This is the sixth installment in
MIT’s series of Jean Nicod Lectures, a series that is based on the lectures
delivered annually in Paris by a leading philosopher of mind or philosophically
oriented cognitive scientist, with past participants including Jerry Fodor,
Donald Davidson, John Searle, and Daniel Dennett.  Millikan’s lectures were
presented in 2002 and address the problem: "What are the varieties of
meaning? And what do they have in common, so as to be treated together under
one cover?" (ix).  Millikan’s answer, briefly put, is that there are
various forms of meaning — as is evidenced by the fact that we say that James
meant to harm his sister, that a chainsaw is meant for cutting wood, that the
word ‘evil’ means ‘wicked or highly immoral’, that black clouds mean rain, etc.
— and although there is no single thing that all these varieties of meaning
have in common, they cannot properly be understood without grasping their many
interconnections.   

I.  Purposes and Cross-Purposes

The book is divided into four
parts, the first of which consists of two chapters in which Millikan
concentrates on the notion of ‘purpose’.  She begins (in Chapter 1: "Purposes
and Cross-Purposes of Humans") by dismantling the commonsense view that a
principled distinction can be drawn between ‘real’ purposes and merely
metaphorical purposes (or between a person’s purposes and merely biological or
natural purposes), as when we say that a person’s consciously formed intention
not to blink is a genuine purpose, whereas the purpose of the eye-blink reflex
(to prevent foreign objects from entering the eye) is a purpose only by analogy
or metaphorically.  Millikan argues, instead, that consciously formed purposes
and biological purposes are on an equal footing, and purposes of all kinds are
entirely natural insofar as all "have their origin in adaptation by some
form of selection" (13).  Millikan then (in Chapter 2: "Purposes and
Cross-Purposes of Memes") turns to ‘memes’ (i.e., cultural artifacts that
are reproduced by imitation rather than genetically — e.g., dress fashions,
ideas, values, and forms of expression), arguing that the underlying purpose or
function of memes is to facilitate social coordination.

II.  Natural Signs and Intentional Signs

The second part of the book begins
with a discussion (in Chapter 3: "Local Natural Signs and Information")
of ‘natural signs’, signs which cannot be false or unsatisfied (unlike ‘intentional
signs’, which can).  For example, red spots all over Tommy’s face mean (or
naturally signify) that Tommy has the measles, but only if he does in fact have
the measles.  Natural signs, obviously, carry ‘natural information’, that is to
say, information about a segment of the natural world, and they do so — in Millikan’s
view — because in each case "it is possible for a true belief to be
reached about one thing [e.g., Tommy’s health] from knowledge of the other
[e.g., red spots]€¦, where the truth of the belief reached will not be
accidental because the connection in thought correctly and nonaccidentally
tracks a dependency in nature" (p.37).  For a correlation between As
and Bs to be nonaccidental, Millikan adds, there must be a reason why
the correlation persists or replicates itself over other domains of time and
space.

Millikan then
proceeds (in Chapter 4: "Productivity and Embedding in Natural Signs")
to discuss two features of natural signs: their productivity, that is, their
capacity to say new things or provide new information; and their embeddedness,
that is, the capacity of one natural sign to be embedded within another, as
when tracks in the woods are a sign of the presence of quail, while a photo of
these tracks would be a sign of that sign of quail.  

The subject of
discussion then turns from natural signs to intentional signs, beginning with
Chapter 5: "Teleosemantic Theories".  Millikan first points out that
intentionality (or how a representation could be of or about
something) is not explained by a teleological theory, which only aims to
explain how representations can be false.  What account, then, do teleological
theories offer of misrepresentation?  According to such theories, notes Millikan,
"false representations are representations, yet they fail to represent,"
just as "something can be a can opener but be too dull and hence fail to
open cans" (64).  The strategy here is to avoid the reification of special
‘intentional objects’: in cases of false representation, the subject is not
representing something that doesn’t exist, but rather fails to represent
anything at all.  In short, Millikan’s view is that "teleosemantic
theories are piggyback theories" (66).  In other words, such theories do
not seek to provide an account of intentional representation, but instead ride
on some basic theory of representation (e.g., a causal theory, an informational
theory), to which is added: (i) the teleologist’s claim that for a
representation to be an intentional representation it must be a function
or purpose of the system that produced it, and (ii) the teleosemanticist’s
account of what a false representation is, given the underlying theory of
representation.  "That is all teleosemantics amounts to," writes Millikan
(67).

Millikan, therefore, next considers (in Chapter 6: "Intentionality")
what base theory of the representing relation the teleologist could offer on
which to build an account of intentional representation.  In particular, she
examines Dretske’s proposal of placing natural signs and the natural
information they carry at the base of a theory of intentionality — on this
proposal, natural signs whose function it is to carry natural information come
to represent intentionally, they become intentional representations.  But
Millikan finds fault with such a proposal, on the grounds that it wrongly
assumes that the intentional representation-producers have as their purpose to
produce natural signs; rather, on Millikan’s view, "when they [intentional
sign-producers] perform their functions by their normal mechanisms they produce
natural signs" (76).  Millikan then proceeds to outline her theory of
intentional representation, identifying in the process three kinds of
intentional representations: descriptive (which represent or describe what is
the case), directive (which represent or direct what is to be done), and
pushmi-pullyu (which are both descriptive and directive).

Millikan closes Part II with a discussion (in Chapter 7) of
intensionality, with particular reference to intensional contexts. 
Customarily, intensional contexts are described as contexts in which
coreferential expressions are not substitutable salva veritate. 
Millikan, however, holds that there is a more illuminating way of thinking
about intensional contexts: "The phenomenon here is not that one cannot
substitute coreferential terms without change of truth-value, but that the
grammar alone does not prove that one can.  Whether one can or not is a
pragmatic matter" (95).

III.  Outer Intentional Signs

The underlying
thesis of Part III is the claim that "conventional signs used for their
conventional purposes usually are read in exactly the same way that natural
signs are read" (109).  The opening chapter (Chapter 8: "Linguistic
Signs Emerge from Natural Signs") kicks off the project of highlighting
the similarities between the two kinds of signs by outlining how conventional
language signs evolve from natural signs without losing their character of
being also natural signs.  This is followed by one of the more provocative
parts of the book (Chapter 9: "Direct Perception through Language), where
Millikan advances the claim that understanding language is at root another form
of direct perception of the world.  She hopes that this claim, when conjoined
to the view that perception is a way of understanding natural signs, will help
support the view that understanding language is very much like understanding
natural signs. 

Millikan begins
by rejecting the traditional notions of ‘direct perception’ (according to which
sense impressions are directly and infallibly represented by the mind) and ‘indirect
perception’ (according to which representations of the outer world are derived
by the use of inference).  She nevertheless retains the distinction between
direct and indirect perception, but reconceives it in terms of translation
rather than inference, so that a representation of the outer world that is
derived by a series of fallible translations from one sign into another counts
as an indirect perception.  Millikan then proceeds to her central thesis, that "coming
to believe something by being told it is so, in the typical case, is the
formation of a direct perceptual belief" (120).  Suppose, for example,
that I come to believe that my son, Johnny, has just arrived home from school. 
A number of factors may compel me to adopt this belief — e.g., I could simply
observe Johnny coming through the door, or I could hear him shout, "Dad, I’m
home!", or I could hear my wife say to me, "Johnny’s home".  On Millikan’s
view, each of these ways of forming the belief, ‘Johnny has arrived home’, are
equivalent in directness of psychological processing.  This may or may
not be the case — I leave that for psychologists to decide.  The problem,
however, is whether ‘perception through language’ (e.g., hearing what is going
on in the world through the medium of other people’s speech transmission
systems) can plausibly be assimilated to ‘ordinary perception’ (i.e., seeing
what is going on in the world through the medium of normally surrounding
light).  For one thing, the reliability of the latter form of perception is
often greater than the reliability of the former kind (though Millikan does try
to meet this objection on pp.124-25).

In the following
chapter (Chapter 10: "Tracking the Domains of Conventional Signs") Millikan
argues that, just as you can tell what you are seeing through a pair of
binoculars without thinking about what is inside the binoculars and why they
work, so you can tell what a speaker’s words mean without thinking about the
speaker’s "insides" (his mind or intentions) — grasping the speaker’s
meaning only requires you to think the same content that the speaker
purposefully communicates.

The remaining chapters of Part III
concentrate on different ways of drawing the semantics-pragmatics distinction
(with Millikan arguing that this distinction typically supervenes upon another,
and necessarily vague, distinction between conventional and non-conventional
uses of language) and on the semantics of indexicals, demonstratives, and
referential descriptions.  

IV.  Inner Intentional Signs

In the final section of the book
Millikan looks at ‘inner representations’ and in particular at the question of "how
and why, during the evolution of perception and cognition, organisms have
acquired inner representations that are more sophisticated than pushmi-pullyu
signs" (157). 

Much of the opening chapter of this
section (Chapter 13) is preoccupied with the various features of pushmi-pullyu
signs, or P-Ps, which represent facts and give directions or represent goals,
both at once.  Bee dances, for example, tell where the nectar is and at the
same time tell where the watching bees are to go.  Indeed, Millikan conjectures
that all intentional signals used between nonhuman animals are P-Ps.  One of
the disadvantages, however, of being a purely ‘pushmi-pullyu animal’ is that "such
an animal does not represent its goals in a format that enables it to know
whether or when it has reached them" (169).

Next (in Chapter 14: "Detaching Representations of
Objects"), Millikan focuses on the dorsal and ventral visual systems in
order to show how the descriptive part of a P-P (e.g., the ventral system,
which detects what objects the animal confronts) may come apart or be detached
from the directive part of a P-P (e.g., the dorsal system, which directs the
movements of the animal).

After a brief discussion (in
Chapter 15) of space and time — where Millikan marshals experimental support
for the view that "just as many animals construct representations of the
spatial layouts in which they live, many construct representations of
regularities in the temporal layouts in which they live" (187) — the
subject turns to goal state representations.  In Chapter 16 ("Detaching
Goal State Representations"), Millikan argues that P-Ps can give way to "a
common system of mental representation€¦in which projected goal states,
objectively represented future states, and objectively represented present
states can all be expressed" (197).  Millikan then investigates (in
Chapter 17) how representations of projected goal states are generated.

The last two, and very interesting,
chapters delve into the ways in which the representational capacities of humans
differ from those of nonhuman animals.  In Chapter 18 ("Limitations on
Nonhuman Thought"), Millikan observes that, unlike humans, "[nonhuman]
animals perceive the world only as a subject of practical concern, not as a
subject of theoretical judgment" (219).  This is evidenced by the fact
that human, but not nonhuman, animals collect facts and develop skills for
which they have no practical uses ("we appear to be compulsive collectors
of all kinds of representational junk" (215)).  However, this ability to
represent things far removed from practical activity requires a distinct
representational system that is unlike anything found in nonhuman animals.  In
fact, the kind of representational system that is required, argues Millikan in
the final chapter, is one which infuses a subject-predicate structure into
representations, where the predicate is sensitive to a negation
transformation.  And so, Millikan conjectures, it is the development of human
language that has allowed for the development of theoretical thought.   

This, in brief, is a wide-ranging
book, covering a diverse range of issues at the intersection of the philosophy
of mind and the philosophy of language, but always bringing the scientific
literature in such fields as psychology and biology to bear on the issues at
hand.  Indeed, one of the great virtues of the book is that it regularly defers
to and illuminates recent experimental studies, as well as providing a rich
array of examples, images, and analogies drawn, in the main, from the world of
biology.  Be warned, however, that Millikan’s (often compressed) writing style
and thinking will be challenging to those not familiar with her work or the
current debates in teleosemantics.  Challenging, I say, but not impenetrable —
and the rewards in insight repay the outlay in effort.

© 2005 Nick Trakakis

 

Nick Trakakis,
Department of Philosophy, Monash
University, Australia.

Categories: Philosophical