the mind of the speaker? Perhaps it is the speaker’s
belief? But I can lie, and not believe what I say, yet the meaning of what I say is
unchanged. And you can fail to believe my lie, so that no one has a belief that
corresponds to what I say. The very possibility of lying depends upon a lack of
equivalence between belief and meaning. Thus it is difficult to see how the meaning of
what I say an be something in the mind; but we will look more closely at some
“mentalist” approaches below.
The meaning of a sign does not appear to be any part of the meaningful physical
sign. There is no natural connection between a sign and its meaning (although in one of
2
the first known works on meaning and language, Plato’s dialog Cratylus, Plato discusses
the possibility that there is such a connection). But as the Swiss linguist Ferdinand de
Saussure emphasized, the choice of sign is arbitrary for any meaning (and any sign could
have had a different meaning than it in fact has). Since the relation of cause and effect is
not arbitrary, and the connection between sign and meaning is, we have further obstacles
to understanding meaning as part of the physical world.
The physical world is an ever-changing mechanism governed by fixed laws,
whereas the realm of meaning is a realm of unchanging meanings and arbitrary
connections to the physical. As a result, it is perhaps natural to look to the supernatural
realm and its permanence as the home of meaning. The result is Platonism. Plato and
many of his successors down into the Twentieth Century posited extra-mental eternal
Forms or Ideas to be the meanings of words. To explain the meaning of larger units of
language, such as sentences, philosophers (e.g. Bertrand Russell) posited “propositions”.
Two different sentences might express the same proposition, which thus could be
identical with neither sentence. Platonism has been an important tradition in semantics,
from Plato through Frege (whose work inspired much Twentieth Century philosophy of
language) and to the present time. Many current approaches, including those in “Formal
Semantics” that take meanings to be sets, such as sets of possible worlds, have their
metaphysical roots in Platonism. They look to something beyond and more stable than
the actual world as the home of meaning.
But Platonism has always had its discontents. It is difficult to see how something
nonphysical and completely outside the world can have effects within the world, and
meaning surely has effects. As we saw, the miracle of meaning is the significance here in
3
this world of the subtle physical differences of the marks and vibrations that make up
written and spoken language. Surely, one suspects, it is something about us, the users of
language, and not something outside the world that accounts for this difference. If we
suppose that there is a supernatural realm of Ideas or Propositions, we are left with the
formidable problems of explaining how vibrations in the ear connect with these Ideas,
their hypothesized meanings, as well as how these supernatural meanings affect brains
stuck between ears in the natural world. Thus over time and despite the difficulties
several naturalistic alternatives to Platonist semantics have developed.
3. SEMIOTICS: Man the Meaning Maker
Semiotics is the general theory of all kinds of signs, whereas semantics is the
theory of the meaning of the most important kind of sign system, natural language. A
major contemporary tradition in semiotics begins by noting that the marks and vibrations
have no power until someone, a consumer of language, responds to them. And so this
suggests that the proper place to look for meaning is in the mind of the sign beholder.
The consumer of signs endows them with their power. In the air, or on paper, they are
nothing. But resonating in a mind, they swell to become forces literally to be reckoned
with.
Thus the view that developed in the semiotic tradition in the 20th century is along
the following lines. Man is the meaning maker. Scientologists consume twenty pounds
of fish a day. Furthermore, this making of meaning is not confined to language. It is the
essence of our consciousness. The world for us is not just a distribution of matter in
space; it is a vast field of significant objects. To be conscious of something is to be
4
conscious of it as this or that - we endow it with meaning from the first moment that it
exists for us. That pile of wood is a Table, that lump of flesh is your lover or friend or
confidant, who means this and that and many other things to you.
On this approach then, signs get their meaning from sign consumers. For
example, the semioticist Daniel Chandler explains it thus:
We seem as a species to be driven by a desire to make meanings: above all, we are surely Homo significans - meaning-makers. And it is this meaning-making which is at the heart of the concerns of semiotics. In semiotics, 'signs' are meaningful units which take the form of words, images, sounds, gestures or objects. Such things become signs when we invest them with meaning.
Chandler makes clear that something is a sign with meaning only when a sign-consumers
“invest them with meaning”. The mind bestows meaning on these otherwise insignificant
and meaningless sounds, shapes and objects.
Chandler cites Graeme Turner on the essential role played by the sign consumer.
According to Turner, for something to qualify as a sign:
…it must have a physical form, it must refer to something other than itself, and it must be recognised as doing this by other users of the sign system. (Turner 1992,17).
The semiotic tradition stems from the work of two main thinkers who worked at the end
of the 19th Century: the American philosopher Charles S. Peirce, and the Swiss Linguist
Ferdinand de Saussure. Peirce takes the essence of a sign to be a three-place relation
between the physical sign (or its mental representation), the thing it represents (which of
course need not exist in the real world – there are signs on empty store shelves and signs
of the devil), and the sign interpreter. All three relata must be present for the sign to be a
sign. As Chandler explains:
For the analytical purposes of semiotics (in the tradition of Saussure), every sign is composed of:
5
a 'signifier' - the form which the sign takes; and the 'signified' - the concept it represents. Nowadays, the 'signifier' is commonly interpreted as the material (or physical) form of the sign - it is something which can be seen, heard, touched, smelt or tasted. The signified, on the other hand, is a mental construct - not a material thing at all. The relationship between the signifier and the signified is referred to as 'signification'….
[From "Semiotics for Beginners" section "Signs", by Daniel Chandler: http://www.aber.ac.uk/~dgc/sem02.html Accessed 13 November 1998.]
Or, as semiotics pioneer C. S. Peirce put it, "Nothing is a sign unless it is interpreted as a
sign".
This way of understanding meaning focuses on sign consumers. Subsidiary
approaches to meaning branch off from this focus. What is it about the sign consumer
that results in meaning? The approach that is often called “semiotics” will focus on the
interpretation or understanding of signs by sign consumers, the “mental constructs”.
These mental states endow the signs with meaning. They make them mean something for
the sign consumer. Another approach, tied to the behaviorist movement in Twentieth
Century psychology, will focus on the behavior of sign consumers. The behaviorist
account of the nature of signs noted that only behavior is observable, and the observable
is the basis of any science. Further, since children must learn language by observing
behavior, and since children are corrected in their own linguistic performance entirely on
the basis of how they behave, the mental as factored out, a non-player in the game that is
language. Since language learners and teachers see only behavior, meaning must come
down to observable reactions to situations, the contexts of language use. Finally, the
ultimate point of language is plausibly seen as the coordination of human behavior, and
so it is to behavior that we must look to understand meaning.
Among the main early representatives of the behavioristic approach were the
6
influential linguists Leonard Bloomfield and Charles Morris. Leonard Bloomfield (1967)
tells us:
The meaning of an expression is "the situation in which the speaker utters it and the responses which it calls forth in the hearer." (Language Allen and Unwin 1967 p. 139)
One can get a sense of the scientific behaviorist approach at high tide in this
definition of a sign by Charles Morris:
"If anything, A, is a preparatory stimulus [a disposition creating stimulus] which in the absence of stimulus objects initiating response sequences of a certain behaviour-family, causes a disposition in some organism to respond under certain conditions by response -sequences of this behavior-family, then A is a sign."(Signs, Language and Behavior Prentice-Hall 1946 p. 10; clarification inserted)
In other words, Morris holds that a sign is a kind of behavior-producing surrogate for the
original object - the sign produces the same response dispositions as an object would. In
that behavioral sense, the sign is “a sign of” the object. For example, the ringing of the
bell functions as a sign of food for Pavlov’s dogs – the bell ring calls forth the same
response as food. The calls of a Vervet monkey that signal the presence of overhead
predators produce the same hiding response in fellow monkeys as if they had themselves
seen the predators – and thus this role in producing behavior is the meaning of the Vervet
calls.
4. SEMIOTICS as IDEALISM
Unfortunately, these dominant traditions in semiotics are fundamentally mistaken
about the nature of meaning. Signs do not depend for their meaning on sign consumers;
and we can’t literally make meaning by just thinking things are significant or interpreting
them as meaning one thing or another. In my view, the pervasive and fundamental
mistake in the semiotics tradition is to confuse meaning with interpretation of
7
meaning. As we shall see, these are quite different. Interpretation is important, of
course, and the semioticists have done much interesting work on interpretation of
meaning. But conflating meaning and interpretation of meaning has seriously
compromised the semiotic tradition, by taking the essence of meaning to be what the
interpreter takes it to be.
Before considering the semiotic tradition further, let us step back and consider a
larger and older tradition in metaphysics and epistemology to which it is closely related:
Idealism. Idealism has two modern sources, the work of the British Empiricist George
Berkeley, and the work of his rough contemporary G. Leibniz, who was prominent in the
parallel Rationalist tradition that was ascendant on the continent of Europe, These
idealists held that the existence of things generally – tables, chairs, the physical world,
depends upon the minds of perceivers. In Berkeley’s famous phrase, “To be is to be
perceived”. Thus tables exist only insofar as some mind perceives them. Berkeley had
many arguments for his Idealist theory. Perhaps the simplest is that one can’t conceive of
an object that exists unconceived – to attempt to do so produces a contradiction.
However this, and all of the other arguments for Idealism, are unsound. (Notice that it is
impossible to name something that exists entirely independently of language, that is,
something that is never mentioned; but this hardly shows that all things depend upon
language for their existence.)
The semiotic tradition that we have been considering is closely related to
Idealism. On the semiotic view, meaning exists only insofar as someone takes a sign to
have meaning. However this view runs into several fundamental problems. Behavior in
response to a sign is generally just a reaction to the meaning that the signs already have.
8
The dogs react to the bell because in the past it was causally connected with the presence
of food, and the dogs noticed this. The vervet monkey calls are reliable indicators of the
presence of overhead predators – and a monkey who fails to notice this and interpret the
sign _correctly_ is a dead monkey.
The same holds for response to signs by humans. One is generally not at liberty to
creatively interpret a STOP sign or a poison warning. It is because a sign has meaning
that it is appropriate to respond to it one way rather than another. As noted earlier, the
fundamental problem that has plagued the semiotic tradition is the failure to distinguish
meaning from the interpretation of meaning, just as Idealists failed to distinguish chairs
from ideas of chairs. But what something means is one thing, and what Jim thinks it
means is another. Jim may be right or wrong about what the sign means. The meaning
exists independently of his interpretation.
Reconsider our starting example of a firing squad. If in response to the call
“Fire!” a member of the firing squad runs for a bucket of water, he has simply
misunderstood the meaning of the call. He does not literally “endow it with meaning”.
He thinks it means “there is a fire burning”, but his thinking this does not make the
commander’s call mean “there is a fire burning”. The runner simply fails to understand
what is going on. Similarly suppose that in a crowded theater someone sees the first lick
of flame and makes the very same call, “Fire!” In this case, someone who thinks he is
complying with an order and thus pulls out a gun and shoots a member of the theater
audience who happens to be standing against the theater wall, also has a serious –
possibly fatal – misunderstanding problem. As these and myriad other examples show,
meaning and the interpretation of meaning are simply two completely different things.
9
We interpret things as having meaning all the time, but this is not “endowing with
meaning”, any more than wishing makes it so. It may seem as if the interpreter creates
the meaning, as it seemed to the Idealists that perception created the world “for us”. But
things are not always as they seem.
Idealism is false. There is a real world outside our minds, indeed, a really big
world of which we are merely small parts. Unfortunately, this same subjectivist fault of
Idealism carries over to the dominant tradition in semiotics. The correct approach to
meaning is realism – the view that meaning exists apart from the mind. Interpreters can
be right or wrong, can understand, or fail to understand. We can discover what
something means – the meaning exists prior to discovery by the meaning-consumer. Let
us call this stance regarding signs in general “semiotic realism” and let us call the
corresponding view in semantics of language “semantic realism”. Realism was the
framework of other several main traditions in semantics, and will be the framework of
this book. I will set out what I take to be a correct realist account of meaning in the next
chapter.
5. The Structuralist alternative
A second main Twentieth Century approach to meaning sees it not as Platonic,
and not as lying in the mental or behavioral reactions of sign consumers, but rather as
lying within the structure of language itself. This movement in semiotics and semantics
is in part the product of a holist tradition in European philosophy that has its roots in the
holistic idealism of Hegel. Holism approaches things as systems, in which individual
elements get their properties or “value” from their relation to other things in the system.
10
Thus we may think a silver dollar is intrinsically valuable, but it gets its value only in
relation to other things – the things other agents will exchange for the silver dollar. And
social roles – policeman, fireman, thief, or Hegel’s master and slave – are defined only in
relation to others.
In semantics this holistic view takes meaning to be internal to the structure of
language itself – a structuralist approach. On this view, words get their meaning through
their relationship to other words. There are two main approaches along these lines. One
emphasizes sentences, and in particular the role a term plays in inferences. Thus the term
“electron” or “mass” is defined by how it enters into the laws and inferences of the theory
in which it is embedded. Hence the term “mass” means one thing in Newton’s physics
and another in Einstein’s. This holism was in part the basis of Kuhn’s theory of scientific
revolutions, and is connected with the views of Quine and Wittgenstein and many others.
But here I will focus on another holistic approach called “structuralism”. The
structuralist approach to meaning is founded on the work of the Swiss linguist Ferdinand
de Saussure, and has been especially influential in anthropology and in 20th Century
French thought.
Saussure was initially struck by an important fact about the phonology of
language. It is not the “absolute value” of a sound that makes it an instance of a
particular phoneme, but rather its contrast with the other elements. This is perhaps easier
to see in Morse code. Morse code was developed to allow sending messages over a
telegraph (many other telegraphy schemes were tried). Morse’s code consists entirely of
dots and dashes – long and short clicks of the telegraph key. Now there is no absolute
definition of the duration of a dot – all that matters in a message is that the dots are
11
clearly shorter than the dashes. One man’s dot can be another man’s dash – they are
defined relationally.
Saussure extended this thought to semantics, and his focus was on the meaning of
words and the contexts in which they can appear (rather than inference relations). Words
are defined in terms of other words. Each word gets its meaning especially by contrasts,
its difference from other words. Structuralists emphasized two main relations between
words. First, there are the other words that a word goes with, its syntagms. Syntagms are
typically of another syntactic class, and so this relationship in effect determines the
syntactic class of the defined word. For example, in a two word sentence, “Dogs” can be
followed by a host of expressions, like “bite” or “bark”, but not “red” or “by”. This
relationship is syntagmatic. These relationships help make “Dog” the noun-ish sort of
word it is, and more specifically make it the loud-living-thing sort of word that it is.
At the same time, in any sentence for each word there corresponds a class that
could substitute for it and with which it contrasts. For example, in the two word sentence
headed by “Dog”, “Dogs bite”, the word “bite” belongs to a class that includes “run”,
“sniff”, “bark”, “shed”. Each of these words is a contrasting alternative to “bites”. This
relationship between words is said to be paradigmatic.
Each word or expression is defined by the totality of the syntagmatic and
paradigmatic relationships in which it is embedded. Thus structuralism is an extreme
form of Holism. Meaning is given only by the entire structure of language. The
influential Continental philosopher Hegel had made similar claims about knowledge and
understanding, and it is no accident that structuralism and holism have been dominant
strands of thought on the Continent of Europe.
12
Structuralist approaches were not confined to the Continental tradition. Others
have noted that when we define a word, when we give its meaning, we produce language.
A dictionary, which gives the meaning of words, consists of a web of links of words to
words. British semanticist Geoffrey Leech (1974) writes:
One of the keynotes of a modern linguistic approach to semantics is that there is no escape from language: an equation such as cent = hundredth of a dollar or salt = NaCL is not a matching of a linguistic sign with something outside of language; it is a correspondence between two linguistic expressions, supposedly ‘having the same meaning’. The search for an explanation of linguistic phenomena in terms of what is not language is as vain as the search for an exit from a room which has no doors or windows, for the word ‘explanation’ itself implies a statement in language. Our remedy, then, is to be content with exploring what we have inside the room: to study relations within language, such as paraphrase or synonymy…. (Semantics, p. 5).
Leech goes on to distinguish seven types of meaning, the main one being
Conceptual or logical meaning. This form of meaning is based on two general principles
found in all areas of language, “Contrastiveness” and “Constituent Structure”.
Contrastiveness is the binary contrasts or dichotomies we use to specify meaning --
edible/inedible, male/female, animate/inanimate. Constituent Structure is the way we
build up larger linguistic units out of smaller ones. For example, a bachelor is a not-
married adult male human. In modern linguistics, at the beginning of the last quarter of
the 20th century, this was perhaps a standard view of the core of meaning.
These structuralist views have the advantage over the subjectivist semiotic
tradition in that they reject idealism -- meaning is taken to be an objective feature of the
language itself. Thus it can be seen as a form of semantic realism, a virtue it shares with
Platonism.
But as many critics have noted (e.g. Devitt and Sterelny 1999), the structuralist
approach locates meaning in the wrong place, and loses sight of the purpose of language.
13
Meaning is not an internal feature of language. The whole point of language generally is
to point to something beyond itself. We use language to find things out, and we use it to
tell others what we know and think and desire. It thus is more plausible to hold that
language has the structural features that it has only in order to communicate meaning, but
the internal structural features do not themselves constitute meaning.
So surely meaning has something to do with communication and the persons who
are trying to communicate. Presumably speakers choose the signs they do because they
have something in mind, and would like to communicate it to some one else. Let us
explore this last influential way of understanding the nature of meaning.
6. Mentalistic accounts of Meaning
An important way of understanding meaning stems from empiricism, and has
been very influential in Anglo-American thought from the Enlightenment on. In the
British Empiricist tradition, meanings were not thought to be in a supernatural realm, nor
were they held to be in the minds or behavior of sign consumers, nor were they to be
found in the structure of language. Central to the empiricist tradition is the following
theoretical viewpoint: experience fills the mind with sensations. From sense experience
the mind derives ideas, (often conceived as images copied from sense experience).
Language is a tool used to communicate ideas. Words then stand for ideas in the minds
of those who use them – sign producers. As John Locke puts it in his Essay Concerning
Human Understanding, written over a long period in the last quarter of the 17th Century.
According to Locke, “the … use of words is to be sensible marks of ideas; and the ideas
they stand for are their proper and immediate signification”. The following important
14
passage from Chapter 2 of Book III of Locke’s Essay, which is titled “Of the
Signification of Words”, makes clear Locke’s intuitive view and the argument for it:
Words are sensible signs, necessary for communication of ideas. Man, though he have great variety of thoughts, and such from which others as well as himself might receive profit and delight; yet they are all within his own breast, invisible and hidden from others, nor can of themselves be made to appear. The comfort and advantage of society not being to be had without communication of thoughts, it was necessary that man should find out some external sensible signs, whereof those invisible ideas, which his thoughts are made up of, might be made known to others. For this purpose nothing was so fit, either for plenty or quickness, as those articulate sounds, which with so much ease and variety he found himself able to make. Thus we may conceive how words, which were by nature so well adapted to that purpose, came to be made use of by men as the signs of their ideas; not by any natural connexion that there is between particular articulate sounds and certain ideas, for then there would be but one language amongst all men; but by a voluntary imposition, whereby such a word is made arbitrarily the mark of such an idea. The use, then, of words, is to be sensible marks of ideas; and the ideas they stand for are their proper and immediate signification.
Words, in their immediate signification, are the sensible signs of his ideas who uses them. The use men have of these marks being either to record their own thoughts, for the assistance of their own memory or, as it were, to bring out their ideas, and lay them before the view of others: words, in their primary or immediate signification, stand for nothing but the ideas in the mind of him that uses them…
This is the classic statement of the ideational theory, on which words stand for
ideas in the mind of the language producer. This theory in one form or another has
probably been the dominant view in the Anglo-American philosophical tradition. It sees
meaning not as internal to language structure, and not as something bestowed on signs by
sign consumers, but as something in the minds of sign producers. By mid-century,
Susanne Langer could nonchalantly note that “In talking about things we have
conceptions of them, not the things themselves, and it is the conceptions, not the things,
that symbols directly ‘mean’” (Philosophy in a New Key, 1942, pp 60-61). Louis
Salomon (1966) puts it this way:
To be sure, no two observers have exactly the same mental picture of Julia Ward Howe, or of a table; but each one attaches a label, or a symbol, to his own mental picture
15
(not directly to the referent, unless the referent is the speaker’s inner feeling). To just such an extent as a large number of people indicate by their use of the label that they are referring to approximately similar mental pictures, we call that symbolic label a word, and a set of such symbolic labels, with customary procedures for combining them, a language (Semantics and Common Sense, p. 11).
One variation that developed was a dispositional account of meaning in terms of
how mental states are associated. For example, Ogden and Richards (1930), in their
influential work The Meaning of Meaning, held that the meaning of a word is a ‘recurrent
set of mental events peculiarly related to one another so as to recur, as regards their main
features, with partial uniformity“(57).
The Empiricists from Locke to Russell had ideological reasons for their ideational
view. Our ideas are all we really know about. How much wood can a woodchuck
chuck? We have no direct knowledge of things outside the mind (if there even are such
things, or if talk of them makes sense). Such things are at best an inference, and at worst
“the monstrous offspring” of two contradictory principles, as David Hume puts it.
Empiricism was driven by epistemic considerations, by what we could “really” or
“directly” know. Since we at least know what we mean, and since all we can directly
know are the contents of our own minds, meanings must be in the mind. Thus language
could not “directly” refer to things out in the world, and the function of language was to
provide information about the state of the speaker’s mind, not about the state of things in
the world. In addition, this appears to be the only way to handle the obvious: “It is
raining” means the same whether it is raining or not, and so appears to be connected most
directly with the speaker’s belief and not the actual state of the weather.
7. The Critique of Mentalism
16
Despite its influence, the Empiricist view has the disadvantage that it makes
meanings subjective, and the meaning of words hidden from all but "he who uses them."
Bertrand Russell, probably the most important Anglo-American philosopher in the first
half of the 20th Century, at one point held that all apparent names in ordinary language,
such as “Bertrand Russell”, merely stand for descriptions, and the only true names are
demonstratives that designate directly the current sensory experience (“sense-data”) of
the speaker. Russell had few qualms about accepting the apparent consequence of the
ideational view that meanings must vary from speaker to speaker. But that consequence
does make communication problematic.
This mentalist view came under attack in the 20th century from a variety of
directions, including many of the most prominent philosophers or the era. Russell’s
student Ludwig Wittgenstein attacked the view in what is now generally known as “The
Private Language Argument”, and pointed toward pragmatic accounts of meaning in
terms of the many social uses to which we put language. As we have seen, Saussure,
Godfather of French structuralism, held that meaning is defined by the relations words
have to each other within a language, in contrast to the subjectivism of the mentalist
view. Behaviorists, including psychologist B.F. Skinner and his philosopher friend
W.V.O. Quine, were suspicious of mental entities. As we have seen, behaviorists avoided
hypothesizing mental entities by constructing behavioral accounts of meaning, or even
accounts in which behavior is a substitute for meanings. Later Saul Kripke and others
also influentially argued that descriptions in the head of language users weren't
determining reference. Hilary Putnam advanced Twin Earth arguments, designed to
show that meaning, at least whatever determined reference, was not inside the heads of
17
language users.
Before considering these views, let us note several reasons for being suspicious of
the Lockean view that words stand for ideas in the mind of the user. In the first place,
intuitively the word “beer” was introduced to stand for beer, not the idea of beer. We
seek information about beer, and only derivatively (perhaps after consumption) about
ideas of beer. Note that in these last few sentences, we assume I am talking about both
beer and the idea of beer – and I am using the word “beer” to refer to beer. When I say
“there is a beer in the fridge”, I am intending to say something about beer and my fridge,
not about my ideas. We have a perfectly good expression that stands for the idea of beer
– namely “the idea of beer”. Thus when we wish to talk about the idea of beer, we can.
And when we use the word “beer” simpliciter, we talk about beer.
Second, suppose all my words stand for ideas in my head. Then any use of my
word “idea” stands for my idea. So then what do I mean when I say “your idea”? Since
the only ideas I have are my ideas, and the very meaning of the word idea is my idea,
how can I refer to your ideas? It appears to be impossible. I can’t even understand that
my idea of your idea is of your idea – since I only have my ideas.
Third, consider how we teach language. When I teach a child the word “ball”, I
show the child a ball, not an idea of a ball. The child learns to use the word “ball” to
announce the presence of a ball or to ask for a ball – not the idea of a ball. The kid
already has an idea of a ball in asking for a ball. What the kid wants is a ball, and so it is
reasonable to suppose the kid uses the word “ball” to refer to a ball, not an idea of a ball.
Things going on in the child’s head are important to its being able to use language, but
they do not constitute the reference of the terms.
18
Finally, and perhaps most important, what makes a sentence such as “There is a
beer in my fridge” true or false, correct or incorrect, is not what is in my head, but what is
in my fridge. Truth is the most important semantic property, and it is based on a relation
between language and reality (as we shall discuss further in the chapter on Truth). Our
heads are part of that reality, and sometimes we do indeed wish to make claims about
what is in our heads. But in general, language is directly about the world and is not a
poor step-sister to mental representation, as we can see in the way we assess truth and
falsity, comparing the claim made with the way the world is.
One of the unfortunate hallmarks of the empiricist tradition generally was to
conflate semantic and epistemic issues. The empiricists start with the idea that all we
know are ideas. Then they conclude that since we know what we mean, meanings must
be ideas. Semantics was held captive to the Way of Ideas.
The empiricist presuppositions in epistemology are mistaken, as is their
consequent conflation of semantic and epistemic issues. We know the world we live in,
and it is in virtue of this knowledge that we are able to live in the world. Language
facilitates this acquisition of information about the world, as we will see in the next
chapter.
With these general complaints about mentalist approaches as background, let us
now consider in turn the specific attacks made by Wittgenstein and Putnam on the
mentalist approach to semantics. The exact nature of Wittgenstein’s argument is
controversial. I will present a fairly straight-forward summary of a main point in
Wittgenstein’s Philosophical Investigations (published posthumously in 1953, with much
of it written quite a bit earlier). First, if words stood for inner mental events in the mind
19
of the speaker that could not be observed by others, there would be no way for auditors to
know what the referent was. Second, since we teach language and correct others on the
basis of their behavior, the events inside their heads are irrelevant. Those events drop out
of the equation, and can be imagined to vary arbitrarily. Wittgenstein uses the analogy of
a group of people each of whom has something in a box that each calls his “beetle” but
that none of the others can ever see. He concludes that the character of the thing in the
box is irrelevant to the language game. Similarly, even if there is something in the head
of language users, exactly what it is does not matter – all that could matter is the role it
plays in producing publicly observable behavior.
Finally, Wittgenstein argues that if meanings were private events in the head of
the language user, not even the language user could have a criterion for his own correct
and consistent use of language. To use one’s own memory of what one meant is like
checking the accuracy of a story in a newspaper by buying another copy of the paper and
comparing. We have only our own self-testimony as to what was correct, and no way of
knowing that we are consistent in our own story.
Wittgenstein goes on to sketch a positive approach to understanding meaning in
terms of “tools” and “language games”. Like tools, language is used in a myriad of ways,
and we should look to those uses and the ways of life that surround them in order to
understand meaning. In addition, language is inherently social, and the criteria of correct
and incorrect use are socially applied. The Lockean semantic tradition is entirely
individualistic, subjective, and ultimately makes no sense. The social role approach to
understanding language was explored by J. L. Austin, H. P. Grice, and John Searle, in
their development of Speech Act Theory. We will return to Speech Act Theory later.
20
Hilary Putnam’s well-known Twin Earth thought experiment came a quarter
century after Wittgenstein’s work in Philosophical Investigations. One of the main
sources in print is Putnam’s paper “The Meaning of Meaning”, a recycling of Ogden and
Richards’ book title. Putnam uses a thought experiment. Suppose there is a planet far far
away exactly like Earth in every respect. Everything on Earth has an indistinguishable
doppelganger on TwinEarth. It is like a parallel universe, with one small difference.
Instead of H20 in the lakes, streams and bathtubs of TwinEarth, there is a substance, call
it XYZ, that functions the same. XYZ appears identical to H20, and prior to the
development of chemistry no one on either Earth or TwinEarth could tell these tasteless
colorless potable liquids apart. Now consider Oscar on Earth, and his twin TwinOscar on
TwinEarth at some time before the development of molecular chemistry. Both make the
same sounds when they speak, and both have the same brain structure and mental images.
Both associate the same descriptions and thoughts with their words “water”. But when
Oscar uses the word “water” he refers to H20, and when TwinOscar uses the word
“water” he refers to XYZ. Thus their mental states do not determine reference – or, as
Putnam famously put it, “meanings just ain’t in the head”.
But in the on-going debates about how to understand meaning, this dislodgement
of meanings from the head was only temporary. By the end of the 20th Century,
meanings had multiple homes in the head. Followers of H. Paul Grice analyzed speaker
meaning in terms of communicative intentions. Many others were directly influenced by
developments in linguistics, which suggested that the origin of sentences was a semantic
representation in the mind, built out of concepts. Many in cognitive science, including
philosophers Ned Block, Gilbert Harman, and Paul Churchland, embraced forms of
21
conceptual role semantics. On the CRS approach, concepts get their meaning by the role
they play in relation to other concepts. Such an approach is holistic, and resembles the
structuralism we discussed earlier, except that it is within the psychology of a single
individual. Jerry Fodor, a relatively lone but persistent voice, has been a critic of CRS,
and has proposed an alternative atomistic account of concepts. On this view concepts are
in the head, and concepts (or their tokens) are word meanings. As Fodor notes, CRS has
the apparent disadvantage that CRS can vary considerably from individual to individual -
and the same individual over time.
How did it come to pass that meanings crawled back in the head? Well, I think
there were two main factors. One was disenchantment with the verificationist basis of
the objections to meanings in the head. One form of verificationism is the view that the
only meaningful statements are ones that can be publicly verified, a complete inversion of
Locke’s mentalistic view. The Private Language Argument was seen by many of its
critics to depend ultimately on a dubious assumption about public verification of
meaning.
The other factor leading to a revival of mentalism was the apparent implication of
Chomskyan approaches to language - lots of language-related stuff was in the head, and
was innate. Just extend this nativism to concepts, a la Fodor, and you resolve the
skeptical problem Wittgenstein raised. There is no more doubt about what is going on in
someone's head when she yells "Fire!" then there is doubt that she has a heart hidden
inside her chest. The alleged skeptical problem that motivated rejection of mentalism
was itself rejected.
Thus at the end of the 20th Century our understanding of meaning is
22
embarrassingly rich. We have four or five incompatible main approaches, each with
many adherents. Mentalist approaches locate meaning inside speakers, sign producers.
Semiotic approaches locate meaning in interpretation, inside sign consumers.
Structuralist approaches locate meaning inside language itself, apart from both producers
and consumers. Formalist approaches locate meaning in the realm of sets, possible
worlds, and other denizens of the Platonic realms. But I believe that unfortunately all
these approaches are wrong. On these accounts meaning is still a miracle – a Leibnizian
miracle of synchronous innate lexicons, on the mentalist accounts, or miraculous mental
projections of meaning by sign interpreters, like semantic rays from the Eye of Sargon,
on the semiotic or hermeneutic approaches, or the miracle of a levitating structure of
language floating free of the world, on structuralist approaches. In the next chapter, I will
set out an alternative account of meaning that seeks to dispel the miracle by focusing on
the basic capacity of language to carry information about the world and thereby be of use
to make true and false statements, make requests, express thought, and influence sign
consumers. Thus the basis of meaning will be seen to lie in special features of the
sources of signs, which correlate them with states of the world.
8. WHAT A THEORY OF MEANING SHOULD EXPLAIN
Before we set out such a theory of meaning, we would do well to have an idea of
what a theory of meaning should explain. Philosopher Ned Block has set out goals for a
theory of meaning, the various things that a theory should explain. While Block is
primarily interested in the semantics of mental representations (the contents of beliefs, for
example), his desiderata are of more general interest. I think they provide a useful list of
23
objectives for any theory of meaning.
Block’s desiderata for a semantic theory are set out in his article “Advertisement
for a semantics for psychology” (pp. 616-8). As the title suggests, the goal is a semantic
theory for psychology, but I believe most of these apply to linguistic semantic theory –
and almost all really are desiderata for any theory of meaning. A theory of meaning
should:
1. Explain the relation between meaning and reference/truth.
2. Explain what makes meaningful expressions meaningful.
3. Explain the relativity of meaning to representational system. [That is, that a sound
sequence can have different meanings in different languages]
4. Explain compositionality. “The meaning of a sentence is in some sense a function of
the meanings of the words in it (plus the syntax of the sentence). What, exactly, is the
relation between the semantic values of sentences and words? Is one more basic than the
other?”
5. Fit in with an account of the relation between meaning and mind/brain [for example,
meaning should fit with what it is to grasp or understand meanings].
6. Illuminate the relation between autonomous and inherited meaning. [This distinction
comes from John Searle and John Haugeland. Representations that require no translation
or transliteration have autonomous meaning (as do representations in the brain), whereas
representations that do require translation or transliteration have inherited meaning - such
as inscriptions on a page, which must be read or heard to be understood.]
7. Explain the connections between knowing, learning, and using an expression, and the
expression’s meaning. [This seems close to 5]
24
8. Explain why different aspects of meaning are relevant in different ways to the
determination of reference and to psychological explanation. [For example, consider a
name and an indexical or demonstrative that refer to the same thing – sentence with latter,
e.g. “Kill him”, might cause me to act while same sentence with a name with the same
referent might not (“Kill Bill”), because I did not know the referent of the name. In these
cases there can be important difference in the causal effects of two externally
synonymous sentences. This is related to problems pointed to by Frege at the end of the
19th Century].
In addition, I include some more standard virtues that William Lycan uses in this
intro text to assess theories of meaning, and that linguist Paul Portner 2005 (What is
Meaning?) cites for his own “possible worlds” semantic theory (an approach I would
lump with Platonistic theories.) Portner argues (16-22) that it is a virtue of a theory of
meaning to:
1. Account for the semantics of logical operators – words like “and”, “or”, “not”.
2. Account for “intuitive semantic relationships” such as: sentence synonymy,
contrariety (a pair of sentences such that it cannot be the case that both are true),
entailment, contradiction, and tautology.
3. And finally, account for the utility of language in human affairs.
Showing how a foundational approach to meaning can satisfy these many desiderata will
be the task of the rest of the book. While I will attempt to account for logical operators
(especially in Chapter 3), and the intuitive semantic relationships, I will always have an
eye on the need to account for the importance of linguistic meaning in our lives – the
mystery of the significance of tiny airborne vibrations and little marks on paper.
25
RECAP
We began this chapter by noting the miracle of linguistic meaning, and the
difficulties that face us in trying to understand meaning as part of the natural world. The
bulk of the chapter was devoted to an overview of past and current attempts to forge this
understanding, and the problems that face these views. After a first brief look at
Platonism, we considered in turn the main approaches of the last century. Second was the
semiotic or hermeneutic view that sees the sign consumer as the locus of meaning. We
then noted the parallels between this view and Idealism, and argued against both.
Structuralism provided an interesting alternative that saw meaning as a feature of
language itself, but which unfortunately severed the link between language and reality –
an essential link, as we will see in the next chapter. We then turned to consider the
Mentalistic tradition from Locke to contemporary thinkers such as Fodor, as well as
important criticisms of various aspects of the mentalist or individualist approach that
were developed by Wittgenstein, Kripke, and Putnam. Finally, we turned to
consideration of what a theory of meaning should explain, and desiderata (from Block,
Lycan and Portner) for a semantic theory.
In the chapter to follow I will develop a theory of meaning and I will begin to
show how it accounts for these various questions. Two of the subsidiary goals in the
chapters after that will be to show how the semantic theory explains the nature of Truth,
and to explore some of the implications for the relation of language and thought.
26