quine's theory of ontological commitmentdigitool.library.mcgill.ca/thesisfile47630.pdf · as a...
TRANSCRIPT
QUINE'S THEORY OF ONTOLOGICAL COMMITMENT
•
•
QunrE1S THEORY 01!' ON!OICGICA.L
COMMI!MEI!r
A. !beais sulmd. tted in partial :tul.i'ill•ent of the requirelllents f.or the degree of Jlaster of .Arta in the l'acul ty of Gre.duate Studies and Research of JlcGill trni.versit7.
April 1966 Montreal, P. Q.
TABLE 01!' CONTENTS
page
Abstract.
Foreword ••••••••••••••••••••••••••••••••••••••••••••••••••••• Introduction ••••••••••••••••••••••••••••••••••••••••••••••••• Chapter I. The Origin of the Method ••••••••••••••••••••••••• Chapter II. The Extension of the Method •••••••••••••••••••••
I 1 :; 10
Chapter III. Chapter IV. Chapter v.
Some Objections •••••••••••••••••••••••••••••••• 20 The Conceptual Priori ty of Na tura1 Language • • • • • 30
The 11E-problem" ••••••• • •• • • • • • • • • • • • • • • • • • • • • • • • • 40 Conclusion ••••••••••••••••••••••••••••••••••••••••••••••••••• 46 Bibliography ••••••••••••••••••••••••••••••••••••••••••••••••• 47
FOREWOBD
The :f'ollowing rW.es of' ref'erence have been observed in this the.sis:
1. .All quotations :t'rom, and ref'erences to, worlœ by' Quine, and worlœ whiCh mention Quine, ref'er to the bibliograpJv' :f'ound at the end of' this thesis. Th:u.s, tor instance, "Strawson (2~ p. 196', means: "Strawsoa., Individu.als, p. 196."
2. Journal articles wbich have been reprinted in anthologies are not present in the bibliograpby. !he-,·titles of' 'BUCh arti:eles occur on1y in the te:x::t anà/or f'oot-notes. !J!huai· tor instance,
"Warnockt "Metapbysics in Logic" in Flew {lo) p;:re,'!lneans, "l!'lew CLdJ P• 78." · ·· · "
I.
INTRODUCTION
The invention of !'u.:nctional calculua, and more generally the rigoroua
development of s,ymbolie lo~c in the second half of the nineteenth centur,r
and in the beg1nnjng of the twentieth century have haaa great impact on
philosopb\Y. However, nowa.d.qs there ia a controvers;y concern:l.ng the
relevance of techniques of symbolie logic for solving philosophie problems.
:Both aides fi8ree that as a resul t of these techniques, philosophera have
ga.ined a better insight into their problems, thus oJJ.ly the rat~ge of appli
cation of these techniques is at issue.
The problem of ontolog;y is probably the most appropriate testing
ground for this contronrs;y, not oJJ.ly because this philosophie problem
has recei ved great attention in twentieth ce:ntu.ry logical theory, but
also because it is thought to have been most affected b;y the earlier
mentioned developments in logic.
In 1113' thesis I examine Quine' s the ory of ontological commi tment in
the light of the controvers;y which it haa generated. I first give one
of the earliest attempts at applJing the techniques of s;ymbolic logic
to the problem of existence (Ch. I). Then I outline Qu,ine1s treatment
of the problem (Ch. II Sect. I) which is generally thought to be the
most ~orous contemporary attempt to bring s,ymbolic logic to bear on
problems of ontology.
2.
I argue against Quine 1s position, the position that the o:nly philo-1
sophical problem o~ existence is t~t which is coJmeeted wi th the exis-1
tential qua:ntifier. After having argued that Quine i:nte:nds to solve
the general philosophical problem of existence by bis criterion o~
ontological eommitment (Ch. II, Seet. II) I argue, ~ollowing Strawson,
that Quine's thesis concer.ning the elimdnation o~ siDgular terms involves
a misconception. (Ob. III, Sect. I).
Next I outline the objections o~ Carnap and Warnoek to Qaine 1s
notion o~ ontological commitment. (Ob. III, Sect. II).
In J'If3" ~ourth chapt er, I diseuse how Quille does reply to Carnap 1 s
objection and could reply to Warnock1s, and I argue that Quine 1s views
on reco:nstructiDg our natural language rely on untenable assumptions.
I conclude J'If3" thesis by arguing (Ch. IV) that Quine' s theory' o~
ontological commitment, involving a reconstruction o~ our natural
language does not deal adequately wi th the philosophical problem of
existence.
I
THE ORIGIN Ol!' THE METHOD
I
As a preliminary to my main topic, Quine 1 a theory of ontological
commi tment, I shall diseuse Russell' s the ory of descriptions; al though
I shall introduce Strawson1a opposition to Russell 1s programme, I will
not offer any definitive conclusions about the controvers.y.
Rasaell's chapter in Introduction to Mathematical PbilosopbY en-
1 2 ti tled "Descriptions" and Strawson 1 a article "On Referring" will
receive most of my attention. Nevertheleas, in my exposition I shall
also make use of Rusaell 1s "On Denoting"3and Strawson1s Introduction to
Logical Theo;r4•
Russell 1 a theor.r of descriptions is intended to solve the problem of
nunreality". He saya: "It ia argued, e.g. by Meinong, that we can apeak
about "the golden mountain" ,, "the round square" and so on; we can asaert
two propositions of which these are grammatical subjecta; bence they must
have some kind of logical being, since otherwiae the propositions in which
they occur would be meaningless"5• For Russell Meinong's argument shows
1. Russell (20) p. 167-180. 2. Caton (6) P• 162-195. :;~ Peigl (9) p. l03-ll8. 4. Strawson (25) 5. Russell (20 p. 169.
4.
a lack of "tbat feeling for reality wbich ~t to be preserved even in 1 1
the most abstract etudies" • He be~ieves that logicians like Meinong
were led to such conclusions by their und.ue emphasis on the grammatical
forme of propositions as guides for analysia; they lacked the apparatus
of propositional tunctions~. But, ":in deal.i.ng with propositions, we are
dealing in the first instance wi th sy.mbols, and if we attribute signifi-
canee to groups of symbole which have no significance we,shall fall into
the error of admi tting unreali ti es. n3 "Logic must no more admit a uni-
corn than zoolog;y can; for 1ogic is concerned wi th the real world just
as truly as zoo1og;y, thougb. with its more abstract and general features". 4
The Russellian posi tlhon is that i t is significant but false to say
"I met a unicorn", since the proper 8Il8.lysis reveals that "a unicorn" is
not a constituent of the proposition, though the concept "unicorn" is5•
To illustrate this point with another examp1e; we might ask for the
meaning of: ( 1) "A student knows his algebra"
(2) "student"
but not of: (3) "a student"
The 1ogical for.m. of (1) isz "x knows his algebra" where x can be-,planted
by a name or a description.
Now, a word is appropriate about namee and descriptions. It is
Busse11' s v:Lew that in the world there are things which we sometim.es nam.e
and sometimes de scribe. The question whether a name names a:n;ything is
senseless because to be a name.!!. to name something. That which it names
is i ts meaning. Consequently, "% is real• or "x is unreal" mak:e no sense
1. ibid. P• 169. 2. ibid. P• 168. 3· ibid. p. 170. 4. ibid. p. 169. 5. ibid. p. 168.
5.
when the values of x are names; on the other band when values are descrip-
ti ons the propositions make sense, they can be true or false.
The distinction between names and descriptions is seen from an exami-
nation of identity contexte. 1\'or example, the :f'unction: "x = x" will
become a true proposition whenever the x is replaced by a description.
Wha.t does this .mean?
Russell divides descriptions into
A. Defini te descriptions
B. Indefinite descriptions
Roughly, the difference between A and B is that in À there is an impli-
cation of uniqueness whereas in B there is not.
Thus:
Members of À bave the form "the so and so".
Members. of B have the form "a so and so".
Russell wants to say that this difference is intuitively clear so that
when I say "The man next door is noisy" I am taken to mean:
(a) "There is a man next door."
( b) "There is only one man next door."
( c) "There is nothing which is a man next door and is not noisy."
and when I say "A man next door :ls noisy." I am taken to mean:
( d) "The re is a man next do or."
(e) It is not the case that all men next door are not noisy."
1 One chooses between A and B mainly on epistemological grounds. But,
definite or indefinite description may or may not describe. A does not
describe if one of (a), or (b) is false. B does not describe if (d) is
1. In this example, one would choose between A and B according to one's information about the source of the noise.
6.
false. Bence, if in the identi ty context we replace x by a defini te
description which does not describe definite~ the resulting statement
of identity will be false. "The round square is round" will be false.
From this, it follows that substituting a name for the x in a tunction
will result in a proposition different from the one which resulte from
substitttting a description for the same x. For example, consider "K
wanted to know Whether x was Scott".
"K wanted to know whether Scott was Scott" is different from ffK
wanted to know Whether the author of Waverl;y was Scott", Where "Scott"
is a name and "the author of Waver~" a description.
As far as existence is concerned the thesis is as followsz
I. "x exista" is senseless When the value of x :ia a name.
II. "x èxists11 or "x is so and so" is a sig:Difican.t proposition if
the value of :x: is a description and i ts truth, or talai ty depends on the
truth or falsity of the a's, b1s and c 1s corresponding to the description.
Àceordingl;y, "the present KiDg of ]'rance is wise•• will be a signifi-
cant but false proposition beeause in
a0 = "There is a King of France at present"
bo = "There is at most one XiDg of France at present"
co = "There is nothing wbich is a KiDg of France and is not wise"
ao is clear~ false, therefore the eonjunet ao and b0 and c0 is false.
Thus, we have a decision procedure for determôning the truth or falsity
of existential propositions where a description is a eomponent; we re
phrase it as a eonjunct of the a•s, b1s and c 1s, and determine the truth
values of the conjunets by empirical means. If they all turn out to be
true, the erlstential proposition is true; if one or more of them turne
out to be false, the proposition is false. A corollary to Bussell's
solution of the problem of ttunreali"t;;r" is that whenever there ie a genuine
problem as to whether an e:x:istenti4 proposition ie tru.e or false, its
subject term is treated as a deeeri~tion. For e:z:am.ple, einee there ie 1
a genuine problem whether "Homer exista" is tru.e or false, we treat
"Homer" not as a name but as a shortha:nd for some description or other.
The following is a statement of Strawson's opposition to Russell:
"I think i t is tru.e to say that Russell 1 a theory of Descriptions •••
is still widel;y aeeepted among logicians as giviDg a correct aeeount of
the use of auch expressions" (as 1 the so and so 1 ) "in ordin.a.r.Y language.
I want to show in the tiret place that this theory, eo regarded, embodies
some fundamental mistakes".1 It has been seen that Russell anal;yeed
"The present King of :France is wiselt'~ .. sa:sr s,as ao and b0 and e0 • He
arrived ·, says Strawson, ·at this anal;ysis b;r asld.ng himself what would
be the eircumstanees in which we would sa:y that anyone who uttered the
2 sentence S had made a tru.e assertion.,, &o, b0 . and e0 are onl;y the nec es ..
sary conditions of making a true assertion b;y uttering the sentence B,
but they do not constitute a correct analysis of sentence s.
At the basie of this opposition lie the following statements:
a. "We cannot talk of the sentence being true or false, but onl;y
of its beillg used to make a true or false assertion."3
b. "':Mentioning' or 1referring1 ie not sometbing an expression does:
i t is something that someone ean use an expression to do. n4
Bussell'e confusion about a and b gave rise to th1nk1ng that "referrillg
or mentioning, if it oceurred at a11, mu.st be meaning". Strawson1s con-
Caton (6) p. 16;. ibid. p. 169. ibid. p. 169. ibid. p. 170.
Phrase in parentheses is 1l13' own. 1'
a.
clusion about l.!!he XiDg of li'rance is wise" is that the question of 1 ts
being true or false does not arise. The speaker of the sentence does not
mention a.nytb.irlg or ~body. Thus, he is spuriously using a significant
sentence rather than, as Russell would say, gem.d.nely utteril::tg a false
one. The "!he" of "The auch and such" shows, but does not state, that
we are, or intend to be, referrillg to one particular indi vidual of the
species "Such and such". "The'' used in such a way implies that the
existential conditions described by Hussell are fulfilled. That there
is a King of France is presupposed1 by The King of l'rance is wise· but
does not entail i t becauee the assertion of one and the denial of the
other does not lead to a contradiction. 2
!hus, while Strawson agrees that the position Russell seeks to
reject, i.e. that "round square","the golden mountain" have some kind
of logical being, should be rejected, he maintains that it could be
done more satisfactorily by avoiding the confusion Bussell and his
earlier opponents (e.g. Meinong) share:.3 It I!Dlst be noted that Straw
son's position is not without its critics. 4 levertheless, a few points
remain unassailed:
1. Using an expression to refer to sometb.irlg is not in and of
itself to make any assertions about that thing but to lay dow.n the
grounds for assertions.
and more importantly:
2. :&nphasis should be placed on the different r&les expressions
1. Strawson gives the following def~tion of "presupposition": K is presupposed by L, if and only if , the truth of K is a necessary condition for the truth or falsity of L. Strawson (25) p. 175.
2. Caton (6) P• 185. 3'· ibid. P• 165. 4. Ayer (1), Russell (21), Black (4).
9.
play in a language and not on an analysis of expression as suoh wi thout
respect to their varied uses in a ~ety of contexte. '
!his latter point is a crucial one f ft is in fact discredi ting the
Russellian claim about the importanc' of the apparatus of propositional
:f'unctions, for by suoh an apparatus we can e:mmine only the conte:x:t-in-1
variant :teatures of statements leavüg an examination of their contert-
variant use aside.
Tbns this chapter has served a double r&lea
a) It bas introduced the historical roots from which will spri.ng
Quine's attempt to apply :tormal logical considerations to the
problem o:t existence, and at the same time
b) it has show.n a germ o:t the opposition to such programme.
II
THE EXTENSION Ol!' THE METHOD
10.
I
Gi ving a full amount of Quine' s theory of ontological commi tment
ia a rather complicated taak. The reaaon for this is twofold:
a) It is di:t'f'icult to malte out the extent of Quine' a claims.
b) Quine discusaes many important issues which the reader feels
are somehow related, yet the connections are seldom made explicit.
For Quine the ontological problam can be put simp~ - What is there?
He begins his treatment of the problem by an inves-tigation into the
nature of ontological disputes. He suggests that at some point in a
philosophie discussion one of the participants may assert that there
are certain thil:Jgs which the other llll!q, wi th equal vehemence, de~.
Quine further notes tbat the proponent of the negative side in the onto
logical dispute seems to be at a disadvantage since he cannat admit that
his opponent disagrees with him. The situation is this. A asserts
"Unicorne exist" and if B wants to deey this he ia in the stra:nge posi
tion of naming some things (unicorne) and saying that they do not exist.
Thus, as soon as B tries to express his disagreement with A he seems to
be contradicting himself. Quine calle this the "Pla tonie riddle of
nonbeing". He finds "traditional" attempts to solve the riddle inadequate.
11.
SayiDg ei ther that unicorne erlst in Olle sense and tables erlst in a
different sense; or distinguishiDg ~etween "subsiét~and "existing"
lead to unnecessary theoretical complications.1 In~ case, these
tradi tional answers have "united in ru1 ni ng the good old word • axiat'"
because" we have all been prone to sa;y, in our common-sense usage of
1exist 1 , that Pegasus does not exist, mean.i.Dg simp]J' that there is no 2 auch entity at all". !o dissolve the riddle Quine otfers a solution
in the spirit of Russell' s theo:r:y of descriptions. He wants to have a
. managea ble way of handling ontological disputes, but more importantly
he wants to refute, or at least to discredit Platonism, in the sense
in which Platonism is associated with the alreaqy mentioned problem of
nonbeing and the tradi tional answers to i t. In Quine' s attack on Pla-
tonism, however, the spirit of nominalism is more prominent thau in
Russell's attack on its Meinongian version. It will be remembered that
Russell' a main concern was to maintain a "sense of reali ty", whereas
Quine is also concerned with m1n1m1sing the number of entities presup-
posed b.1 our discoures. This feature of Quine 1s attack is illustrated
by his argument; against talk1ng of "unactualized" and "actualized"
possibles. ~ positing there two additional rea1ms of entities we open
ours el ves to a who le arra.y of add1 tional problems which were not there
in the first place.4 In his view a bloated ontology besides generating
unnecessary problems is, for aesthetic reasons, un.des1rable. Quine' s
programme is as follows:
He takes from Russell the notion that apparent D&Dles, complex
1. Quine (16) p. 4. 2. ibid. p. ;. ;. ibid. P• 4. 4. ibid. P• 4.
12.
descriptive phrases like "the author of Waverly" or "the present
King of France" can be ~zed as tragments of the wbole sentences 1n
which they occur. He goes further: 1
"The virtue of this analysis,"
he says "is tha.t the seemi:Dg name, a descriptive phrase,_i~ paraphraaed
1n context as a so-called 1ncomplete symbol. No unified expression is
offered as an ~sis of the descriptive phrase, but the statement as
a whole Which was the context of that phrase still gets its fUll quota
1 of meaning--whether t:rue or false". Thu.s, to apply Hussell 1 a theory to
"The round square cupola on Berkeley Collage is pink", we get "Something
is round and square and is a cupola on Berkeley College and is pink".
The second statement, says Quin~ceases to contain any component whiâh
even purports to name the alleged entity "round square cupola". The
burd en of the claim is placed on the pronoun "something".
He now extends Russell's analysis to all singular terme, terme
which purport to refer uniquely to enti ti es. lll singular terme can 1n
principle be elim1nated.2 ~ singular term K is equivalent to a des
cription auch as "the x tha.t K1s 11 and can be analysed away "!....!! Russell". The extension of Bussell's analysis consista in saying that
all singular terme can be const:rued as descriptions; for Russell, the
class of s1ngula.r terme was greater than that of descriptions. Quine<1s
thesis in contrast presupposes that 1n principle there is no difference
between nam1ng somethi.ng and listing all the predicates and only those
predicates which apply to that thing; that we can secure uniqueness of
reference to, for example, Socra.tes by listing all the predicates whiâh
apply to Socrates and to h1m only. Nevertheless, Quine's conclusion
l. ibid. p. 6. Note that by"context"he means wha.t Russell meant by "context of the proposition" and not what Strawson did by "context of the assertion".
2. ibid. P• 146.
from the possibility of eJ1m1nsting 'ingular terms is that we can
use singular terms significantly in Jentences w1 thout presupposing that i . 1
there are the entities wbich those t~rms purport to name • Thus, "the
only way we can involve ourselves in ontological commitment (is) by our
use of bound variables"2
• This then is Quine 1 s cri teri on of ontological
commi tment. For an example of how i t might work, take "There are grey
elephants". In first order .tunctional calculus, the above statement
would be symbolised as"3 x (E xct-G x)'; the x is bO\md by the existential
quantifier "3 x". Th.us, one would be commi tted to ail and only those
things which x takes on as values auch as to make the sentence true.
The importance of this cri teri on for Quine is illustrated by the
following quotation. 11We now have a more explicit standard whereby to
decide what a given theor,y or form of discourse is committed to: a
theor,y is committed to those and only those entities to which the bound
variables of the theor,y met be capable of referring in order that the
affirmation made in the theor,y be true 11 •3
Before going on to tu.rther elaborate the criterion, I shall add
a few more words about the elimination of singular terms. Quine says:
"Wbatever we say w1 th the help of names can be said in a language that
shuns names altogether. To be assumed as an entity is,purely and simp1y,
to be reckoned as the value of a variable. In terme of the categories
of traditional grammar, this amounts roughly to saying that to be is
to be in the range of reference of a pronOlm.. "4 That we be able, in
principle,to eliminate singular terme is crucial to his general programme,
for if it were true that singular terme are theoretical1y superfluous,
1. ibid. p. 12. 2. ibid. p. 12. 3• ibid. P• 13. F. 4. ibid. p. 13.
it would follow that we could have a language consisting of quantified
statements alone. Renee, the Quiniap ontological criterion would be
sufficient for all cases.
Also the problem of univeraals cauld be solved by Quinian methode;
we would bec-.nitted ouly to classes because:
(a) All names could be construed as predicates (or as clusters of
predicates);
(b) All predicates ~be construed as relations;
(c) All relations can be construed as classes of classes.
As I have said, the thesis about singular terms is crucial to Quine's
contentions concerniDg ontology. A more detailed exandnation of it
will be offered in the next chapter.
However, for a fuller understanding of Quine's criterion of
ontological commitment two qualifications of it must be taken into
account. The two qualifications Quine off ers are the following:
(a) "What there is does not in general depend on one's use of
language, but what one says there is does111 • This tells us what Quine
means by "ontological commi tment of our dis course": we are not commi t;jed
to entities--what exista, but only to saying that there are auch and auch
entities--saying what there is. This leads to the point that, primarily,
ontological commitment applies not to man but to discourse. The way I
talk may commit me to saying what there is but I ~ disregard ~ commit
ment. I maJ take the attitude of frivolit,y.
Another important qualification connected w1 th (a) is
( b) One can free one self from ontological commi tment by al tering
1. ibid. P• 103.
15.
1 one's discourse • It may be that I can rephrase what I want to ~
into an idiom avoidillg quantification, avoidiDg the use of "fhere is."
!his point is important for Quine because whenever I use "!here is" I
automatical.l7 i.D.vo1 Te JÇ'Self in an onto1ogical commi tm.ent but w1 th
this qua.li.fication more flexibility is allowed.
Quine also says that: "Relative to a reall.1' alien language it 'II1Jq
happen despite the most sympathetic effort, that we cannot make even the
2 roughest and remotest sense of ontological commitmsnt" • lfow, this sug-
gests that there may be cases where we cannot decide in an ontological
dispute alODg Quinian methode, but the question is: What would count
as auch an alien language for Quine? J.Prom the context it appears that
he has amo:ag others, the natural language in mind.
".... the idiomatic use of "there is" in ordinary" language knows
no bounds comparable to those that might be reasonably adhered to in
scientific discourse painstakingly for.mulated in quantificational terms•3•
But, would he accept auch an alienation between a quantified language4
and the œtural language?5
In sum, we see two things happeni:Dg: an attempt at a more or lesa
decisive ~ of bandling ontological disputes, and second, an attempt
at reducing the num.ber of entities which we can, or must sa:r there are.
It has already' been pointed out that Quine is attack:l.ng Platonism, and
that his attack is in the spirit of nomina.Jism.
1. ibid. P• 103. 2. ibid. P• 107. 3. ibid. P• 106. 4. A language expressible in quantiticational form. {Quine otten usee
"echematieed language" to mean '!quantified language".) 5. A more detaUed discussion of this question is g1 Ten in JIJ1' Obapter IV.
16.
II
At this point the difficul ty concer.ning the extent of Quine • s
philosophical claims has to be settled. It might be argued that Quine
1s not concerned with metaphysics, but is merelJ attempting to deal
with some philosophical problems of the exact sciences. Therefore, to
say that he fails to solve the general philosophical prob1em of existence
1s beside the point. This argument 1s supported by the fact that Quine
repeatedlJ points out that he is not worried about what there is, bu.t
what one 1s committed to saying there is. But, this distinction 1tse1f
needs to be clarified.
What Quine says 1n the fol1owing passages seems yet stronger support
for this argument. Quantificational form is a good standard for apprais-
ing onto1og1cal cammitment of one or other of our theories because~he
quantificational form is a convenient standard form in which to couch a
1 theor,y." or again:
" •••• i t is only in this spirit, in reference to one or another real
or imagined 1ogica1 schematisation of one or another part of or all of
1. ibid. P• 105.
17.
science, that we can with fUll propriety inquire into ontological pre
suppositions".1
In these passages he seems to suggest tbat he ie not concerned w1 th
the general problem. However, I shall now show that this is by no means
the whole story.
1) Even in the pages where the two previous quotations appear he sqs
more. He is e:r:plicitly arguing against the "champions of ordinar,y l&DgUage"
who say that "there is" certainly belcmgs to ordinar,y la:r:tguage but "look
aekance at a cri teri on of ontological cODDi tment which turne on a real
or imagined translation of statements into quantiticatioD&l for.m". 2
2) In \ford and Object3 Quine argues against saying that there can
be different senses of "there are".
The giet of both 1) and 2) is that Quine 1s not satisfied with eaying
that there is the problem of "there is" in ordinary language for which
there cannot be a forma! solution, and that there is the problem of "there
is" of sche-.tised language which .!!. for.mally soluble. He is sq1ng rather
that we could and should have a general (formal) solution for both.
3) In llethods of Logic Quine sqs: "There are no ultimate philo-
sophical probleu concerning ter.ms and their references, but on1y con-
cerniDg var.iab~es and their T&J.ues; and there are no ultimate pbilosoph:tcal
problams concerning existence except insofar as existence is expressed
b7 the qU&lltitier (3x). n4
4) Hie discussions and examples range over a wide section of ordi
nary la:Dguage. 5
These four points are convincing enough bat there are other grounds
1. ibid. p. 106. 2. ibid. p. 106. 3. Quine (19) P• 241. P. 4. Quine (17) P• 224. 5. ct. Quine (16), p. 2-12., and Quine (18).
on which one might assert that Quine is maldDg a fairly wide philoso
phical clailll. He also says: "Ontologioal questions •••• are on a par
18.
with questions of natural. science.•1 "Aa an em.piricist I contimle to
think of the conceptuel schem.e of scieBCe as a tool, ultilllately, for
predict1Dg future experience in the light of past experience. :P.tv'sioal
objecta are conceptuall;y im.ported into the situation as convenient inter
mediaries •••• comparable, epistemologically, to the goda of Homer.•2
"Posit1Dg does not stop with macroscopic ~sical objecta. Objecta
at the atomic level are posited to make the laws of macroscopic objecta,
and ultimately the lan of experience, simpler and more DIBnageable."'
"!rotal science, mathematical and naturel. and human, is similarly lJut
aore extremel;y underdetermined by experience. !he edge of the system
mu.st be ltept squared with experience; the rest, w:i.th all its elaborate
11\Yths or fictions, has as its objective the sim.plicity of laws.•4
Oonsequently, when Quine sqs: "!he important thing is to muler-
stand our instruments; to keep tab on the diverse presuppositions of
di verse pollrtions of ·our theor:r, and reduce them where we can", 5 he
is to be taken as maldDg a general claim. about ontolog;y, one that bas
and ehould bave importauce over and above mathematics and logic.
!rhis section was not m.eant to turther elabora te Quine' s cri terion
of ontologioal commitm&Dt but rather to decide its seope. It has been
deeided that Quine intends to solve the general philosophical problem
of existence. !lms, in the remaining cbapters I shall treat his claim.
concerning ontology as a general philosophical clâta.
1. QQine (16) P• 45. 2. ibid. P• 44. 3. ibid. P• 44. 4. ibid. P• 45. 5. ibid. P• 117.
19.
ln the immediately following chapter I shall examine two types of
objection to Quine's theory of ontological commitment, objections to
his thesis about the elimination of singular terme, and objections to
his very notion of ontological commitment.
III
SOME OBJOOTIONS
20.
I
In the previous Chapter I bave suggested that the thesis about the
elimination of singular tems is crucial to Quine' s general programme
and have given reasons why it is so. One f'inds a fuller discussion of this
issue in Strawson' s Indi viduals1 and "Singular Tems, Onto1ogy and Iden
tity"2. Existentia1 quantification has "•••• a role, in language which
is to be brought out or elucidated in contrast with the place, or ro1e,
in language which linguistic singular terms have. No sense can be
attached to the idea that they can have the place they have even if there
is no such place", S81S Strawson.3 This statement is a succinct summar,y
of the elaborate po1emic directed against Quine in "Singular Tems of
Ontology and Identity". The outline of Strawson1a argument there is as
follows:
He begins w1 th three quotations from Quine:
1) "The who1e categor.y of singular tems is theoretically supe~
f'luous ••• there are logica1 advantages in thi nkj ng of i t as theoretically
cleared awa:y."4
1. Strawson (24) p. 195. J.F. 2. Strawson ( 26) 7. In Strawson (24) p. 196. 4. Quine (17) P• 211.
21.
2) "Wbatever we say with the help of names ean be said in a laÙguage
which ehuns names altogether."1
;) ".All sillgul.ar terme aside from the variables that serve as
pronolmS in connection wi. th quantifiera, are dispensable and elimina ble
by paraphrase.n2
Then he points out that while these three statements aeem to s~ the
same thing, yet there is a difference. 1) and 2) state the theoretical
possibility of a situation where we use no siDgular ter.ms at all. ;)
states the possibility of using siDgular terms, but also paraphrasing
them into different for.ms where they are not used.
Thus, there are two the ses, one strcmger and one weaker. The weaker
(W) being 'that we can eJim1nate singular ter.ms by paraphrase language',
and the stronger (s) being 'that we can e]1m1nate singular terme by a
language which could stand b.f itself~ W is weaker because to understand
the paraphrase would involve understanding it "as forma of words from
which singular terme have been eliminated by paraphrase."; S on the
other band will not do because there is no guarantee that the description
of the logical character of the "new language" will remain as it was
within the "old la:cguage" witbout the contert of the "old language".
Strawson's strategy is to show that:
a) Quine produces evidence to support W only.
b) ~e advantages Quine claims follow from S only.
Since W follows from S but S does not follow from W the truth of a)
and b) will seriously damage Quine' s position. Renee, what Strawson
seeks to establish is "that it is impossible in principle tbat the
language of the paraphrase (my 1 new language' ) sbould be interpreted
1. Quine (16) p. 1;. 2. ibid. P• 146 ;. Strawson (26) p. 434.
22.
as Quine and the rest of us interpret i t, unl.ess i t is se en as a para
phrase lallgUàge,i.;.e. unless language also contains singular terms."
Strawson is correct in maintaining that Quine needs S because Quine
wants to say more than just that nin the langu,aae of the paraphrase
all reference to objecta is so narrowed down".
Briefly, the argument is as follows:
Quine's predicates or what Strawson calls universal terms "must
be connected with our experience if any are to be understooa.rr1 Further-
more, these universal terms must be connected with particular bits or
slices of our experience. Renee, if they are to be learnt as predicates
of particulars, they must be learnt as predicates of demonstratively
identified particulars. But, no meaning can be attached to the idea of
their being learnt as predieates of demonstratively identified particu-
lars unless the language cont$8 expressions used for maki ng demonstra-
ti vely identitying references to particulars, i.e. unless i t contains
singular terms for particulars. To put this more generally: It is
maintained that a schema could not have the sense it has for us, the
gra.mme.r it has for us, except in the context of that simpler gra.mme.r
2 against the background of which we can learn it.
Strawson considera the possible objection that •no ways of reading
or interpreting the schema are mandatory, that these are the na.tural
ways only for those whose ordinary language does in fact contain singu
lar terms. 13 A similar objection is envisaged in Individuals: "ÂllS.lysis
must !ether be seen as an attempt, hampered b7 the difficulty of getting
away from the forma of ordina.ry speech, ••• We mnst be liberal and ima
gina.tive 1n our interpretation of it."4 However, his answer is that
1. ibid. P• 446. 2. Against the background of natural language contajning singular terms. 3. ibid. p. 448. 4. Strawson (24) P• 19~.
23.
in such a case we Dll18t fortei t our cla.im to re:ter to particulars at
all; we are then not enti tled to sq, because it does not· malœ sense,
that all reference is narrowed dow.n to the values of our bound variables.
T.hat Strawson has given an adequate representation of Quine's point
can be seen from the :to11owing two quotations from Quine: "!be creative
aspect is involved in the progressive re:tinement of scientific language."1
"!be philosopher' s task was wel1 compared by B'eurath· to that of a
mariner who :mu.st rebui1d his ship on the open sea. n2
I could find nothillg beyond these and other similar}J" vague commente
of Quine which could be tak:en as counter-objections to Stn.wson's·.thesis
concerning singular terme.
Thus, i t JliEJ:Y be conc1uded that Quine cazmot ( 1ogical}J") invi;te us
to consider the possibUity of el1m1netiDg singular terme, for our Ullder-
standing of existentiallJ" quantified statements depends on our ability
to refer identif;ying}J" with singular terme. In rq fourth· chapter I
shall :turther rein'! oree Strawson' s cri ticism of Quine on this point,
by pointing out inadequacies in Quine's views on reconstructing
(regenerating) our :natural {ordiœry) language.
1. Quine (16) P• 106. 2. ibid. P• 79.
24.
II
In this section, I am going to consider two objections to Quine's
propoeal of a criterion of ontological oommitment; namely those of
Carnap and Warnock. Although these objecti.one originate from two
widely different philosophical backgrounds, logical posi ti Tism, and
ordinary laJJgU.age analysis, they are s1m1lar in that both question the
very notion of an ontological commi tment.
For Carnap the problem of ontological commi tment is not a theoreti-
cal, but merely a practieal one. His argu.m.ent is as follows:
!here are two types of existenti.al statements:
a) internal existential statements
b) external existential statements
Statements of type a) are formulated within a given la:nguage. i'hey
assert "that there are entities of a specified kind, (and) can be formu-
lated as simple existential state.ments in a language oonta1ning variables
1 for these entities." Examples of these are "there are tables", "there
are numbers greater than 10". Given the langu.age, these are "usual.ly
1. Schilpp (22) p. 871.
25.
analytic aDd tri vial. n1
Statements of type b}: , on the other haDd "... purport to assert
the existence of entities of the kind 1n question not merely within a
given language, but, so to speak before a language bas been constructed"~
These Carnap would cali pseudo-theoretical statements, they should pro
perly be thought of as proposals for the acceptance of language forms.
Examples of these are: "There are objecta", "There are uumbers". How,
for Carnap, since he does not think a)-type atatements are problematic,
the purported problem of ontological commitment ia no more than a dispute
about the acceptance of language forms. He suggeats an example to illus-
trate the point.
1 Sappose1,he says, 1there are two philosophera Xl and x2 discusaing
in the na.tural la:nguage about two construc-.ed object languages LJ. aDd
L2! These two languages are the eame except the univerae of discourae
D:L of L1 (the range of values of the variables) ia more comprehensive
than D2 of ~· Sappose aleo that Dl containe indiv:f.duals, classee of
indi viduals and classes of classes of indi viduala, but D2 does not con
tain classes of classes of indi viduals. Consequently, the 81Dtactical
rules'of transformation and the semantical rules for L- concepts are
auch that:
1) 11li'or some :x and so:rne y, x is an element of y" is provable L-
tru.e 1n Ll•
2) "li'or enry :x and every y, x is not an element of y" is provable
L-true in L2·
How, both Xl aDd X2 agree that L2 ia simpler than LJ., they both under
stand the ayntactical rules for L1 and ~ only x2 does not understand
1. ibid. P• 871. 2~ ibid. p. 871.
26.
the semantical rules of LJ., but this is not crucial. So far there is no
problem. The problem arises when x2 objecta to L1 on the grounds that
he bas arrived at the following two ontological resulta:
3) There are classes of objecta.
4) There are no classes of classes of objecta, and he says that
he arrived at this result because the semantical rules of ~ allow the
phrase "classes of classes", which ha,s no reference.
For Carnap 3) and 4) would be pseudo-statements because they are
misconstrued by x2 as existential statements of the natural language
"absolutely and objective1y"1 rather than (correetly) as statements
2 "relative to this or that la.ngua&e".
Therefore, when Quine says that your language Lx commi ta you to
§ and ~ beeause your semantieal rules allow ~ and SJ , aceording to
Carnap, he is uttering a pseudo-statement. Thus, Car.nap's point is
that ontologieal disputes as Quine conceives them are spurious. They
are idle disputes. Once we get clear on the semantical and syntactical
rules of the language, such disputes need not arise.
Wa.mock1 s treatment of the problem is different. He takes Quine' s
assumption to be that "if an expression has a designa. ti ve use, there is
something which in that use it designates113 and aslœ for the test whereby
one could decide Wbat expressions are to be taken to have designative
uses. 4
One of these tests is existential generalization. From:
1) "Leeds is a City" one can generalize to
la) "Something is a city" or
1. ibid. o. 873. 2. ibid. p. 873. 3. Warnock, "Metapbysics in Logic", in Flew (10) p. 78. 4. ibid. p. 78.
lb) "There is something which is a city" or
le) "There ia an x auch that x is a city"
But, from:
2) "Valhalla is Il\)'"thological11 one cannot generalize to
2a) "There is an x auch that x is !1\)'"thological"
Thus it is argued that the above shows that "Leeds" does, and "Valhalla"
does not have designative use.
'But, this test', says War.nock, 'breaks down in cases of;
3) "Appendicitis is painful"
4) "17 is a prime number", because it ia not clear whether one
would rule out statements of the form:
;a) "There is something which is painful"
;b) "Something ia a prime number" t
Even if we are liberal and say that clearly we can a.dmi t ;a and ;b does
it follow (as it would for Quine) that we are committed to Platonic
entities? In both statementa (;a and 3b) "somethiDg" is a component,
and treating it as logicians do both statements are of the form:
"3 x( •• x ••• )tt fhe point, however, ia that "something" -does not fonction
in ordinary language in the aame way as i t does in logic. Thus, at least
in the case of ordinary language the first test fails to tell whether an
expression is uaed designatively or not. Consequently, as far as ord.i.nar;r
language is concerned, we can say with Warn.ock that the question of onto
logical commitment is not yet answered.
1 The second test Warnock examines is that of ~1inian "application".
(Quine "Notes on Existence and Necessity" Journal of P.bilosop~ 40, 1943)
This is the converse of existential generalization in that here one is to
1. ct. Quine (18), and Quine (17)
• 28.
find a formula wbich is true for all x' s and see whether i t "applies"
to the term in question. If it does, the term is said to have designated
use. For example, if there is a formula, s~ ( ••• x ••• ) which is true of
all x and applies to a term 'a' then 1a 1 bas a designative use. Let the
formula be (3 y) (x)y). If it is true that (x) (;ry) (x;>y), and that
3 7 (x) 7) is "applicable" to the term 1 a 1 ; i.e. if substi tu ting 1 a' for
'x' in)y (x>:r) )"ields a true statement; then 'a' is said to have a
designative use.
The wealmess of this latter test aceording to Warnock, is that we
could find certain formulas wbieh are both true of all x' s and "appl)""
to every possible term,e.g. •(x = x)•.
To sum up Warnock's position so far: Quine is interested in the
ontological commitment of languages. His procedure for deciding about
commitment is to find wbich terme aan be instances of the bound variables
of quantification. (Those which are, are said to have designative use).
But existential quantification is but a deviee of quan~ificational logic
which as auch bas "little or no clear application to the ordinary words
and idiome in which the problems are initiall)" expressed".1
Jinally, Warnock considera the possible eounter-objection tbat his
own objection might after all be ir:relevant. Could Quine agree wi th all
that Warnock bas said, and reply that the question of ontological commit-
ment is concerned not wi th ordinar.r language but wi th "one or another real
or i:magined logieal schematisation of one or another part or all of science 11?2
If what I stq in the previous cbapter is right, such a charge of ir-
relevance bas already been answered. Bevertheless, the fact remains tbat
Quine's criterion of ontological commitment bas no prima facie application
to ordinary language.
1. Flew (10) p. 90. z. Quine (16) p. 106.
29.
~e purpose of this chapter bas been to outline the objections of
Carnap, and the objections of two ordinary language philosophera: Strawson
and Warnock. In the next chapt er I shall reexamine, briefly, Carnp.p 1 s
objections, and, in detail, the ordinar,y language philosophera' objections.
IV
THE COBCEPTU!L PRIORITY OF BJ.TURAL LANGUAGE
30.
The aim of this ehapter is to elueidate :f'urther some of the difti
cul ties one encounters in attempting to relate the '11U'ious aspects of
Quine's philosophy. Thus, I propose to consider how he does or could
reply to Oar.Dap's ani Warnoek's objections.
Very brietly, the ob3ections are as follows:
(1) Oar.Dap:
Ontological questions arise in certain traaeworks;
answers to them mq be found by log.l.cal methode or by empirical
investigations. However, questions about the erlstence of the :tramework
(uumbers, large size objecta) are at beat questions eoncerniDg convenience
of linguis'tic expression. !hus, Quine'e problem about ontology is a
pseudo-theoretical one.
( 2) Yar.nock:
The efforts of symbolic logicians to clarif'y probl8DIS of
ontology :tail. If, however, Quine agreee that S)'Dlbolic logic can make
no important contribution to philosophie problems of ontology, only to
"one or another real or i:magined schematisation of one or another ~
or all of science", the issue cames to an uncontroversial end.
Quine's general defense against (1) is that Oarœp's distinction
between exte:mal and internal enstential statements is a :misconception. It
1
relies on the "analy'tic-syntheti;c" distinction, a distinction which is
unacceptable. 1
1 What needs to be done now i~ to examine Qttine's reasons for the
1
rejection of the "analJtic-synt~tic" distinction, and Carnap's justi
fication of the "inter.nal~exte~" distinction.
The "analy'tic-synthetic" di~tinction is usual.lJ' stated in the
following way: There are two types of statements, thoae which are true
by virtu.e of certain facts about' the world, and those which are tru.e i
by v:l.rtue of their meaning. Thefie two types of atatements are said
to be "synthetic" and "ana:cytic" respective:cy. Quine rejects the
"analytic-synthetic" distinction because, in his opinion, there are
no adequate criteria for analyticity. He considera the following
attempts at giving a criterion:
A) The e:xplanation of analJ'ticity in terme of syno'1J'3"1117. When
I srq 1 "all bachelors are UXIJDal'l"ied men" is analJ'tic' , I mean that
"bachelors" and "unmarried men" are synoil1m.OUS. More generallJ', a
statement is called analy'tic if i t ean be transformed into a truth of
logic by substi tuting synonyme for synonyme. In our example, this
substitution would result in the logical truth "ali bachelors are 1
bachelors" (x) (Bx = Bx). But, Quine's objection is tbat this will
not do because our notion of synOJJ'3flJY is in as much need of clarifi-
cation as that of analytici ty. In fact, most attempts at definillg
syno~ rely on the notion of ~icity.
B) Expl.anation of analyticity in terme of artificial languages.
Bere i t is held that since ord~ lar:lgWlge is vague we must look
at artificial l.anguages1 which ~ve explicit nsemantical rules",
l. Quine would distiilguish between artificial languages and a reconstru.cted language. CF. in my next chapter: Quine • s use of "ordinar;y language".
32.
in order to lllake the distinction olear. Camap (also) holds this new.
l'or him, a atatemtnt S is said to be ~o tor a laDguage L it it is
true aecordillg to the semantical mlea ot L. Bat what, asks Quine, are
"semantical IUles"? Bow are we to explain .!!!!g? Quine in the end say'S,
"Semantioal IUles determining the anal.J'tio atatements ot an art1tic1al
language are ot interest onl.7in so far as we alreaq understand the
notion of analy'tioity; the7 are of no help in gainiDg this UD.derstand:tng."1
lfow it :f'ollows from B that Oamap's distinction between erlernal. and
internal existential statement relies on the "analytic-s,ynthetic"
distinction because the var.y notion of semantical rule relies on that
distinction.
In sum, the force o:f' Quine's counter-objection is that i:f' Carnap
makes the "internal.-erternal" distinction, the onua is upon h1m to
explain what an artiticial laDguage is, wbat a aemautical rule ia.
Since auch an explanation would have to req on the "anal.y1iic-synthetic"
distinction, and since this distinction is untenable, Oamap1s objection
is without weight.
It must be noted bere that the issue about ane.J.ytic and syntbetic
statements ia b7 no meana a cloaed one. Xevertheless, Quine is justi
:f'ied in his counter-objection. C&rnap does have a rather strict re-
quirement :tor what he considera a l.aDgu.age and,!!!. does not meet e.r:fT
ot Quine'a objections to the "ana:cy-tio-aynthetio" distinction, i.e.,
in the W8J' he lllakes the distinction i t is vu.J.nerable to Quine' s attf4ck. -In spi te o:f' al1 this, the di:f':f'erence between Carœ.p 1 s and Quine' s
position is not as great as one might think. !bis is what Quine 88.18
about their di:f':f'erence concer.ning ontology: "llow Carœ.p bas mai:ntained
tbat this is a question not o:f' matters of fact but of cboosing a
1. Quine (16) P• 36.
conTenient language torm, a cOimmient conceptual acb.eme or tramework:
tor science. ti th this I agree, but oJll1' on the pro'Yi.ao that the same be
conceded regarding scientitic b1Potheses generalq. Carnap bas recognized
that he is able to preserve a double standard tor ontologioal questions
and scientitio Jvpotheaes ODly' by assum.ing an absolute distinction between
the analytio and the s;ynthetio; and I need not sa:r aga1n tbat this ia
a distinction which I reject. •1 Or again: "Carnap, lewis, and othera
take a pragœatic stand on the question of chooaiDg between l.a:Dguage
toma, scientitio framework:a; but thair pragmatism leaves ott at the
iJDagined botmdar,y between the a.Da.l.Ttic and the s;rnthetic. I:n repudiatiDg
auch a botœ.da.J:1 I espouse a more thorougbgo:L:ag pragmatism. tt2 !hus, the
main difference betwêen them is that Quine is more sens:! tin to the
subtle difficulties conneoted with the "analytic-s;ynthetic" distinction,
and as he sqa he is a more "thoroughgoiDg pragmatist" • 'lb.ether auch a
"thorough go:L:ag pragma.tiD" is oalled for, whether it is needed, has
still to be decided •
.As far as Wamock:'s objection ia concerned, I have already dismi.ased
the counter-charge of irrelevance as havillg no adequate basie. !hus, I
propose to re-e::œmine the controversy ~ light of what Quine sqs in his
ohapter "Logic and the Reification of Universals": "Ordinary la:Dguage
rema::ins indeed ftmdamental, not only' genetically, but also as a medium for
the ultilllate claritication, by' however elaborate paraphrase, of auch more
artiticial usages." But, preoccupation with ordirlar,y language passes
over the creative aspect of philosopq,.3
While Warnook: would readily agree w1 th the tirst statement he
would have suspicions about the creative aspect ot philosopD1'• J'or,
1. ibid. P• 46. 2. ibid. P• 46. 3. ibid. P• 105.
be could argue, if it is tru.e tbat ordi:nar,y language is :f'nndamental,
then &rJ'3' irlhovations introduced in a. regenerate language have to be
clarified in ol'diDar;y la:nguage. But aillee the problem of ontology, he
could say, arose from ordi.:Da:r7 la:Dguage, and paradoxes seem to arise in
it, the problem should be solved there. .A regenerate, schematiaed,
la:Dguage which ex hvpothesi does J'lOt bave those problems can at best
be a sketcey summa:cy of what bas been achieved in the natural la:Dguage.
But, 1lOW what is this creative aspect of philosoplv'? In "A.1ms
and Claims of Regimentation1", Quine speaks at departures tram ordiDar;,y'
la:nguage. 2 Som.e departures, he says, might later become ol'diDar;y la:nguage,
because the language which is psychologicallJ' JIIOst fundamental is ordi
Dar.Y J.a:nguage; consequent]Jt, azq departure could la ter become a part of
ordinary language.
Herein lies the basie di vergence between Quine and ord.inary laDgUage
philosophera lilœ Strawson or Warnock. Ord.1.na.r.r language philosophera
would insist tbat tbere is no precise definition (no necessary and
sufticient conditions) for what constitutes an ordinary language, but
the7 would maintain that there are certain conceptual criteria enablir.lg
us to decide which language is ordirJar.r and which is regenerate. Quine
on the ether band tak:i.Dg advantage:.'of this vaguenese fills the gap:
the language wbich is most fundamental peyehologically ~ ordinary
language. OrdiDar,r language philosoplllers, in contradistinotion, maintain
that ordina:ry language is conoeptu.al.ly, not just psycholoôcal]Jt, prior.
Qu,ine's thesia about translation 1:las to do with the issue here.
He denies empbatical]Jt that radical translation is possible, tba.t it
is possible ever to give an emet translation of som.ethiilg in a different
1. "Begimentation" means reconstruction in quantificational form. 2. Quine (19) Section 33.
language. !his he calle a "lexicogm:phers1 JQ"th,. !1!be reason he otters
~or rejecti.Dg the possibil1't7 o~ radical translation is that it presup
poses an inadequate theoey of meanil:lg; that it presupposes the Tiew of
mea:nings as som.e ld.nd of ideal enti ties. On that Tiew sentences eould
be radical translations of one another because they would "sba.re" the
same meanillg.
lfithout going into aJl:1 great detail, we cou.ld grant Quine the
impossibility of radical translations WhiCh.follow from the inadequacy
of~ (Platonist) theor.r of meani.Dg. However, we need not agree with
his conclusions. P1rst, we need not agree that none but the ps,rcholo
gical account of meanil'lg is possible. Second, we need not agree that
because under certain theories radical translation is misconceived, we
must, and need rest content with a loose contact with our natural
la:nguage when we depart from i t. lhen we are faced wi th a departure
from ordina.r,y usage, the questions we, must ask are:
Is the departure called for?
What are the "conceptual wheels" turnèd by the departure?
Oould we turn theae same "wheels" in the old, non-regenerate
language?
Blat, most importantly:
Will this departure be consistent w.1. th other things we want to
sq, and can perhaps say onJ.y in the natural language?
!rhese questions are important, and though they seem too vague in the
abstract, in particular contexte a:nswers to them are not cm.:Q' pessible
but are easential. !rheae questions and answers to them. will, in par
ticular contexte, constitute an analy'sis of the meani.Dg of the d.epartures
in question. Quine off ers two just~ications for his departures fJ:o m
ordiJlarT l.a:Dguage in his theo17 of ontological commi tm.entl t
l. ibid.. P• 158.
1) they aJ:e• aida to clarification
2) they simpli:t,y theory
I shall etart with a discussion of 2).
i'he notion of simplicity is probably one of the most problematic.
So ~ different senses of it are invoked by different people in
various contexte tbat a:ay reference to i t must be treated w1 th great
caution. That there is uneasiness about "simplicity" can be seen
from what :Bu.nge says: "The unqualified demand for econollij" in every
respect, or even in some one respect, is definitely incompatible with
a number of important desiderata of theory constru.ctions-such as, e;;g., accuracy, depth, and external consistency-whence simplici ty tout court
should nei ther be regarded as mandatory nor be coun.ted as an inde pendent
cri terion on a par w::L th othe re-let alone above others. The rules of
simplicity fall under the general norm 'Do not hold arbitrar:y (un
grounded)beliefs'"l~.
That there is further uneasiness about Quine's use of "simplioity"
is evident from what Katz says. Katz argues tbat Qu.ine 1s discussion,
2 in Word and Object , of simplioity does not "offer an independent
methodological justification for preferring the simplest b1Potheses",3
or rather the purported justifications he does offer are wrong. Quine' s
passage in question is to Jlij" knowledge his most detailed and serious
account of simplici ty.
To deal adequately w::Lth the problem of simplicity in science is
beyond the ecope of this thesis. !levertheless, we cau say tbat in some
1. Bange, "The Weight of Simplicity in the Construction and Assayin of Scientific Theories", in lester (12) p. 309.
2i. Quine (19) p. 20. 3. Katz (14) p.lll.
sense of 11 simplicity'" a requeat to aim.plif'.y specifie theories TJ_, T2 •••
ete. is legitimate, beeause usually' we ean get clear on what is asked
for in the specifie cases. But, talk of simplicity of our conceptual
scheme as if i t were some kind of a theory, like the one a we are
familiar w.:l.th in science, is surely' not to be taken at face value.
!hus, Quine's offering simplification of theor.y as a justification for
departures from. ordinar.y language as such is highl;r questionable. He
should not invite us to accept that even .2!!!, of . the aim8 of philosoph;y
is simplification of theor.y unless he ean show us:
(1) How it is legitimate to speak of our eoneeptual seheme as
some sort of a theory.
(2) How the notion of simplicity is operative in this purported
theory.
In as mnch as Quine does not deal adequatel;r with these two points,
his u:rmecessary extension of the notion a theory and his reliance on
the notion of simplicity is not justified.
Let me now turn to Quine' s claim that his programme of ontological
commi tment can aid us 11to understandj :ag the referential work of l.a:n8uage" •1
He remarks that in ordinary language we paraphrase certain expressiou
to reaeh the assent of our interlocutor, i.e. my interlocutor ms;r not
assent to, or understand something I sa:y, but if I paraphrase my state
ment he may assent to an~or understand the paraphrase.--This is true.--
He then goes œ: to say that paraphrasing a certain e:xistential statement
into logical symbole can also have the same resulta .-!his is also
true.--However, this does not mean that all paraphrasing into quantifi
cational form w.:l.ll have that result. That sometimes we can aehieve
1. Quine (19) p. 158.
clarity' 'by our use of e:rlstentiall.y quantitied statements in no W8:3'
justifies sa.yiDg that allf8:3'a, and ollly through existential quantiti-
cation cau we clarif",y the notion of e:rlstence, cl.a.ri.f,y the re:terential
work of our language. It is u:a.sound to make a clear-cut distinction
between how we cOJIDIDlllicate, and what the re:terential work of la:nguage
1 is1 as Quine bas done. For how are we to tmderstand the referential
work of language unless we know how to commun:t ca te wi th referring e::z::-
pressions? In the first chapter i t bas been. seen how illportant i t is
not to be "taken in" by plausible so1màing paraphrases, and how impor
tant i t is to keep the conterts of assertions in mind.
Another claim Quine m.akes, which is meant to lend further support
to the theory of ontological comm.itment, is that • •••• To paraphrase a sentence
o:t ordiDar;y language into lO&ical symbols is '9'irtually to paraphrase i t
into a special part, still o:t ordinar,r or semi-ordi.Dary language; tor
the shapes of the indi vidual characters are UDim.portant. So we see that
paraphrasing into logical symbole is a:tter, lllJ. not unl1lœ what we all do
every- dtq in parapbrasiDg sentences to avoid ambigui ty. • 2 :rirst, "semi
ordinary' l.aiiguage" is nothing but a "long-hanà:" for S)'Jllbolic notation,
as for eza:mple: "There is an ::z:: euch that it X's" (semi-ordinary-) is by
defini ti on &PlOJ31DIOUS wi th 9 ::z::( IX)" ( aym.bolic) • The pro blem for us is not
to get :trom sèmi-ord.inary la:Dguage to symbolic expression, but to get to
semi-ord.inary l.aiiguage i tselt. About this Quine sqs:
"But in the pattest o:t pa#aphrasing one courts contusion and
obscurity' imag1n1ng som.e absolute By'D07J.Y'IIII' as goa1.•3
1. ibid. P• 158. 2. ibid. P• 159. 3• ibid. P• 161.
This counter-objection bas already been dealt with. L&ck of
absolute syno~ does not justif,y Quine 1s refusal to accept &QYthing
short of a psychological cr1ter1on for syno~ for meaning.
A corollary to Quine 1s.thesis on syno~ is that the only notion
of conuni tment whi.ch makes · a:ny sense is to be found in elementary logic, in
first order 1'tmctional calculus. This explains his insistence on Ê!!
cr1ter1on of ontological commitment. Since he found a need to solve
ontological disputes and would not reconcile himself to a method lese
rigorous and precise than quantificational logic, his offer for a solu
tion was to be in tems of quantificational logic. In this respect,
Quine was "ru.thlessly" consistent. However, whi.le one could agree that
quantificational schemata, serving as limited modela for some types of
discourse are useful provided the possible or actual uses of that model
are know.1;1, one would maintain that the task of expla:in:lng, and/or under
standing is not through paraphrase into quantification, but rather it
is through descriptions, redescriptions, examinations of various uses,
and contexte, and other similar methode.
Quine 1s worries over ambiguity are, in a:rrt case, needless. There
1s nothing wrong w1 th ambigui ty and impreciseness as auch. Most of our
concepts are and some need to be ambiguous and imprecise. But, then
what is ambiguous outside a specifie context of discourse need not be
so within it. 'That a term is ambigtlous 1 , means that there are several
wa1s one ean understand it, therefore, one need only ask which way it
was meant to be understood. The problem of ambiguity arises when one
looks at terms or statements as eontext invariant "entities", things which
must do the same job in all contexte. Thus in rq next chapter I shall
investigate, in the spirit of the previous paragra.ph, what regimentation
contributes to our understanding of the problems of ontology.
v
THE "E-PROBLEM"
40.
U'p to œw, I have deal t wi th the presuppositions underlyiDg Quine 1 s
thesis about ontologioal. commi tment. In the follo'Wing chapter I ahall
address m;,ysel:r to the specifie problem of what Quinian regi:mentation ca:n
contribute to our u:nderstandi:Dg of E-problems. I use "E-problems" to avoid
biasing the issue wJ. th a loaded expression. It refere to the oluater
of problems falH:ag under these difi'erent heaà1ngs:
a) !he various uses of 11there is", "exista".
b) !he refer:r.lmg work of l.arlguage.
c) Decià.iDg about ontologieal disputes.
d) Genui:D.e or pseudo-ontologioal disputes.
e) The statua of uni versals.
The temptation is tc aq, as Quine does, that ordina:ry use of 11there
is", "erlsts" lmows no bounds, thus we have to restrict it, to make
it more 118ll8.ge&ble. But, we must go eautious~ here. 1fe must first
distiDguish between sqiDg that we do not UDdersta:nd what is meant b7
the expressions "thare is", "exista", and SEqiDg that their use soma
times generates parado::œs. Quine would sq that disagreementa and
paradoxes are generated 'because we have a vague and imprecise notion
ot ex::Lstence. Renee, tor hia the solution of E-problems cannot be
piecem.eal, va.ry1Dg :trom. contexte to contexte; he requires a general
decision-procedure.
"•••·" now have a more ex.plicit standard whereby" to decide what
ontology a given theor,r or for.m of discoures is committed to."1
41.
But, as I han indicated in..,- preTious chapter, treedoa from
'9118Ueness, and treedom from paradoxes are nei thar sufficient nor neces
eary conditions for freedom from contusion.
!hat Quine ahould come to auch a strict requirement for sol"fi.Dg
E-problems can also be seen from the first question he aslœ about the
problem.: "What is there?" !his is indeed a simple sounding question,
but one which upon scro.till1' becomes incree.si»gly abstruse. It sounds
lik:e "What is on tables?" "What is on one's mind?" ''lhat is the sum
ot 2 and 2?". Y et i t demanda an answer that could satisf7 81l'3' one of
these. !his wtq. of approa.ching the E-problem. prejudices the case from
the outset. It presupposes that there is a general, tormal problem
which needs solving.
Quine's anti-Platonism., in the end, comes to this: Plato has
asked "What is real?" or "What really ex::Lsts?", and his answer •s
that it is the one in the man11 he was led to this conclusion b;r con
centrating on the indiTidual terme of discourse, not realizing that
whatever we sq fapes the tribunal of experience together wi th our
whole conceptua.l schem.e. But depending on our conceptual scheme what
is real may vary.
Quine might be right in his diagn.osis that the source of Pl.ato' s
fallac;r was over-emphasis ot the sim:Uari ti es in grammatical fôr.m.
1. Quine (16) p. 13.
42 ••
Bowever, he does not, in that case, go far enough. He is atill
lookiDg for a general solution to !BE problem of existence. Renee, the
question: "What is there?" Now, &l'JI. auswer to any of the three earlier
question is s, P or R whateTer s, P. or R might be. Yet the auswer to
"lhat ia there?" is drastically d.ifterent. It can haTe only, one possible
answer.
One can uaderstand the question 11What is there, tor theor,r T?tt or
11\Vhat is the ontological comœitment of theor.r T?"• And, the auswer that
"theor,y T is committed to those and on1y those entities which are the
values of bound 'VIU'iables of theor,y T." makes sense. It JD&kes sense,
bu.."f; i t presupposes that theo17 T is expressible in qœntif'ioational
tom.. In reali v, there are Ter,y few, if aJV' auch theories. !J!hus
even the ontological problem of theories is mch more tho1'1'q' than
Quine seems to suggest. But, when we come to E-problems, tb.i..llgs be
come m:u.ch more difticul t.
In the first place, i t is di::f'ficul t to see how Quine can apealt of
the ontoloQ of genèral discourse, how he can sq that when we come to
THE problem of existence we should look upon our conceptual scheme as
some sort of a theor.r, and on our l.anguage as some sort of unifiable, if
not unified, discourse. !he truth is that E-problems are not in all1'
important sense like the purported ontological problems of the sciences.
Our particular sctentitic theories have certain more or lesa specifie
jobs to do. The deployment of modela tor certain scientific theories,
which !! an essential scientitic-ontological task is gtlided by JDalJ1
specifie considerations, arisiDg in specifie conterts of inquir.y. These
tasks are not sbared with our discourse at large.
"!rhere is" &l'll.d "exista" give rise to disagreements and paradoxes
in ordinary di~ourse because the1 are put to l1l.l.lV' and varied uses.
Of course, schematisation will help in some cases, but 1 ~ is only one
(by no means the most important) of many considerations. It is not
t:rue that whenever E-problems arise they are resolvable by the Quinian
criterion; Warnock1s "Are there Appendicites?" and "Are there m;ythical
mountains?" are two e:x:a:mples of that. '
Thus Qu.ine 1s dictum that:
"To decline to explain oneselt in tems of quantification, or in
tems of those speèial idiome of ordinary language by which quantifica
tion is directly explained, is simply to decline to disclose one's
referential intent"1is too strong. E-problems, I• w:lsh to maintain,
are generated by various contexte of inquiry', especially when two cate
gories of dis course, auch as the DJTthical and factual, the formal and
material, are at play in those contexte. But, even then the E-problems
are only secondar;y. The prima:ry pro blems are adequate understanding
and adequacy of descriptions.
Suppose A describes a certain state of affaira and B finds the
description inadequate. Then B may invite A to take another look or
to tr;y to redescribe the given state of affaira. In the end, A and B
:may still come up w:l th two al tema te descriptions. Suppose also that
the two descriptions differ, as Quine would say, in their referential
intent. Bow could A and B understand and decide about this difference
in intent? They would not ask one another for schematisation but
rather ask what turther :moves they would be allowed to make, or pre-
vented from making with their respective descriptions. In other words,
they would have to explore in piecemeal fashion the consequences and
ramifications of the â.l tema te descriptions. Their exploration might
have one of two re sul ta:
1) the referential problem has been dissolved with the mutual
1. Quine (19) p. 243.
44.
acceptance of one description
2) the difference in referential intent hae been underetood and
nei ther of them would, for some understandable, particular reason change
his description.
The second alternative is a possibility, but even then schematisa-
tion would not help because the divergence would be due to their dif-
ferent treatment of the particular situation.
:N'ow, Quine would say that they could have gone on and on, until
eventually they would have agreed on a certain schematisation. But,
this is what I find unacceptable, for it
a) places too much emphasis on the referential problem
b) puts too much faith in schematisation.
But more seriously,
c) it assumes the possibility of tully resolving all ontological
(referential) disagreements.
We could say more or lesa clearly what outeide considerations are
relevant to our particular problem, in the particular univeree of die-
course. But, to suggest that there is ultimately only one univeree
of discourse, a quantifiable one, where ontological probleme can be
resolved, seems at beat gratuitous.
I have earlier suggested that Quine 1s treatment of the E-problem
ie motivated by his dissatisfaction with the Platonic view of universals;
in accordance w1. th the law of parsimoDY he attempts to reduce the number
of universals. In this sense, he is attacking the Platonic theory from.
1 wi thin. However, in recent philosoph;lcal discussions , the concern
over uni versals has diminished, not tJ:!,rough complacency, but be cause
1. Stra.wson, "On ReferriDg" in Caton (6); Scriven, "Definitions, Explanations, and Theories" in :Peigl (8); Wittgenstein (27)
the new ways of looking at our concepts, our language, is thought to
have undercat the Platonic view of universals. Quine, if m:r arguments
in these last two chapters are valid, failed to show us how what he
talees to be the ontological problem is a problem, and furthermore, if
he meant to contribute to the solution of E-pro blems how his theory
of ontologioal oommitment oontributes to such a solution.
CONCLUSION
!he conclusion reached in this s'tudy 1s that Quüae's attempt to
apply techniques of s.ymbolic logic bas~ at best, ~ a limited ap-
plication to the problems of existence. His criterion of ontologi
cal commitaent, relyillg on the notion of the e:d.stential quantifier
of s.ymbolic logie, is mch too restricti n to be of use in sol Ting
the philosophie problem. J'urthem.ore, his reliance on syabolic
logic is misconceived, tor an adequate discussion of existence mat
include a cri tical e:œmtnation of the logicians' notion of ex:Lsten
tial qualification. It ought not start wi th the assu.m.ption that
their notion is the correct one.
46.
BIBLIOGRAPHY
47.
BIBLIOORAPHY
Only those works which have been of use to me in the present thesis, are mentioned here.
{1) AYER, A. J.
(2) AYER, A. J. (ed.)
(3) BEN'ACER.U', P. and PUTNAM, H. ( eds.)
(4) BLACK, M.
(5) BUNGE, M.
{6) CATON, c. E.(ed.)
(7) CHURCH, A.
(8) PEŒGL, H. etc., (eds.)
(9) l!'EIGL, H. and SELLARS, w. (eds.)
(10) :FLEW, A. (ed.)
(11) :FLEW, A. (ed.)
(12) FOSTER, M. H. and MARTIN, M. L. (eds.)
: The Concept of a Person,_ London, MacMillan and Co. Ltd., 1963.
: The Revolution in Philosophy, London, MacMillan and Co. Ltd., 1963;.
: Philosopbx of Mathematics, New Jersey, Prentice-Hall Inc., 1964.
: Problems of Anal;rsis, London, Routledge and Kegan Paul, 1954.
: The Mfth of Simplicity, New Jersey, Prentice-Hall Inc., 1963.•
: Philosoph;y and Ordinary Language, Urbana, University of Illinois Press, 1963.
: 110ntological Commitment". Journal of PhilosopSr, Vol. 55, 1958.
: Minnesota Studies in the Philosopbl of Science, Vol. II, Minneapolis:, University of Minesota Press, 1958.
: Readi.:n.gs in P.bilosophical Anal.ysis, New York, Appleton-Centucy-Crofts, Inc., 1949.
: Essaye in Conceptual Apal.:rsis, London, JlacMillan and Co. Ltd., 1956.
: Logic and Language, Pirst Series, Oxford, Blackwell, 1960.
: Probabili ty, Confirmation and Simplici t:r, New York, The Odyssey Press, Inc., 1966.
48.
(13) HARRE, R. : Theories and !J!bings, London, Sheed and Ward, 1961.
(14) KATZ, J. J. • The Problem of Induction and its Solution, • Chicago, The l.TniTersity' of Chicago Press, 1962.
(15) LIISKY, L. (ed.) . Semantics and the Philosophy of Language, • Urbana, University of Illinois Press, 1952.
(16) QUillE, w. v. • l'rom a Logical Point of Viewi New York, • Harper and Row, Inc., 196~.
(17) QUINE, W. V. . Methods of Logic, New York, Holt, Retnehart • and Winston, Inc., 1964.
(18) QUINE, W. V. "Notes on Existence and Necessity", Journal of Pbilosopgr, Vol. 40, 1943.
(19) QUINE, W. V. : Word pd Objec!, Cambridge, The M.I.T. Press, 1960.
(20) RUSSELL, B. • Intro!uction to Mathe:mat.i_ç,!Ü Philoeop,&, . IDndon, George Allen and Unwin Ltd .. , 1953.
(21) RUSSELL, B. • "Mr. Strawson On Referr.l.l:lg", :Mind, Vol. LXVI, • No. 263, 1957.
(22) SCIILPP, P. A. (ed.) . !he Pbilosoph,y of Rudolf Oama;2, La Salle, • Open Court, 1963 •.
(23) STRAWSON, P. J!l. • "A Logician's Landscape", Philoso'Dhv, • Vol. 30, 1955.
(24) STRA.WSON, P. J!l. . Indi'Vidue.ls, London, Methuen, 1964. •
(25) STRA.'ISON, P. J!l. . ~tro!!!ction to Losical Theo~, London, . Jlethuen, 1963.
(26) STR!WSOllt P. 1'. : "SiDgular Tems, Ontology and Identi ty", Bïnd, Vol. LXV, No. 260, 1956.
(27) WIT!GE!fSTEill, L. • P.bilospEhical Investii!tions, Oxford, • Blaeltwell, 1963.