<< . .

( 29)

. . >>

Or a disposition, right now, if I try, to proceed to get into a position to
do A if I try . . . and so forth. Clearly the edges are not sharp here ei-
ther. But the distinction between knowing how and actually being able
may cast some light on how we should understand what knowing what
one is judging about entails. Perhaps we should draw a similarly vague dis-
tinction between having a concept of an object and being able to re-
identify it in practice. Then we would note that there are degrees of
knowing how to reidentify. I may command myriad ways or only a few.
Also there are degrees of being able to identify in practice. I may get
thousands of opportunities I can utilize or none. But however we divide
things up, it is clear that knowing what I am thinking of is always a
matter of degree. No one is in principle infallible at reidentifying any-
thing, even should they happen always to succeed in practice. Nor, un-
like the case of many more ordinary abilities, need they always know, ei-
ther immediately or at all, whether they have succeeded.
Not only are substance concepts always imperfect abilities, the means
used for many of them may be in our command only temporarily.
Sometimes this is because our memories are short. For example, al-
though I can usually remember C sharp for five minutes so as to re-
identify it when I hear it again, I always forget it overnight “ I don™t
have “absolute pitch.” Luckily my concept of C sharp has other con-
ceptual means. I know how to reidentify it by going to the piano, or the
flute, or by asking my friend Brian who has completely infallible pitch.
Because I know C sharp™s name, I also know how to mark sameness for
the thought of C sharp in new grounded judgments as I hear or read
about C sharp. I may read what its frequency is, or about the difficulty
of playing this or that instrument in its key.

New faces can be, for me, a bit like C sharp. I can identify them over
the next hour, but not always over the next week. If I have forgotten
both the new face and the name, I may have to act very fast to pick up
the scent. Descriptions are sometimes useful, but their usefulness may be
short lived. I may need to call on the memories of others quite soon,
before they forget who fits the description, before they forget, say, who
all was present on the occasion. Sometimes there may be no way to
pick up the scent, and I will never “know who that was” that I met,
never “know who it is I am remembering.” But as a limiting case, of
course, I do know who I™m remembering. I would know how to pick
up the scent were I to come across certain kinds of information. I have
relevant abilities. It™s just that I haven™t had, and perhaps won™t get, a
chance to apply them.


To lack the ability always to identify correctly “ to be disposed some-
times to error “ is part of the human condition. Merely to be disposed
to error, however, is not yet to have made an error. And it is only in so
far as one actually makes errors in identification that one™s thoughts be-
come equivocal.
One result of mistaken sameness markings may be invalid inference.
For example, suppose that I see that something in the tree is a squirrel
and then I see that “it” takes off and flies. I may conclude that some
squirrels can fly. But the inference will be invalid if I unknowingly had
lost track of the squirrel and some other “it” did the flying. Since invalid
inferences can lead from true premises to false conclusions, misidentifi-
cations can of course give rise to false beliefs. More interesting however,
is when misidentifications give rise to confused or equivocal thoughts.
A person who has very basic misinformation about a thing may be
said to be “confused” about it, but that sort of confusion is not equivo-
cation. “Do you see that woman?” Jane says, pointing to Ann, and then
she tells you a whole kettle of lies about Ann, all of which you believe.
Perhaps you never learn anything else much about Ann, but once in a
while you do see her again on the street, and then you review in your
mind all those dreadful lies. You are woefully confused about Ann, but
your thought of her is not equivocal. For it is definitely about Ann that
you are confused, and not about something ambiguous. Perhaps you
even believe a lot of wrong individuating descriptions of Ann, but that

is not equivocation either. As long as you don™t actually use any of these
descriptions in such a way as to result in actual misidentifications of
Ann, you will not yet have an equivocal thought.
Suppose, on the other hand, that you mistake Carol whom you see
on the street for the city mayor. Having heard that the mayor was in
Washington just this morning, you conclude that she must have taken a
plane home.That WHO must have taken a plane home? “ Carol, or the
mayor? That is what I mean by an equivocal thought. It isn™t a thought
of the one woman any more or less than it is of the other, but hovers
On a descriptionist theory (Section 3.5), an equivocal thought
might, I suppose, be one that was governed equally by two or more def-
inite descriptions that were not coreferential. Or on a conceptionist
theory (Section 3.5), an equivocal thought or concept might cor-
respond to a disposition to recognize incompatible things as part of the
same extension. I™m not sure that either of these suggestions is entirely
coherent, but the point I would make is that where the conception is
taken to determine the extension of a concept, equivocation, should it
exist, would be found in mere dispositions of the concept user. My claim,
on the contrary, is that equivocation is found only where actual infor-
mation, derived from distinct sources, is marked as being about the
same. One always has dispositions to misidentify things, given sufficiently
awkward conditions.
Suppose that I think John is the president of the local chapter of the
AAUP, but he™s not. Bill is instead. Do I have an equivocal concept, or
just a false belief about John? That depends on whether or not I have
gathered information about the president of the local chapter of the
AAUP and applied it to John. Nor need this mean that I have applied
information derived from Bill to John. More likely, I have gathered gen-
eral information about the presidency of the AAUP. I know what the
president™s duties, privileges, and some of his probable locations are
(e.g., at the meeting in Mannly Hall on Friday).The president (timeless)
of the local chapter of the AAUP is a rough sort of substance. It is
something I can learn about and the information will remain valid over
time. If I have falsely identified John with the current presidency, I have
probably mixed information about two things together. I have probably
made wrong mediate inferences, inferring, for example, that John will
be in Mannly Hall on Friday. On the other hand, if I merely take John
to be the tallest man in the room and he™s not, granted I have no

information about the tallest man in the room, as such, there is not as
yet, and likely there won™t be, any hint of equivocation in my idea of
John. False information is not, as such, equivocation.
To have two things or more confused in one™s mind is surely a com-
mon condition. For example, much of the history of science might be
told in these terms. What is astonishing is not that it happens, but that
in dealing with common objects, properties, and kinds it doesn™t happen
more often. What is astonishing is how good we usually are at keeping
track of those ordinary things in our world that (unlike which glass is
which in the cupboard “ Section 5.5) matter to us “ at not mixing
them all up together. What is astonishing is how good our concepts
tend to be, despite the fact that they must operate, as must our other
abilities, on principles resting not just on the character of our minds, but
on the structure of the world outside.

How Extensions of New Substance
Concepts are Fixed:
How Substance Concepts Acquire


In the first section of this chapter, I will use the results of the last chap-
ter to explain more exactly how the extension of a substance concept is
fixed. I will be concerned, especially, with how the extensions of new
substance concepts, acquired directly on first meeting with their refer-
ents, are fixed.The rest of the chapter will be devoted to fitting the the-
ory of substance concepts defended here into the more general theory
of intentional representation developed in Millikan (1984, 1993a).
Evans concluded, about the man with the memory of the ball he was
unable to identify, that the man did not have the capacity to think of
that ball at all, but only to think of a ball (Section 13.4). But, I have
argued, the question whether he could think of that individual ball
doesn™t turn on whether he was actually able to reidentify it. It doesn™t
turn on whether his situation was right for reidentifying it. It depends,
rather, on whether his thought was produced by his cognitive systems
in such a way as to have, as its first assigned function, that it be coiden-
tified, specifically, with thoughts of that particular ball. And to that ques-
tion the answer would seem to be yes.
As Evans describes the case, when the man was looking at the ball he
understood it to be an individual ball, indeed, one that happened to be
steel and shiny. It is not in question, then, whether he had the general
ability to think of individual balls. Presumably at that time he applied a
relevant sort of primary-substance (individual-object) template to the
ball and was conceptually ready to track it for purposes of collecting

certain sorts of information about it. He had various skills in place for
tracking individual physical objects, and some understanding of what
might be learned about this particular physical object. The cognitive rep-
resentation produced by seeing the ball “ the thought it produced “ was
thus designed to be taken up by an interpreting system having, among
other functions, the function of conceptually tracking and reidentifying
this particular ball. Had the system tracked and reidentified this partic-
ular ball, it would have been working entirely in accordance with prin-
ciples it was already designed (selected for, trained, or tuned) to instanti-
ate.1 Thus the man™s idea of the ball was a fully intentional cognitive
representation of it. The fact that he also saw, but then forgot, a similar
ball at another time does not change that matter. Nor is the matter
changed by the fact that in this case he couldn™t perceptually discriminate
that ball from the other (Section 14.1). Nor is it changed by the fact
that no one happened to show up later to explain to him which ball it
was (Section 13.4). Similarly, if I spot a new kind of lizard in the grass,
one that is completely unfamiliar to me, and propose to find out what
I can about its species, granted that the general abilities I have already
acquired with regard to tracking lizardlike species are adequate, my con-
cept of this species may already be completely determinate in extension.
Also, any new substance concept that I acquire just by remembering
the name of that substance and understanding in what general category
it falls (Chapter 6), may immediately have a perfectly determinate ex-
tension. I have an ability to recognize this same name when I encounter
it again. Suppose then that the name has a determinate referent in the
public language. This requires that others in the language community,
past or present, have or have had an ability to recognize this referent,
and have sometimes or do sometimes broadcast information about it by
using this name. I then have the (fallible) ability to identify this as be-
ing information about the same. My concept has the same referent as
the public term, and by holding the steadying hand of language, I have
the ability to learn how to extend my conception of this referent to in-
clude nonlinguistic means of identifying it. Thus, Burge (1979, 1982) is
right that what I mean by a term may depend directly on what others
in my language community mean by it, indeed, that what I am thinking
of can depend on what others mean.

1 In the terminology of Millikan (1984), it would have been performing an “adapted proper
function.” For details on the derivation of adapted proper functions, see also Millikan (in
press b).

Contrast these cases with one where an apparent cognitive represen-
tation has no referent at all. In one place Evans remarks that “[a]n in-
formational state may be of nothing: this will be the case if there was no
object which served as input to the information system when the in-
formation was produced” (1982, p. 128).2 Similarly, it certainly is
possible that the cognitive systems should sometimes be triggered to
produce apparently intentional representations, apparent concepts, by
sources entirely foreign to any proper use of them. It is possible to seem
to perceive and to seem cognitively to represent objects that are not ob-
jects at all.That is, there is no way that the cognitive systems could pro-
ceed to reidentify an intentional object in this case that would consti-
tute these systems proceeding in accordance with principles they were
designed to instantiate.
As it first begins to develop, then, a substance concept may have a
completely determinate extension, or it may be determinate that it has
no extension at all. Between these two possibilities are others in which
the germ of a substance concept might develop normally from here in
any of several ways. These are the cases in which a reference has yet to
be “focused” (Perner 1998) thus hovering between possible extensions
(Section 6.3).
Also, as has been a central theme of this book, if the content of any
substance concept proceeds actually to be misidentified, and the mistake
is not corrected, so that information about two things is bound together
as though they were one, although the representation is not empty, it is
at least to a degree equivocal. Correctly identifying a substance in
thought has the logical form of contributing to the solution of a coor-
dination problem or, say, of directing one light ray toward a focus with
others. Identifying correctly is like focusing the eyes. When both eyes
are open but not focused on the same thing, neither eye sees anything
clearly, but when both are focused on the same thing, then both see,
both see the same, and both see clearly.


According to Russell™s principle, thinking of something is always
grounded in some sort of “acquaintance” in the Russellian sense imply-
ing a kind of knowledge of what that thing is. There is no such thing
as an object of thought the identity of which is unknown to the

2 On Evans™ use of the term “information,” see Appendix A.

thinker. It cannot be that I first think of something and subsequently
grasp (or fail to grasp) what it is I am thinking of.This claim is correct,
I have argued, for conceptual thought, so long as we understand know-
ing what one is thinking of as a fallible ability, compatible with the pos-
sibility that one can make mistakes about the object of one™s thought
(Section 13.5 and Section 13.6). What I am thinking of conceptually is
not determined prior to my ability to reidentify, though it is determined
prior to my actual acts of identifying, which may not properly express
that ability.
Russell™s principle is mistaken, however, if “thought” is understood to
cover all forms of mental representation. It is mistaken if taken to be re-
quired of intentionality more generally. Conceptual representation is
marked off by requiring a matching capacity to reidentify its content,
but not all intentionality involves concepts. Perception, for example,
may have content that the perceiver has no need or ability to reidentify.
A full treatment of these more general claims about intentionality is of-
fered in Millikan (1984, 1993a). Here, I will abbreviate only some as-
pects of that work, enough to show how the claims about substance
concepts fit into the overall picture of intentional representation
sketched in those earlier essays.
Natural information, as this notion has usually been understood, is
contained in an output signal pattern that covaries with the pattern of
input from some source of information, through some physical
medium or “channel,” according to physical law. I call this sort of in-
formation “informationL” (“L” for “law”).3 This is not the sense of
“natural information” used in most of this book, that sense being ex-
plained in Appendix B, but it is likely to be familiar to the reader, so I
will use it as a starting point in explaining the general notion of inten-
tional representation.
InformationL is ubiquitous, both in animate and inanimate nature. It
has, merely as such, nothing to do with intentionality, nor is it, just as
such, of any use to an organism. There are certain conditions, however,
under which an output signal containing informationL about a config-
uration at its source is also an intentional representation of that config-
uration. To make this the case, the signal carrying informationL has to
carry this information “intentionally” in the following sense. It must
carry it in accordance with the natural purpose or function of some

3 This is roughly the way in which Dretske, Fodor, and Gibson use the term “information.”
For discussion, see Appendix B

transmitting or relaying mechanism, a mechanism that has been selected
or trained for exactly that job.
If a mechanism has been selected or trained for the job of producing
or relaying a certain kind of informationL, that will be because some
cooperating or coordinate mechanism, perhaps another phase of the
same mechanisms, has a use for this informationL. But informationL, ar-
riving from the environment, relayed through an animal™s sensory or-
gans to its brain, arriving in a certain code or vehicle, cannot be of use
to the animal unless the animal is designed or has learned to be guided
by this sort of vehicle in ways appropriate to the configuration at the
information™s source. That is, informationL is of no use to an organism
unless it can be “interpreted” through the arousal of inner or outer ac-
tivity of the organism that is appropriate to the state of affairs the in-
formation concerns. In sum, informationL that is embodied in an inten-
tional representation is produced or channeled in accordance with the
proper functioning of some designed mechanism, where a further
proper function of that mechanism is to cooperate with a correspond-
ing “interpreting” mechanism to guide that interpreter in accomplishing
some (ultimately practical) function or functions beyond, under the cir-
cumstances represented.
Now let me generalize this idea. Only some mechanisms designed
to produce intentional representations are designed to do so by pro-
ducing or transmitting natural informationL. This may well be how di-
rect perception of the spatial layout of the immediate environment is
normally accomplished for purposes of direct action guidance. Gibson
was probably right about that. But this is not, for the most part, how
cognition is accomplished. In its more general form, intentional repre-
sentation requires only that there be a mechanism designed to produce
items bearing a certain correspondence to the distal environment, corre-
spondence in accordance with some definite rules (“semantic rules”),
which items (“intentional representations”) are to be used to guide an-
other system (the “interpreter”) in the performance of certain of its
Now an enormously important feature of designed mechanisms is
that they do not always accomplish the functions for which they were
designed. Many mechanisms designed either by natural selection or by
learning manage to perform their assigned tasks only occasionally. Only
occasionally does the slap of a beaver tail on the water actually perform
its proper function of saving the beaver™s relatives from a danger. In part,
this is because its timing only occasionally corresponds to the timing of

any real danger. Mechanisms designed to produce intentional represen-
tations that correspond by a rule to the distal environment can have
these functions yet frequently fail. Compatibly, mechanisms designed to
produce intentional representations may be designed to do so, not by
the use of channels of natural informationL, where output form corre-
sponds to input form in accordance with natural law, but by the use of
quite unreliable statistical methods (see Appendix B). The output of the
system may correspond to the distal environment just often enough to
be more useful than if the organism had no such system in place at all.
What makes intentional representations “represent” their intentional
objects is thus quite different from what makes photographs “represent”
their subjects. Intentional representations represent what they would
need to correspond to for their interpreting mechanisms to use them
productively in accordance with design. If they fail to correspond in this
way but the interpreting mechanisms are guided by them in the normal
way, the result will be unproductive. It is unproductive for beavers to
waste energy and time diving under when there is no danger present. It
is unproductive for a worker bee to fly off in the direction a bee dance
says unless there is nectar in that direction. But that the intentional rep-
resentation needs to correspond to the world by a certain rule in order
to be productive does not imply that there exists any method by which
it can be produced that would guarantee this correspondence. It does
not imply, for example, that there must be some method by which its
object could always be discriminated from all other objects in accor-
dance with natural law. To represent something it is not necessary that
one be able to tell it from all other things.4
The “correspondence” that a system producing intentional represen-
tations is designed to establish between these representations and their
representeds can be thought of as an abstract isomorphism, in this way.
Transformations (in the abstract mathematical sense) of the representa-
tions correspond to transformations of what is represented, such that
different representations map different representeds in a systematic or
“productive” way. Intentional representations have, as such, not ordinary
extensions but truth conditions.5 They are not analogous to names or
open sentences. Rather, intentional representations always make claims.6

4 For discussion, see Appendix B.
5 More accurately, they have satisfaction conditions. I have omitted discussion here of repre-
sentations that are imperative rather than indicative. See Millikan (1984, Chapter 6).
6 Indicative ones do. See note 5 above.

If they did not make claims, they could not be such as to guide activity
appropriately given the existence of their extensions. About danger, for
example, there is nothing to be done, nor is there anything to be done
about here or about now. But about there is danger here now there may
well be something to be done. Similarly, about nectar fifty yards southeast
of here there is nothing to be done, unless one reads this as an assertion

<< . .

( 29)

. . >>

Copyright Design by: Sunlight webdesign