Universals and Particulars in a Phenomenalist Ontology

Author(s): E. D. Klemke
Source: Philosophy of Science , Jul., 1960, Vol. 27, No. 3 (Jul., 1960), pp. 254-261
Published by: The University of Chicago Press on behalf of the Philosophy of Science
Association

Stable URL: https://www.jstor.org/stable/185968

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

The University of Chicago Press and Philosophy of Science Association are collaborating with
JSTOR to digitize, preserve and extend access to Philosophy of Science

This content downloaded from

143.106.203.12 on Sun, 14 Apr 2024 23:32:25 +00:00
All use subject to https://about.jstor.org/terms
 UNIVERSALS AND PARTICULARS
IN A PHENOMENALIST ONTOLOGY*

E. D. KLEMKE
De Pauw University

A phenomenalist philosophy which employs the Principle of Acquaintance (PA)

plus the Principle that what exists are the referents of certain meaningful terms, defined
by PA, cannot include either universals or particulars in its ontology, but is limited
to instances of universals as constituting the range of ontological existents. Universals
must be omitted since they are repeatable and, hence, never wholly presented or
contained, whereas the objects of direct acquaintance are wholly and exhaustively
presented. Furthermore, no entities beyond characters (qualities and relations) are
given in direct acquaintance; hence, particulars, too, must be omitted for the
phenomenalist who relies on PA.

I would like to raise some questions about universals and particulars in a

phenomenalist type ontology. Part of the discussion will consist of termino-
logical clarification; part of it will be more substantial. Various conclusions
which I reach will, I believe, be damaging to certain features of a phenomenalist
type ontology. But this need not be catastrophic, for I hope to show that these
features may not be essential to a phenomenalist position.
First, some preliminary remarks are needed. (a) 'Ontology,' in the following,
refers to what exists. I use 'exist' in a special sense, in contrast to cthere is'
or 'there are.' 'There are' is a commonsensical notion. Hence, one may say,
as in ordinary usage, 'There are eight chairs in this room,' or 'There are no
beergardens in Evanston.' 'Exist,' on the other hand, is an ontological term.
What exists are the simples or ultimates designated by an ontology. The
undefined, primitive (descriptive) terms of one's ontology refer to what exists.
For a phenomenalist, e.g., what exists are such things as sense-data, but not
chairs. (b) The term 'individual' is used with respect to existents of a certain
type, namely, what many philosophers have called particulars. Here, particulars
will be the linguistic counterpart, i.e., terms. Particulars, then, refer to existent
individuals. (c) I omit the entire problem of values from consideration. Hence,
in this paper, 'universals' will refer strictly to non-value universals. Whether
there are also value universals, I shall not attempt to decide.
I shall now state my first thesis:

(T1) A phenomenalist philosophy which employs the Principle of Acquaintance cannot

include universals in its ontology.

The Principle of Acquaintance (abbreviated PA), one may remember, in an

empiricist formulation, states that, in order for terms to be meaningful
(sometimes 'significant' is used), they must refer to those things with which
one is directly and immediately acquainted, or which are directly presented

* Received July, 1959.

254

This content downloaded from

143.106.203.12 on Sun, 14 Apr 2024 23:32:25 +00:00
All use subject to https://about.jstor.org/terms
 UNIVERSALS AND PARTICULARS IN A PHENOMENALIST ONTOLOGY 255

(some say 'wholly presented') to one. For the phenomenalist, the things
with which we are directly acquainted are usually held to be primarily sense-
data. But such things as awarenesses, objects of memory, etc., may be included.
PA has usually been defended hand-in-hand with another principle which I
shall call the Ontological Principle (OP). OP states that what exists are the
referents of certain meaningful terms (meaningful by PA), namely, primitive,
undefined, descriptive terms. From these two principles, it follows that (in
this view) what exists, in the sense of being ontological simples, are those
things and only those things with which I am directly acquainted.
Some philosophers' hold that the solution of philosophical problems may
be achieved by talking about an ideal language. In this view, PA and OP
need to be slightly revised. PA now states: All primitive, descriptive signs
of the ideal language refer to entities with which I am directly acquainted.
And OP says: What exists shows itself by the inventory of the kinds of un-
defined, descriptive signs of the ideal language. For the issues which are here
being discussed, either of the formulations-linguistic or non-linguistic-
will do. I prefer to speak primarily in terms of the latter. With minor changes,
what I shall say applies to the ideal language version as well.
The espousal of PA alone, i.e., without OP or a principle similar to OP,
serves no purpose. Hence, as I mentioned, OP, or a similar principle, usually
accompanies PA. In this paper, I shall often speak simply of PA, as in (T1)
above. Frequently, when I do so, I shall mean PA in conjunction with OP
or a principle similar to OP.
The reader's reaction to (T1) may be: "This, of course, implies, or is nomi-
nalism." I answer: Not necessarily. (T1) suggests either of two ontological
positions: (i) 'Only individuals exist; universals do not' (nominalism); or
(ii) 'Instances of universals exists; universals may or may not'. In (ii), 'in-
stance of a universal' does not mean 'individual.' An instance of a universal
is, for example, a quality which an individual may possess, or, simply a quality.
My reason for not calling this a universal will be seen below. I hardly know
what to call the second position. It surely is not identical with nominalism,
for, oddly, it does not imply that individuals must exist. It says nothing of
individuals. That is, particular-terms may have no referents in (ii); hence,
what we might call individuals will be, say, mere fictions, lacking ontological
status. (More of this below.) A third alternative, realism, is what (T1) holds
to be ruled out for a phenomenalist ontology. I shall attempt to show this
shortly.
Before turning to my argument, some definitions are needed. The definitions
which I shall provide are not real definitions. The definitions of 'universal'
and 'individual' are largely reportive. The definition of 'instance of a universal'
is partly stipulative, but upon the basis of what seems to follow from the
definition of 'universal.'
One kind of universal, that is, the "Platonic" universal, is of course, ruled

' E.g., G. BERGMANN, The Metaphysics of Logical Positivism. Longmans, Green and Co., 1954.

This content downloaded from

143.106.203.12 on Sun, 14 Apr 2024 23:32:25 +00:00
All use subject to https://about.jstor.org/terms
 256 E. D. KLEMKE

out immediately for the phenomenalist, since he is not directly acquainted

with such entities. Hence, our definition of 'universal' must be such that
universals need not be non-spatial and/or non-temporal entities. I suggest
the following definition:

Universal: the repeatable character, whose instances are commonly called qualities
of individuals or relations among individuals; hence, that which is never wholly
contained in any one presentation, unless it exists only in that one presentation
(which could never be known with certainty to be the case).

By 'quality,' I mean nothing mysterious. A synonymous expression might

be: 'the kind of thing to which predicate terms refer.' If one senses a red,
square datum, then in the sentence 'A is red and square,' 'red' and 'square'
refer to qualities. Our other two terms may now be defined. I shall give two
definitions of 'instance of a universal,' for reasons which will be evident later.

Instance of a universal:
(A) That which is a quality of an individual or a relation in which two or more
individuals stand, and with which we may be directly acquainted.
(B) That which is a quality, or a relation which unites two or more qualities or two
or more conjunctions of qualities.

Individual: That about which a quality or qualities may be predicated, or that which
may stand in a relation to other individuals.

Now for my argument. Universals, if they exist, cannot (with the possible
exception noted) be the objects of any kind of direct acquaintance. I cannot see,
hear, etc., (either in the sense of perceiving or sensing) any universals. What
I do sense are instances of universals. Neither can I remember universals, or
become reflexively aware of them, etc., since (if PA is accepted) such "acts"
depend upon earlier sensings. Hence, if universals exist, then, in order for
me to know them, or be in any way acquainted with them, I must know them
in some way other than direct acquaintance. This follows from the fact that
universals are (with the possible exception noted) repeatable, and, hence,
never wholly presented or contained, whereas the objects of direct acquaint-
ance, for the phenomenalist, are wholly and exhaustively presented. If, for
example, I sense a red datum, I have no reason to believe that I have been
presented with the universal red. The most that I can know is that I now
sense a specific quality which I call red. If I very soon after sense another
red datum, then I know that I cannot have been acquainted with the universal
red in the first sensing. And, similarly, I have no basis for knowing that I
am now acquainted with it, as I can anticipate the possibility of another
sensing of red.
It should be noted, by the way, that the existence of instances of universals
does not imply that there is anything over and above the instances, i.e., that
universals exist. Universals, as well as instances, may exist, but the mere
acknowledgment of instances does not necessitate the acknowledgment of
universals ("in themselves"). In this case, the term 'instance of a universal'

This content downloaded from

143.106.203.12 on Sun, 14 Apr 2024 23:32:25 +00:00
All use subject to https://about.jstor.org/terms
 UNIVERSALS AND PARTICULARS IN A PHENOMENALIST ONTOLOGY 257

is, perhaps, an unfortunate one. Very well, call it something else. I shall
continue to use the term, however, and this will do no harm as long as one
remembers that the existence of instances of universals does not imply the
existence of universals. For example (if I may express it in this way), although
there may exist this red and that red, etc., this does not necessarily mean that
redness also exists. Whether it does or not must be determined by some
other kind of probing.
I take it, then, that (T1) has been established: A phenomenalist philosophy
which employs PA cannot include universals in its ontology.
Some questions might arise at this point.

(1) "Why distinguish universals from instances of universals in this

fashion? Why not call the instances of universals 'universals' ?" This question
reflects a kind of muddleheadedness. The answer is: Simply because instances
of universals are not universals. Again, universals, if they exist, are entities
with which I cannot be directly acquainted. Instances of universals, on the
other hand, are objects with which I am directly acquainted-qualities like
(this) red, specific cases of relations like to-the-left-of, etc. Hence, even though
we are not acquainted with universals, we must distinguish universals from
instances of universals upon the basis of our knowing that, if universals exist,
they are not the sorts of things which instances, in fact, are. We can sense
instances of universals, and sensing is a form of direct acquaintance. If
universals exist, we could never sense them, or in any other way be directly
acquainted with them. We would have to know them by some other kind of
apprehension-conceiving, or whatever it may be.

(2) "Why distinguish instances of universals from individuals? Are not

individuals precisely instances of universals?" So someone else might ask.
No, such assertions are again the result of confusion. (I am aware that they
have been often stated, but this does not make them less confusing.) Indi-
viduals, if they exist, are, or would be, what instances of the quality-type can
be predicated of (i.e., something which has a quality) or what instances of the
relational sort relate (i.e., something which is in a relation to something else).
Thus, it hardly needs to be pointed out that, if individuals exist, then they
are distinct and different from instances of universals. Some philosophers
have said that individuals are, or would be, what "stand under" the qualities.
They are those objects which qualities can be qualities of, or they are those
things which stand in relations.
Someone might say, at this point: "Ah, but this violates PA too. We are
not acquainted with individuals in addition to instances of universals. Such
individuals, if they exist, would have to be reached by understanding, reason,
etc., i.e., by some acquaintance or apprehension other than direct acquaintance.
Very good! That is my next thesis. Incidentally, it will now be noted why
I always used such phrases as 'if they exist' when I discussed individuals
above. And this is why the second definition of 'instance of a universal' was
needed. I suggest, then, that a second thesis reads:

This content downloaded from

143.106.203.12 on Sun, 14 Apr 2024 23:32:25 +00:00
All use subject to https://about.jstor.org/terms
 258 E. D. KLEMKE

(T2) A phenomenalist philosophy which employs the Principle of Acquaintance

cannot include individuals in its ontology.

This may seem inadequate, for two reasons, one having to do with the
notion of individuals, the other, with the notion of universals.

1. We are, in fact, acquainted with individuals of a sort. Consider two

white squares, one to the left of the other. One might say that it is simply
given to me that there are two of them. Hence, individuals exist as well as
instances of universals. However, what is given to direct acquaintance are
not individuals which are things (entities, etc.) underneath or with the qualities
white and square, i.e., not any other objects, but simply two (white-square)s.2
In every case of direct acquaintance by sensing, I find a quality, or a con-
junction of co-present qualities, or a relation of such. I do not, in direct
acquaintance, apprehend an object called the individual, as over against the
qualities.
Someone may suggest: "Of course, for the phenomenalist, there can be no
substantial individuals, but there can be non-substantial individuals. Let us
call the quality or conjunction of co-present qualities non-substantial indi-
viduals." If this suggestion were fruitful, then we might revise (T2) to read:
"A phenomenalist philosophy which employs PA cannot include substantial
individuals in its ontology." But this won't do, for two reasons. First, we
blur the distinctions which we have already achieved if we call qualities or
conjunctions of qualities 'individuals,' albeit non-substantial individuals.
Second, non-substantial individuals (of the above sort) are not individuals
in any ontological sense. They "lack being," as some say. If, for example, we
examine one of the white squares (mentioned above), we find nothing over
and above (or under and below) the qualities white and square (not whiteness
and squareness!) If these qualities did not exist, there would be nothing with
which to be acquainted. Hence, we can hardly include non-substantial indi-
viduals in a phenomenalist ontology.
Nevertheless, I believe that it will be possible to employ the term 'individual'
in a phenomenalist philosophy. But it will be a merely fictive term, a kind of
shorthand or technique for designating instances of universals or conjunctions
of instances of universals, or for distinguishing one instance of a universal
or one conjunction of such instances from another. One conjunction of white
and square, e.g., may be called'a' (or 'this' or 'John'), and another 'b' (or
'that' or 'Henry'). Particulars (terms), then, will be merely pointers. They
have no ontological referents. In this sense, they are similar to logical rather
than descriptive signs. Hence, (T2) still stands as above.

2. So much for the case of individuals. One might still feel a dissatisfaction
with respect to the status of universals. And there is a problem here which
may cause a radical difficulty for the phenomenalist. Incidentally, this has

2 I hardly know what notation to use here.

This content downloaded from

143.106.203.12 on Sun, 14 Apr 2024 23:32:25 +00:00
All use subject to https://about.jstor.org/terms
 UNIVERSALS AND PARTICULARS IN A PHENOMENALIST ONTOLOGY 259

been pointed out before. The argument runs something like this. Take two
white patches. I have shown that the white with which we are acquainted is
an instance of a universal, and not a universal. Now suppose that I say about
the two white patches-call them a and b-'a resembles b' (or 'a is similar
to b'). A philosopher who claims to intuit, apprehend, or conceive universals
(or, at least, who posits them) will perhaps say that the two white patches
are white because they possess the universal character whiteness, or because
they participate in whiteness, etc. And he will say that they are similar to each
other and to other white patches by virtue of the universal similarity. Since
PA prevents the phenomenalist from admitting genuine universals in his
ontology, he could hardly say anything of this sort. Yet he will, perhaps, want
to have some way of knowing and stating that the two white patches are
similar. Presumably, he could select a paradigm case, some object-call it
w-which could be used for the purpose of giving an "ostensive definition"
of white. Very well, suppose that I now say that a and b resemble (or are
similar to) each other because they both resemble w. It has been held, by
Russell, among others, that the latter resemblance will have to be a (relation
universal. "Since there are many white things, the resemblance must hold
between many pairs of white things; and this is the characteristic of a uni-
versal. It will be useless to say that there is a different resemblance for each
pair. For then we shall have to say that these resemblances resemble each
other, and thus we shall be forced to admit resemblance as a universal. The
relation of resemblance, therefore, must be a true universal."3
I shall return to this problem presently, but, first, I would like to consider
an assertion which is often made at this point (and by Russell too). The
usual argument is: "If you admit one universal, why not all ?" This is not
an argument, but an appeal. The answer is: Because one cannot. Instances
of universals are still instances of universals, and not universals. Everything
which I said above still holds for them. In other words, even if we are forced
to allow similarity as a universal in our ontology, this does not permit us to
think that instances of universals are universals, since they clearly are not,
for reasons stated above. Nor can we permit other universals, since, as long
as we accept PA, we cannot include such universals in our ontology. The
crux of our problem is: which do we give up, PA or similarity ?
The phenomenalist may wonder: Must we give up either ? Granted that
we need similarity (or resemblance). However, do we need it as a universal ?
Isn't similarity as an instance of a universal sufficient ? His critic will return:
"But then we are involved in the infinite regress of resemblances of resem-
blances, etc." Well, let's see. Consider two white patches again. That a is
similar to b is surely given to me, or directly presented to me, in the same
sense as a's being to the left of b is directly presented to me. To-the-left-of
is an instance of a universal. Likewise, similarity is an instance of a universal.
And, like to-the-left-of, it is simple and unanalyzable, as any referent of a

3 The Problems of Philosophy (Oxford: Oxford, 1912; reset, 1946), p. 96.

This content downloaded from

143.106.203.12 on Sun, 14 Apr 2024 23:32:25 +00:00
All use subject to https://about.jstor.org/terms
 260 E. D. KLEMKE

primitive, descriptive term is. I am, then, directly acquainted with similarity.
Furthermore, no infinite regress is involved as long as we stay at this level.
For this similarity is directly and wholly presented to me. I need not compare
this case of similarity with another case in which, say, I hold c and d to be
similar; nor need I refer both cases to some paradigm case above it.
"But it makes sense to ask if this case of similarity (a and b) is similar to
some other case of similarity. And if we do ask it, aren't we forced to admit
a universal similarity here-a second-level similarity which relates the two
first-level cases of similarity ?" I see four possible approaches toward resolving
this problem. The first two go about the settling of the issue in more tradi-
tional terms. The second two give a linguistic turn to the problem.

(I) We might say that there is really no problem here, after all, that it
makes no sense to ask if two cases of similarity are similar. Similarity is a
relation like identity. Suppose that A is identical with B (call this identity I)
and that C is identical with D (call this 12). Now what in the world does it
mean to ask if J1 and '2 are identical ? We can ask it, but it is very strange to
do so. Furthermore, it isn't necessary. We have to go no farther than particular
cases of identity. The same holds for similarity, if we are talking of actual
similarity and not similarity in some respects or nearly similar, etc. On this
view, then, we have no occasion to go beyond first-level similarity relations.

(II) Or we might say that we do require one universal, similarity. Since

PA won't allow us to include it in our ontology, PA is too restrictive. This
does not necessarily mean that we must throw out PA. We could keep it,
in this view, for a large share of our ontological inquiry. But at a certain
point, we shall require some form of acquaintance other than direct acquaint-
ance. Our present problem is such a point. For if we keep PA, our phenomen-
alist ontology is inadequate to the case of universal similarity.

(III) Or we can take the "Oxford" approach. You are asking questions
like 'Do we need similarity?' or 'Does similarity exist?', says the Oxford
man. You seem to think that your question is like 'Do we need jet-planes?'
or 'Does Winston Churchill exist?' You are confused and misled by grammar.
In asking 'Does similarity exist?' you merely mean: Are there certain words
which are general words, rather than proper names ? And, of course, the answer
is Yes. Certain general words, like 'red' and 'square', are at a more direct
level. Others, like 'similarity' and 'identity' are presupposed by the former.
But both types are general words. And both are needed or, at least, are con-
venient for discourse.

(IV) Finally, we might take the ideal language attack. Here, again, it
makes no sense to ask 'Do universals exist (or subsist, etc.) ?' If you go on
asking the question in this traditional way, how would you ever find an answer ?
However, a question like this has a special philosophical core which can be
preserved if we reconstruct the question, and thereby the answers or possible
answers. The reconstruction reads: Does an ideal language-one which

This content downloaded from

143.106.203.12 on Sun, 14 Apr 2024 23:32:25 +00:00
All use subject to https://about.jstor.org/terms
 UNIVERSALS AND PARTICULARS IN A PHENOMENALIST ONTOLOGY 261

allows us to say (commonsensically) everything that needs to be said-

require predicate terms as well as subject terms?' If you say No, you are a
reconstructive nominalist. If you say Yes, you are a reconstructive realist
Which is the right answer? Well, try it out. See which version works. Trying
it out reveals, according to some, that predicate terms are needed.
This is nice, but it doesn't eliminate our problem. For we are directly
acquainted with the referents of predicate words like 'red' and 'square' and
relational words like 'similarity' when we say 'a is similar to b.' But, again,
what if we ask whether or not a's being similar to b is similar to c's being
similar to d? Are we directly acquainted with second-level similarity ?4
Certainly not. Hence, a complete reconstructive phenomenalist ontology
breaks down as long as we accept PA, or, rather, as long as we refuse to grant
a form of acquaintance or apprehension in addition to direct acquaintance.
(III) gets no where, as far as I can see. It doesn't solve the problem; it
dissolves it, as Oxford philosophers are fond of saying. But I fear that too
much gets dissolved. (IS) and (IV) deserve to be taken seriously. Both involve
a breakdown of phenomenalism, whether traditional or reconstructive, unless
there can be some changes made. We do not need universals as long as we
stay on the level of qualities such as red and relations like to-the-left-of.
But once we get to second-level similarity, e.g., (and these positions do not
prohibit us from "getting" there), we need a universal and not merely an
instance of a universal. But we cannot have universals as long as we refuse
to go beyond PA. So, if we keep PA, our phenomenalism breaks down; that
is, it is inadequate to resolving all philosophical problems. If we keep phenom-
enalism, we have no choice but to allow other forms of apprehension besides
direct acquaintance. Hence, PA is insufficient. Whether some further kind
of apprehension can fall within the framework of phenomenalism would
now be the problem for anyone who adopted alternative (II) or (IV).
We avoid the latter problems if we select alternative (I). And I think it
is a good alternative. First, as in the case of identity, it does seem rather
strange to ask about resemblances of resemblances, if we mean exact similarity,
and not the sort of similarity whereby baby blue is similar to navy blue, just
because both are blue.5 Second, PA is a cornerstone of phenomenalism. If a
philosopher, in this case, a phenomenalist, adopts PA, then he also adopts
a somewhat more limited ontology, one which is not populated with numerous
kinds of entities, but which is adequate to the "narrower" problems with
which he deals. I believe that the first alternative meets the needs of the
phenomenalist. Of course, a more rationalist type philosopher will find all
of this to be inadequate. But perhaps he has more ambitious aims than those
of the empiricist-phenomenalist. Whether such aims are worthy of our approval
I shall not here attempt to say.

4 I.e., with universal, not this instance of it. (This is a meaningful question for this position.)
5 In the latter case, the question is even more inappropriate. Or, if not, then it could perhaps
be dealt with by the use of paradigms.

This content downloaded from

143.106.203.12 on Sun, 14 Apr 2024 23:32:25 +00:00
All use subject to https://about.jstor.org/terms