<< . .

( 29)

. . >>

minate modes of presentation. Characteristically, each dot will be coor-
dinated with multiple ways of identifying, multiple ways that the
thinker would be able to recognize incoming information about that
referent. But the various pieces of information attached to a given dot
are not associated with any one of these ways more than another. True,
each bit of information may have found its way to the dot by just one

path of recognition. But the Strawson system keeps no record of which
information entered by which path. Besides, on this model modes of
presentation are not supposed to be just ways a thing has historically been
recognized by the thinker, but ways s/he knows to recognize it. Many
modes of presentation grasped for it may never have been used in the
forming of particular beliefs. Certainly these will not be modes under
which anything is believed about it. Note also that if the terms of a
thinker™s belief are not each characterized by a determinate single mode
of presentation, but by many modes at once, and if these various modes
should happen not, in fact, all to determine the same object, then, as be-
fore, it seems that the thinker™s thought might in all innocence be
A more serious problem with taking Strawson modes of presentation
to be merely ways of identifying things so as to channel information
about the same arriving in different packages to a single focus or “dot,
concerns the difficulty of individuating ways of identifying a thing, such
as to form distinct modes or senses. Such a view is implicit, I believe, in
Gareth Evans™ discussion of “dynamic Fregean thoughts,” so I will use his
analysis to exemplify its weaknesses. Evans™ view that there are dynamic
Fregean thoughts, if pushed to its limit, yields exactly the same paradox-
ical results as does the naive Strawson model of modes of presentation.


Evans (Evans 1981, 1982, pp. 174“6, pp. 194 ff) proposed that when
you are tracking an object perceptually, say, keeping it in view as it
moves and you move, if you continue to believe over this period of
time that the perceived object has a certain property, this should not be
considered to be a sequence of similar beliefs that you have, but a single
belief that persists over time.3 You continue to think of the object un-
der the same mode of presentation, as long, that is, as you haven™t un-
knowingly lost track of it. Evans calls this sort of thought a “dynamic
Fregean thought,” and he says that in such cases the relevant “way of
thinking of an object” is a “way of keeping track of an object” (p. 196).
Now, if you do not merely persist in the same belief about the object
over the tracking period, but continue to collect new information
about it from perception, noting, say, its way of moving, what it looks
like from the back, what it sounds like, how large it is, and so forth, pre-

3 See also John Campbell (1987/88).

sumably this will not change the fact that you continue to think of it
under the same mode of presentation, as long as you don™t lose track.
Note the isomorphism with the Strawsonian analysis of the same track-
ing event.You continue to keep many old predicates attached to the dot
while you funnel in various new bits of information to attach to the
very same dot.
And what if you should unknowingly lose track of the object? You
thought it was one little minnow “ you named him “Primus” “ that
nibbled first your toe then your ankle, but there were actually two. In
that case, Evans claims, “we have not a case of misidentification but a
case where the subject has no thought at all” (Evans 1982, p. 176). For
in the absence of “an ability to keep track” of the object, “it is not pos-
sible for a subject to have a thought about an object in this kind of sit-
uation at all” (p. 195).
But an ability, I have argued, is not, in general, something one either
has or has not. Most abilities come in degrees. One of my surest abilities
is my ability to walk, but there are still times when I trip. Hence there
seems another response possible for Evans.You do have an ability to keep
track of things like minnows, only this time you tripped (Chapter 4).This
particular dynamic mode of presentation of yours is indeed part of a
thought, but the thought happens to be equivocal. It hovers between the
two minnows, presenting both as if one. Taking another example, imag-
ine a person losing track and apparently, but wrongly, perceiving the same
squirrel eating first six and then seven more Brazil nuts. The result is an
indelible memory of the squirrel who ate thirteen whole Brazil nuts at a
sitting. Surely this is not a case of no thought at all, but a case where two
contents have been blended, a case where thought is equivocal.
True, Evans is wedded to “Russell™s principle” “ “that a subject can-
not make a judgment about something unless he knows which object
his judgment is about” “ and he interprets this to mean that the subject
“has a capacity to distinguish the object of his thought from all other
things” (Evans 1982, p. 89). But Evans gives no argument anywhere for
the soundness of this principle used this particular way. I am happy to
agree that if a dynamic mode of presentation were sufficiently equivo-
cal, not just mixing little minnow Primus with Secundus, but also
rolling in, say, Sextus, Septimus, and Octavius, indeed, a large random
sample of other minnows in the school, it would be odd to consider it
as determining a thought of any minnow at all. It should probably be
considered “a case where the subject has no thought [anyway, of indi-
vidual minnows] at all.” In earlier chapters, I, too, have argued that some

ability to track them individually is necessary to having thoughts of
individuals. This is parallel to the result we got on the naive Strawson-
inspired model: If enough mistakes in identifying were made, the result-
ing thoughts would be effectively empty. But I have never heard an ar-
gument anywhere that no equivocation at all is ever possible in thought.
Whatever one decides about that, however, surely the case of error-
infected naive Strawson-inspired modes of presentation and the case of
error-infected Evans-inspired dynamic modes of presentation must be
decided in the same way, for the parallel is exact. The parallel can be
shown, indeed, to be a structural identity.
Consider the dynamic mode of presentation involved as you percep-
tually track a person, Kate, to whom you have just been introduced at
a party. For a brief moment “ not much longer, suppose, than a saccade
“ you divert your eyes to the face of a friend, but immediately pick up
Kate™s face again.Then a large fat man, excusing himself, passes between
you and Kate, but again you immediately pick up the track. Looking at
Kate and hearing her voice, you perceive these as having the same
source, as locating the same person. Now Kate passes for a moment into
another room, but you continue to hear her voice “ though of course
there are spaces between the words “ and she soon emerges again. By
now she is beginning both to look and to sound quite familiar, so that
after stepping outside for a moment, you immediately find her again.
The time interval was longer this time than between her words, but
short enough for her voice still to be “in your ears.” Compare this, for
example, to the way a bloodhound tracks a person by smell, at moments
losing but then picking the scent up again, or the way one tracks an ob-
ject visually, seeing it as the same object as it emerges after passing be-
hind a tree.You should not think of the bloodhound as merely repeating
a particular way of recognizing the person over and over as the scent is
lost and regained. Nor do you merely repeat a way of recognizing the
visually tracked object.You keep track of it by tracing and anticipating
its natural projectory in space.4
Now suppose that Kate looks and sounds familiar also an hour later
and then a day later when you meet her again, first in the lobby, then on
the street. Probably you would not have recognized her, however, had
you met her in Singapore “ in some radically disjoint context. Similarly,
Evans tells us in his chapter on recognition that though by using your

4 Compare Evans: “ . . . demonstrative Ideas will shade off, without a sharp boundary, into
Ideas associated with capacities to recognize objects” (1982, p. 176).

recognitional ability alone you might not be able to tell a certain sheep
you are thinking of from every other sheep in the world, still, because
you can keep track of the neighborhood in which the sheep is likely to
be and you can also keep track of where you yourself are, you can main-
tain an ability to reidentify the sheep (Evans 1982, Section 8.3). Further
now, suppose that Kate™s name has become familiar, and as more time
goes by you often pick up information about her from friends. Again,
you usually know which “Kate” they are talking about from the context,
from anticipating her possible projectories, and the possible projectories
of various kinds of information emanating from her.
When did you stop tracking Kate? When did you stop following her
spoor, the trail she left of ambient energy structures bombarding your
sensory surfaces? When did the original mode in which she was pre-
sented to you come to an end?
A dynamic mode of presentation that never came to an end would be,
functionally, exactly the same as a “naive” Strawson-style mode of presen-
tation. Each of the peculiar, distinctly unFregean traits that I have described
for the latter modes would characterize the former as well.Whether Evans™
dynamic modes really differ from Strawsonian modes in function depends,
then, on whether a clear principle of individuation “ of sameness and dif-
ference “ could be drawn for ways of tracking or abilities to track. When
did you leave off one way of tracking and start using another ability to
track, or some different kind of ability to “know which object you are
thinking about,” as you collected information over time about Kate?


The concept of a substance consists, in part, of an ability to reidentify
its object or to track it conceptually. One™s conception of a substance, I
have said, concerns how one is able to do this. To describe someone™s
conception of a substance is to tell how they would go about reidenti-
fying or tracking it. If modes of presentation of substances were ways of
tracking substances, then when the same substance is presented to a per-
son through a variety of different modes of presentation and they un-
derstand these as presenting the same substance, their conception of the
substance would have to be divisible into subconceptions, or discrete
means of tracking. Each of these means would then be a distinct sense,
capable, by itself, of uniquely determining that substance as referent.
First is the problem of division.The general method used for reiden-
tification of every substance is the same. One relies on certain expected,

that is, projected, continuities over times or occasions of encounter, for
example, continuity in spatial projectory, in color, in shape, in odor, in
general arrangement of parts, in manner of motion. Or one may rely on
continuity in identity of parts or other associated features. Perhaps it is
the shape of your face, or just your eyes that I recognize immediately as
indicative of you, or your voice, or your signature, or your walk, or your
humor, or your name, or your pocket watch with its distinctive pop-up
cover (see Preface), each given the right context. The means employed
to recognize a substance thus embed prior or more general means, not
means that are unique, usually, to this substance. I follow the object™s
projectory, granted I can do this for objects generally, under a certain
variety of conditions. My ability to track with my feet as well as my
eyes and head may be involved here, hence my ability to walk over
rough ground or to avoid slipping on ice. I can recognize blueberries
partly by their color because I can reidentify colors generally under
these and those sorts of conditions. I recognize squirrels by their shapes
and characteristic motions, granted I can reidentify shapes and motions
generally, big shapes, small shapes, motions in the open, motions par-
tially obscured, shapes and motions close at hand, shapes and motions in
the distance. I recognize many people “by their voices,” where this
means catching the same regional accent again, along with the same vo-
cal quality. I recognize individual old fashioned pocket watches by their
shapes and markings, such as initials engraved on them. I identify indi-
viduals and kinds by their names, or by descriptions of them, where this
requires recognizing the same word again, recognizing the same de-
scription put in different words, or put in a different regional accent, or
a different language. Recognition by one such complex means or an-
other of enough of these sorts of continuities, all reinforcing one an-
other, often suffice for the practical act of reidentifying a substance. But
that the various methods, actually or possibly employed, making up my
ability to reidentify some particular substance, might be divided nonar-
bitrarily into discrete, countable “ways of identifying” simply is not co-
herent. Just as many integrated skills go into even an act of playing
“Twinkle Twinkle Little Star” on the violin or, catching a ball, many in-
tegrated general skills, added to specialized bits of knowledge, go into
any particular act of recognition or particular act of keeping track of a
That a person™s various means of recognizing a substance cannot be
divided and counted does not imply, of course, that they cannot be de-
scribed, or individually designated. Means are like places. There is no

answer to how many places there are in London, but I can describe
where I am, with greater or lesser exactness, and I even can designate
this place entirely exactly, as, just, the place I am now in. Similarly, I can
describe my means of identifying Kate on some occasion more or less
exactly, as by sight, say, or as by noticing the shape of her nose, or I can
refer to a way of having identified her as the way I used last Thursday
in the park.
The second problem with identifying ways of recognizing with
modes of presentation concerns the requirement that each mode of
presentation be capable, by itself, of uniquely determining its referent:
one mode yields one referent.What would it be for a way of identifying
a substance to determine that substance uniquely? This would require
that it be an infallible method of determining that particular substance,
never catching another substance instead. But a second traditional re-
quirement on modes of presentation is that the rational thinker always
grasp the sameness when employing the same mode of presentation
again, for the rational thinker must never make contradictory judgments
about the same grasped through the same mode of presentation. We
would need then to individuate ways of recognizing such that (1) each
cannot fail to net always the same object and (2) this fact is guaranteed
a priori. But ways of recognizing are always in principle fallible, or at
the very least cannot be known a priori to be infallible, because they
depend on certain external conditions being in place. This is because
“ways of recognizing,” in this context, are not ways of holding an ob-
ject up before the mind, but ways of knowing when one is receiving in-
formation about an object. And there is no such thing as an ability to
interact in a given way with a distal object that isn™t in principle falli-
ble. That is the infamous Achilles heel of verificationism.
Perceptual evidence never guarantees its sources. Perceptual tracking
is always fallible. There cannot be an a priori guarantee that one has
kept track, or even that there is anything actually there to keep track of.
The same is true for conceptual tracking.You are surely able to identify
each member of your immediate family in myriad ways, some ways of
which “ a long look full into your spouse™s face in full daylight, for ex-
ample “ may (barring removal of your brain to a vat) actually be infal-
lible. But if that is so, it is because the world, not anything in your
mind, is constructed so as to make it so. It is because there is not in fact
any other person in the world who looks just like that in the face (and
no one actually able, and desirous of, putting your brain in a vat) “ a
convenient fact but not one guaranteed a priori.

Similarly, recognition using identifying descriptions is never infallible.
First, that the description is unique is always contingent. There might
always be, within limits of discernability, two tallest or two oldest, for
example, so that neither is really tallest or really oldest. And one can al-
ways make a mistake about which one is tallest or oldest because one
perceives wrongly, or because one infers wrongly, or because one is in-
formed wrongly by others. True, it has seemed to many that an identi-
fying definite description is the surest sort of tool one could use to
make fixed what one was thinking about. But the job that must be
done by a method of recognition for incoming information is not to fix
a thing before the mind. It is to effect actual reidentifications, to direct
actual incoming bits of information about the same to a focus, so that
they will interact with one another in inference. Used for this purpose,
most definite descriptions are of severely limited value.
I conclude that given Strawson™s model of sameness marking, there is
no way to salvage the notion that there are such things as modes of pre-
sentation that will do all, indeed perhaps any, of what Frege wanted
them to.

Rejecting Identity Judgments and
Fregean Modes1


I would like to understand what the basic principles are that distinguish
the vision of thought we have generated using the Strawson image of
sameness marking from Frege™s original vision of thoughts as exempli-
fying modes of presentation. The first conclusion I will reach is that,
surprisingly, the way the Strawson markers mark identity plays no role
in determining this difference. The interaction of Strawson™s image of
sameness marking with Frege™s vision of modes of presentation yields
strikingly unFregean results. Yet these results are not merely an artifact
of the Strawson model.They follow given any model of sameness mark-
ing. Strawson™s way of marking identity highlights a general feature im-
plicitly present in all other models as well. It will take a while to argue
for this conclusion. I will place particular emphasis on the equals
marker, and on the image of thoughts as sentencelike, in which the
equals-marker model is embedded. For initially it is quite unintuitive
that this particular model is isomorphic to the Strawson model. Such is
the hold that the mental sentence image of thought has on all of us,
with its careful but, as I will argue, illusory distinction between dupli-
cates markers and equals markers, that is, between graspings of necessary
identity and contingent judgments of identity.
Further search is thus needed to understand the division between the
vision of thought we have generated and Frege™s original vision of

1 Parts of this chapter are revised from “Images of Identity” (Millikan 1997b), with the kind
permission of Oxford University Press, and from “On Mentalese Orthography” (Millikan
1993b), with the kind permission of Blackwell Publishers.

thoughts as exemplifying modes of presentation. What exactly is the
source of the difficulties we have encountered in trying to interpret
what a mode of presentation might actually be in a thinking mind or
brain? I will argue that the classical notion of modes of presentation
rests on two assumptions, both of which are mistaken. One classical
source is an implicit denial that the way the mind uses the thoughts or
ideas that it harbors has any bearing on their intentional contents. In
particular, as suggested in Section 8.7, the Fregean model invokes the
passive picture theory of the act of understanding sameness of content.
What the mind does with the pictures is not involved in determining
their contents, or in determining whether they are thought of as same
or different in content. The second source is an internalist view of
thought content, that is, a denial that the natural informational content
carried by a thought has any bearing on its intentional content.


Begin by considering duplicates markers. How will a system consisting,
say, of mental sentences, and that uses only duplicates markers, come to
realize that Cicero is Tully? It must put all the Cicero and Tully infor-
mation into sentences using the same mental name, either “Cicero” or
“Tully,” choose which. Just as one of the dots has to go on the Straw-
son model, one of the mental words has to go on the duplicates model.
So if it should turn out later that Cicero is not in fact Tully, whichever
mental name got chosen will be equivocal, nor will the news that Ci-
cero is not in fact Tully represent, for the system, any definite instruc-
tions for separating the information again into two piles. Duplicates
markers do not differ from Strawson markers in function.
A moment™s reflection shows that the synchrony model, the Christ-
mas lights model, and the anaphor model of sameness marking also have
this result. Each merely binds all the things known about an object into
one bundle. Each performs acts of identifying merely by merging bun-
dles, so that no particular information about any subject has a particu-
lar or different mode of presentation from any other. The only method
of marking we considered that does not obviously have this effect is the
equals model of marking. We find this last model illustrated, implicitly,
in Frege himself.
Frege might be interpreted as having supposed that the mind uses, in
part, a duplicates system of sameness marking. For although senses were

not supposed to be psychological entities, graspings of them surely are
dated, psychological occurrences, and Frege seems to have held that it is
awarenesses of duplicate graspings-of-sense that keep us from contradic-
tion and govern the performance of rational mediate inferences (Chap-
ter 9). If so, this constitutes, I have argued, a substantial psychological
claim. A perverse deity might have made our minds otherwise. Imagine,
for example, the same sense coming into mental view twice simultane-
ously as subject term of contradictory judgments, but the demon has
determined that only synchronously vibrating viewings of the same
sense will move the mind to visage sameness of reference.
It is clear, however, that Frege did not view the result of an identity
judgment to be the elimination from use of one of the two kinds of
senses grasped, nor did he suppose any senses were equivocal. Rather, he
took there to be two kinds of identity judgments, the “informative”
ones such as “Cicero is Tully,” and those such as “Tully is Tully,” which
are not. We might suspect, then, that it is the introduction of this new
way of marking identity, used for identities not known a priori, that al-
lows the Fregean thinker to identify referents without merging his
thoughts of them together. This second identity marker, we suppose,
functions like a mental equals sign. It marks two thoughts as being
thoughts of the same, not by merging or destroying either, but simply
by flagging them for use together in mediate inference.
The suggestion that there is something like an equals sign in thought
which marks identity comes, more generally, from modeling thoughts

<< . .

( 29)

. . >>

Copyright Design by: Sunlight webdesign