<< . .

( 29)

. . >>

extension. But then the prior question can be raised, what determines
which features or properties are the ones you are representing and trying
to recognize? This cannot be determined by your representing prior
properties that the represented properties must themselves have, without
regress. Nor can it be determined merely by your dispositions to recog-
nize these properties, since no one is infallible at recognizing properties
any more than substances.
A standard reply is that we do recognize certain properties infallibly
when we are in “normal conditions,” and that this is what defines what
properties our property concepts represent. For example, red is whatever
appears red to me under normal conditions for seeing colors. But how
then do we define “normal conditions,” such that they are, for example,
appropriately different for seeing the shapes of big things like mountains
and the shapes of small things like fleas, appropriately different for hear-
ing loud sounds and soft ones, and different for hearing sounds, for see-
ing colors and for identifying tastes, and so forth (consider how tea
tasters must prepare themselves)? We must take care that “normal con-

ditions” do not turn out to be, just, the conditions under which one
perceives each of these various properties correctly, for that would be
marching in place. On the other hand, if there is any noncircular way of
defining “normal conditions” for the perception of various properties,
we should be able to use exactly the same technique to define “normal
conditions” for keeping track of various kinds of substances. The two
problems are exactly parallel. How then should we understand these
“normal conditions”?
Biologists are usually concerned, first, to understand normal func-
tion. They may be interested in diseases or other abnormal functions,
too, but these are defined relative to normal function. I have proposed
that normal function, in this context, is best defined relative to a history
of natural selection,5 but you may supply your own favorite theory of
normal function, should you have one, and it will serve the argument as
well. My claim here is only that whatever normal function is, philoso-
phers and cognitive psychologists, too, are, or should be, interested in it
as well as biologists. I add that for the most part biological items require
to be in rather definite conditions in order to perform normally. My
own preference is then to define “normal conditions” relative to selec-
tionist history also “ as conditions under which that function was per-
formed historically such as to be selected for Millikan (1984,1993a).
But if you can supply a better definition, there is no objection. The
point is that if we can give a definition of normal biological function
and normal conditions for performance of this or that function, we can
apply it also to performance of psychological functions, such as the de-
velopment of substance concepts and their application.We can say what
normal conditions for keeping track of substances are.
Grant then that there is a normal way or normal ways for develop-
ment of substance concepts to occur, perhaps different for different sub-
stance domains. Or what is the same thing, grant that normal develop-
mental cognitive psychology is a viable field. There will be a normal
way or ways that the child or adult first recognizes the manifestations of
a new substance impinging on their perceptual organs and a normal
way or ways that they attempt to keep track of that substance, or learn
to keep track of it better, and so forth. And there will be normal con-
ditions for success in keeping track, in building conceptions adequate to
the substance. When everything goes exactly normally, then, there will

5 1984, Chapters 1 and 2; 1993a, Chapter 1 and 2.

be no question what the concept is of, even if there is a disposition to
apply it incorrectly under conditions that are not normal for expression
of these abilities.
But problems can arise when things do not go exactly, when they de-
viate from the ideal. It can happen that each of two substances is kept
track of in a normal way over a variety of encounters, but that there are
also mistakes made so that information gathered from both gets col-
lected under the same concept. For example, one might have two people
“mixed up” or “confused” in one™s mind. Similarly, mass and weight
were not distinguished throughout most of the history of science. More
than two substances might also be entwined under one concept. If it is
not definite which among various similar, closely related, overlapping or
nested substances was the one primarily responsible for the information
that has been gathered and/or for the tuning of the (would-be) tracking
dispositions, then the concept is equivocal or vague. Two or more are
being thought of as one. It is likely that normal development of many
kinds of concepts involves a process of differentiating between substances
originally confused together. Perner calls this a process of “focusing ref-
erence” (Perner 1998). It is tempting to interpret much of the history of
science as an attempt to focus reference, for example, distinguishing
weight from mass and oxygen from other oxidizers.
We also can imagine much more serious confusions than simple
equivocation where it is not clear what if any substance or substances
have been kept track of at all. Biological items are, in general, defined
relative to an ideal. A diseased or damaged or malformed heart is a heart
none the less because of historical relations it has to hearts that per-
formed normally. But having described how normal hearts are struc-
tured and how they function, it is of no interest to biologists how far
away from that ideal a thing has to be before one stops calling it a
“heart.”There are no exact borders of the substance heart in nature, and
the biologist is concerned with nature. Similarly, I suggest, to press the
question, in sufficiently abnormal cases, “But please, really, what is the
extension of this person™s substance concept?” is pointless. What, if any-
thing, for example, was the referent of “phlogiston”?6, 7

6 I take these matters up again at the end of Chapter 14.
7 I am grateful to Andrew Milne for some very helpful suggestions on this chapter.

More Mama, More Milk
and More Mouse:
The Structure and Development
of Substance Concepts


The bulk of a child™s earliest words are concrete nouns, including names
of individuals, names of concrete kinds, and some names for stuffs
(“milk,” “juice”). These are acquired in a rush by the dozens between
about one and one half and two years old: “this vocabulary spurt is of-
ten called the naming explosion to reflect the large preponderance of
nouns that are learned” (Markman 1991, p. 81).1 Adjectives come later
and more slowly, and abstract nouns later still. This suggests that the
ability to distinguish concrete individuals in thought and the ability to
distinguish concrete kinds and stuffs may have something in common,
and that concepts of properties and of other abstract objects may not be
required for these tasks. There is much independent evidence that chil-
dren come to appreciate separable dimensions, such as color, shape, and
size, only after a considerable period in which “holistic similarities”
dominate their attention (see Keil 1989, for discussion). Thus, concepts
of properties again appear as less fundamental than those expressed with
simple concrete nouns.
We can interpret this data as suggesting that concepts of substances
are the easiest for a child to obtain, and more surprising, that the on-
tological distinction among individuals, real kinds, and stuffs does not
produce a difference in ease of early learning. I have proposed that, de-
spite their obvious ontological differences, individuals, real kinds, and

1 See Gentner 1982 and Ingram 1989 for reviews, Dromi 1987 for some reservations.There
is evidence that Korean children may usually have a “verb spurt” a month or two before
their “noun spurt” begins. Still the number of nouns soon overtakes the number of verbs
(Choi and Gopnik 1993).

stuffs have something important in common that bears directly on
what it is to have concepts of them. I will now propose that this sim-
ilarity makes them all knowable in a very similar way, and prior to
properties. Though concepts of individuals, real kinds, and stuffs are
traditionally considered to be quite different in structure, I believe that
their root structure is in fact identical. Only as they become more fully
developed do defining differences appear among them. This is because
their corresponding substances have an identical ontological structure
when considered at a suitably abstract level, and because it is possible
to have unsophisticated substance concepts that rest only on this ab-
stract structure.


I have argued that different domains of substances are differentiated ac-
cording to the kinds of ontological grounds that hold them together,
supporting successful inductions over encounters with them. One can
learn on one encounter with Xavier what to expect on other encoun-
ters with Xavier for a different reason than one can learn from one en-
counter with the element silver or with the species dog what to expect
on other encounters. And I have argued that we do not always have a
correct understanding of the grounds of induction that underlie our
successful substance concepts. For example, it is only recently that we
have come to the understanding we now have, as opposed to the un-
derstanding Aristotle had (Section 2.2), of what holds the various bio-
logical species together, and we still don™t have many details in the case,
for example, of asexual animals and easily hybridized plants. Although,
ontologically speaking, individuals are space-time worms while real
kinds are, instead, collections of similar space-time worms, to have the
capacity to understand this ontological distinction would require a grasp
of space-time structure and temporal relations of a sort not acquired by
children until years after they are proficient in the use of both proper
and common names (Nelson 1991). It seems it cannot be necessary to
having the concept of a substance that one understand its ontological
ground. This suggests that it may be possible in general, even though it
may not be ideal, for the concept of a substance to rest merely on its
most abstract structure as a substance, hence for primitive concepts of
substances within the various domains to have a common structure.
Small children might have concepts of individuals, real kinds, and stuffs

prior to any understanding at all of the differences among the ontolog-
ical grounds that in fact organize these domains, and these various con-
cepts might all be formed in much the same way.
What would seem to be primary in the early experience of a child
is merely that milk and mouse retain many of their properties and po-
tentials for use or interaction over various encounters with them exactly
as Mama does. Given this, we might expect the child, indeed we might
expect any animal, in the first instance, to learn how to relate to, and
what to expect from, each of these various items in much the same way.
Putting it Quine™s way, the child™s first recognitions (and those of the
dog) are merely of more Mama, more milk, and more mouse (Quine 1960,
p. 92).To grasp this possibility, however, it is crucial to keep in mind the
differences between classifying and identifying explored in Chapter 3. If
a child™s primary concepts of, for example, milk and mouse were classi-
fying concepts, then of course they could not have the same structure
as her concept of Mama.
The child observes things about Mama when she encounters her, not
things about samples or instances of Mama. The child identifies Mama;
she does not classify Mama. The psychological structure of classification
is the structure of subject-predicate judgment. To classify an item re-
quires differentiating the item to be classified in thought and applying a
predicate to it.To do the latter, the child would need to have prior con-
cepts of instances of Mama “ concepts of timeslices of Mama, I suppose.
But concepts of timeslices of Mama clearly are analytical concepts, rest-
ing on prior concepts of Mama and of times, and thus we go around in
a circle (Section 3.4).
In a similar way, to learn things about milk, the child need not think
of or keep track of instances or portions of milk. And the very point of
having the concept mouse would seem to be that under it, one does not
distinguish Amos from Amos™s brother, but thinks of them as the same.
Classifying animals as mice versus cats versus dogs would involve
thoughts of Fidos and Spots, Felixes and Macaffees, Amoses and broth-
ers of Amoses, each individual to be judged a member of its proper
kind. But there is no more reason to suppose that this is the way the
child first conceives of mouse than that this is how she conceives of
Mama. She need not have concepts of individual mice in order to rec-
ognize mouse again. Early substance concepts, even when what they are
of, ontologically, is kinds, need not be predicate concepts applied to
prior subject concepts. They need not be understood as descriptions
of anything.

In Chapter 4, I discussed how abilities typically rest on alternative
means. My ability to get from home to school rests on many alterna-
tive means, as do my abilities to swim and to tie my shoes. Similarly,
the capacity to identify a substance typically rests on a variety of alter-
native conceptual means. The number of ways I can identify each of
my daughters is nearly innumerable “ through appearance of body or
body parts from a hundred angles, by voice through many mediating
conditions, by posture, clothing, sounds of feet, handwriting, character-
istic habits and activities, various nicknames, and hundreds of identi-
fying descriptions. On the other hand, as with my other abilities, none
of these ways is infallible. Having an ability does not require one un-
failingly to recognize when the required conditions for its means ob-
tain, or even that one understand what these conditions are. Surely
none of these methods of recognizing my daughters, either taken col-
lectively or taken singly, constitutes a definition or criterion of any of my
daughters. Concepts of one™s friends are not analytical concepts but
synthetical, nor are they “recognitional” concepts (Section 3.4). Exactly
the same is undoubtedly true for the infant™s concept of milk and of
mouse. Indeed, the same is true of any adult™s concepts of these things
in so far as they operate as substance concepts. There are lots of ways
to recognize milk quite reliably: by look, by taste, by context, by the
sorts of stains it makes, by chemical analysis, and so forth. Similarly,
there are lots of ways of recognizing mice. None of these ways plays
any role in defining milk or mice. (These are defined in nature.) And
none, of course, is infallible. (Failure to understand these points is called
Concepts of Aristotle™s secondary substances, such as mouse and dog,
show themselves in English grammar as simple subjects of judgment by
vacillating among singular, plural, definite, and indefinite forms to ex-
press the same secondary-substance thoughts. Thus “The lion is tawny,”
“Lions are tawny,” and “A lion is tawny” have, in most contexts, exactly
the same meaning. The apparent determiners and the indication of
number are doing no semantic work in these sentences. The grammar
of English mandates use of some determiner with every count noun,
but the name of a secondary substance does not, as such, express a
counting thought. Nor is there a quantifier implicit in these sentences.
They are not equivalent to “All lions are tawny,” “Some lions are
tawny,” or “Most lions are tawny,” despite the efforts of elementary logic
texts to force them into one or another of these molds. Many lan-

guages, for example Bengali, Finnish, and Japanese, are more semanti-
cally transparent in this respect, using secondary substance names bare
to express this kind of thought.


What the infant identifies is more Mama, more milk and more mouse.
Better, since the idea is not that the infant mistakes Mama and mouse
for stuffs (that would require grasping the ontology of stuffs), what the
infant understands is “here™s Mama again,” “here™s milk again,” and
“here™s mouse again.”2 But various substances and various domains of
substances differ, of course, in the types of knowledge they afford. The
child™s individual highchair retains its overall shape hence its sitting-on
capacity over encounters, but Mama does not (you cannot sit on Mama
when she is standing). Milk and Mama retain their color, while cat does
not. Cat retains its overall shape and rough dimensions, but milk and
wood do not. These various subjects of knowledge must be grasped as
grouped into rough ontological categories according to the kinds of in-
ductions likely to apply to them.
Even for the very young child, surely, a casual look at a new piece of
furniture on the one hand and a new uncle on the other, must indicate
which can be counted on to retain its current climbing-up-on affor-
dance and which may grow tired of the sport. Similarly, preschoolers
know that what is sleepy might also be hungry, but not made of metal
or in need of fixing (Keil 1983).The conception one has of a substance
is not merely the ways one knows to identify it, but also the dispositions
one has to project certain kinds of invariances rather than others from
one™s experiences with it. One pole of a substance concept consists of
more or less reliable means by which to recognize the substance, the
other pole is a rough grasp of an applicable substance template or tem-
plates (Section 2.6). An essential part of grasping a new uncle™s identity,
of acquiring a concept of him, is grasping that he is, at least, a certain
kind of physical object (in the broad sense) but better, that he is a hu-
man being. This must be grasped not as a set of properties Uncle has
but as a sense for things possible to learn about Uncle.

2 Some commentators on “More Mama, more milk and more mouse” (Millikan 1998a) did
take me to be claiming infants think of Mama and mouse as masses or stuffs.

In the same way that the child differentiates between (My) Highchair
and (My) Uncle, both in her methods of keeping track and also in the
invariants she projects, she differentiates among individuals, kinds, and
stuffs. She has, perhaps, a concept of Mama and also a concept of
women. She uses different methods to keep track of these, and projects
different invariants over encounters with them. Tracking Mama is one
of the means of tracking women. If it™s Mama again it™s a woman again.
But the concepts are entirely separate, not at all confused together. Sim-
ilarly, knowing what to expect of a connected physical object and
knowing to expect something different of a pile of sand (see Blum
1998, and references therein) shows that the child is capable of distin-
guishing between the domains of application of corresponding sub-
stance templates. And obviously the child™s methods of conceptual
tracking have to be entirely different for objects, kinds, and stuffs. For
example, her method of tracking cat will allow her to generalize from
the cat on her left to the cat concurrently on her right, whereas her
method of tracking individuals, hence Macaffee, will not. Methods of
tracking for one ontological category will not necessarily work for an-
other. In this way, the child™s concepts of Mama, mouse, and milk do, of
course, have to differ.
The child differentiates among individuals, stuffs, and real kinds, yet
her concepts of things in these domains have a common structure. Each
contains some means or other of tracking its appointed substance and a
grasp of how to project some of the invariants defining this substance to
new encounters. This is the most important fact about the structure of
these concepts because it defines their function. It explains what we have
them for. Substances need not be grasped by understanding the princi-
ples that structure them and hold them together, but merely by know-
ing how to exploit them for information-gathering purposes. Just as one
does not have to be able to describe or even to recognize the conditions
required for exercise of one™s other abilities, for example, just as a child
can swim without understanding Archimedes' principle and ride a bicy-
cle without understanding the laws of dynamics, neither does one need
to understand the ontological principles upon which one™s successful
projections of substance invariances depend. Analysis of the world struc-
tures that permit the possibility of human knowing is not the same thing
as analysis of the inner psychological structure of the knowing.
Tradition, on the other hand, claims that there is nothing common
to the structures of concepts of individuals, kinds, and stuffs, let alone of

“here™s Beethoven™s Fifth again” and “here™s white again” (Section 2.5).
Throughout the history of philosophy and psychology, the tendency has
always been to project the structure of the object grasped by thought
into the mind itself. For example, it is thought that concepts of the sort
we are calling substance concepts can be grasped only by understanding
“criteria of identity” for their ontological kinds. But what would the
relevance be, for example, of a “grasp of the criterion of identity for
persons over time” to a practical ability to recognize the same person
again over time? We can™t always reidentify persons by following their
space-time worms around. Besides, dogs are quite good at recognizing
their masters, and babies at recognizing their mothers, even though it is
quite certain that neither conceives of a criterion of identity for persons
over time. Not that there are no situations in which an explicit grasp of
persons, say, as space-time worms might prove helpful. Cross-examina-
tion to determine where the accused was an hour before the crime il-
lustrates that. But, for the most part, we employ quite different methods
to keep track of one another as substances.
Another venerable tradition argues that it is possible for us to indi-
viduate other individuals in our thought only because each such indi-
vidual is uniquely located relative to us in the same space-time. This is
surely a valid point, but not, as this tradition has it, because conceiving
of other individuals requires us to think of their relations to us, an-
choring our thoughts of them beginning with thoughts of ourselves.
The valid point is that having a concept of any substance at all in-
volves the capacity to keep track of it, which in turn means interact-
ing with it, actively collecting together various manifestations of it that
impinge on our senses or appear in our thought over time. And obvi-
ously one cannot collect together manifestations of something not in
one™s own space-time system. What is true and important is that the
activity of collecting and employing knowledge of any individual can
be accomplished only in so far as our world has a certain space-time
and causal structure in which we too are ingredient and to which we
are attuned. That is, for the most part we can find our way about in it.
This should not be confused with the idea that knowledge of or
thoughts about this structure are required for success in this activity.
The capacity to reidentify Mama and learn things about her is a high
level skill exercised in the world. Thinking of Mama is not done just
in one™s head.

<< . .

( 29)

. . >>

Copyright Design by: Sunlight webdesign