Parsons (1982). Are There Nonexistent Objects
-
Upload
tomas-castagnino -
Category
Documents
-
view
258 -
download
0
Transcript of Parsons (1982). Are There Nonexistent Objects
-
8/10/2019 Parsons (1982). Are There Nonexistent Objects
1/8
-
8/10/2019 Parsons (1982). Are There Nonexistent Objects
2/8
American Philosophical Quarterly Paper presented at the 56th Annual Meeting of the
Volume 19,Number 4, October 1982 American Philosophical Association, Pacific Division,Sacramento, California, March 25-27, 1982
VIII. ARE THERE NONEXISTENT OBJECTS?TERENCE PARSONS
THEquestion that I have been asked to address
is "Are there nonexistent objects?"1 The
answer that I will give is "Yes." I cannot hope to
establish with any certainty that this is the correct
answer. However, I will try to say what the best
evidence is in its favor, and I will try to answer the
chief objections to this view that I have not
discussed elsewhere.2
The evidence in favor of the view that there are
things that do not exist is that this is entailed or
presupposed by things that we believe and that wehave good evidence for. I will give examplesbelow. The two main sorts of objection that I will
discuss here are (1) objections to the effect that the
things that we believe do not in fact entail or
presuppose that there are nonexistents, and (2) ob?
jections to the effect that, whatever the evidence,the conclusion that there are nonexistents could
not possibly be true. Iwill begin with an objectionof the latter sort.
I. Does "There Are" Mean "There Exists?"
One of the most typical respnses to the claim:
"There are things that do not exist" is that this can?
not possibly be true; it cannot possibly be true
because it is inconsistent, and it is inconsistent
because "there are" just means "there exist." Andso the claim: "There are things that do not exist"
just means "There exist things that do not exist,"and this is clearly inconsistent.
Now Iacknowledge
that when we make state?
ments using the words "there are" we very often
communicate the very same information that we
would otherwise communicate by using the words
"there exist." But this is not because the two
phrases are synonymous, and it does not show
that there could not be things that do not exist. In
fact, a proper understanding of our use of
language suggests just the opposite. The first sec
tion of this paper is devoted to an explanation of
this point.I want to begin by examining a typical dialogue
of the sort that makes it appear that "there are"
and "there exist" are synonymous. Here it is:
Speaker A: There is a passage in one of Meinong'sworks in which he gives up the belief in nonexistent
objects.Speaker B: Oh no, that seriously undermines a paperthat Fm writing.Speaker A: Oh I wouldn't worry if I were you; it's
only a nonexistent passage.3
This is a bad joke, but it illustrates a good point.The joke only works because in his first statement
Speaker A communicates the information that
there is an existent passage of a certain sort in
Meinong's writings, and he does this not by saying"there exists" but merely by saying "there is." And
this is typical, I think, of virtually all of the quan?tifier words that we use in ordinary language. By
using "some F," or "an F," or "at least one F" we
usually communicate "some existing F," "an ex?
isting F," or "at least one existing F." And onenatural explanation of this is that the word"exists" is logically redundant, so that these pairsof phrases are all synonyms. However, that is not
the only possible explanation, and it is a wrongone.
Let me apparently change the subject for a mo?
ment and focus on a couple of other dialogues:
C: Did you put out a fork with each plate?D: Each plate? Heavens no There must be millionsof plates just here in California alone. I'm not Super?
man, you know.E: At least two heifers and three steers have the initial
symptoms of hoof and mouth disease.F: Well, have you called the vet?E: What for? Ours are OK. I was talking about onesin China.
I think that these dialogues illustrate the very same
linguistic phenomenon that is operative in the firstone. Speaker C clearly communicates to D a ques
365
This content downloaded from 205 .133.226.104 on Tue, 6 Aug 2 013 22:42:07 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/10/2019 Parsons (1982). Are There Nonexistent Objects
3/8
366 AMERICAN PHILOSOPHICAL QUARTERLY
tion regarding the plates that are then in use bythem: by using the word "plate" he "meant" "platein use by us," and D's pretense that he
misunderstood this is only a bad joke. But this is
not because thephrases "plate"
and"plate
in use
by us" are synonymous; they are not. In fact, if
they were synonymous, D could not make the jokeat all. Likewise E's use of the phrase "two heifers"
communicates the notion "two of our heifers," but
not because the phrases are synonymous. The ex?
planation is rather this. In ordinary linguistic
interchanges a limited range of things are under
discussion, and both speaker and hearer take the
common nouns that they use to be implicitlyrestricted to these things.4 Without such an
assumptioncommunication would be
extremelydifficult, if not impossible.
Now what about "there are" and "there exist"?
Do these often communicate the same informa?
tion because they are synonymous, or because
they are often used in contexts in which it is
understood that it is only existing things that are
under discussion? The answer will be clear if these
phrases function differently in contexts in which
no such assumption is made. Here are three
dialogues that presuppose such contexts:
G: Oneof
the detectivesin this
story (namely,Sherlock Holmes) has inspired more amateur
criminology than any real detective has.
H: You must be pulling my leg. None of the detec?tives in that story exist, and so there aren't any such
things. So none of "them" could possibly have in?
spired anything.
I: It's very strange; that unicorn that I dreamed aboutlast night bore a very strong resemblance to my
psychiatrist.
J: I guess you need more than a psychiatrist; don't
you know that there aren't any unicorns?
K: Meinong thought that some things don't exist.L: How peculiar; how could anyone think that some
things that exist don't exist. Are you sure that the
poor fellow hasn't been mistranslated?5
Each of these perfectly awful jokes depends on
supposing that in the context in question a word
such as "some" means "some existing." If this
equivalence were correct, the dialogues would not
be jokes at all, but would rather contain perfectly
appropriate responses to peculiar statements. In?
stead, it is the responses that are peculiar and the
initial statements that are appropriate. Not only is
the equivalence missing in such contexts, but or?
dinary people withouta
modern philosophicaleducation don't even get the joke when shown
such dialogues.I take such dialogues to show that there are at
least some perfectly ordinary everyday contexts in
which speakers assume that at least some of the
things under discussion do not exist. This does not
by itslf establish that there are things that do not
exist, but it does show that "there are" and similar
phrases are not synonymous with phrases such as
"there exist," and it shows that the claim "There
arethings that
do not exist" is a substantive claim
that cannot be automatically dismissed on groundof verbal inconsistency.
II. When Does Our Use of a Term Commit
Us to a Referent?
I believe that there is an important difference
between (i) referring to a thing that does not exist,and (ii) failing altogether to refer. We succeed in
referringto a
thingthat does not exist when we use
the name "Sherlock Holmes" or the description"the chief detective in the Conan Doyle novels."
We fail to refer at all when we use a descriptionsuch as "the chief detective in Alice in Wonder?
land.99
Free logic is sometimes explained as the study of
"singular terms which fail to refer to existing
things." If I am right, free logic is a study of two
quite different kinds of singular terms. Some
terms which fail to refer to existing things do so
because they fail to refer altogether; examples are
"the existing present king of France" and "the
chief detective in Alice in Wonderland.99 Other
terms fail to refer to existing things because theysucceed in referring to things that do not exist; ex?
amples are "Sherlock Holmes" and "the chief
detective of The Woman in Scarlet.99 These terms
behave in different ways. Those that succeed in
referring to nonexistent things may appear in con
This content downloaded from 205 .133.226.104 on Tue, 6 Aug 2 013 22:42:07 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/10/2019 Parsons (1982). Are There Nonexistent Objects
4/8
ARE THERE NONEXISTENT OBJECTS? 367
tingently true identities (e.g., "Zeus=
Jupiter,""Sherlock Holmes = the chief detective of the
Conan Doyle novels"), they may appear in true
atomic (nonnegative) sentences (e.g., "Pegasus is
winged"),and
theyare
subjectto normal
logicallaws such as universal specification (e.g., from
"All fictional detectives are cultural artifacts" and
"Sherlock Holmes is a fictional detective" we mayinfer "Sherlock Holmes is a cultural artifact").
Terms that fail entirely to refer to none of the
above. I suspect that many (though certainly not
all) of the ongoing controversies in free logic are
ongoing because of a failure to appreciate this
distinction.
So far I have been making claims without sup?
porting them. How can we tell which singular
terms refer and which do not? I think that this isbest divided into two questions: (1) How can we
tell which of the claims that we make using
singular terms are true?, and (2) How can we tell
which of the true claims that we make using
singular terms commit us to there being referents
for such terms? The first question is a very generalone; I will return to it in the last section. In this
section and the next I will discuss the question of
which uses of which singular terms commit us to
those terms having referents. To give you an idea
of where I am heading, my intuitions on this mat?ter are that:
(1) If I say "Sherlock Holmes is a fictional detective"then I am commited to their being a referent for"Sherlock Homes."
(2) If I say "The detective in that story does not exist"then I may or may not be committed to a referent for
"the detective in that story." I might be attributingnonexistence to a certain fictional detective, in whichcase I would be so committed; on the other hand I
might instead be denying the proposition that suchand such a detective does exist, in which case Iwould
not be so committed. (In short, I agree with Russellthat such sentences have the two interpretations thathe would attribute to the description's having either
primary or secondary occurrence.)
(3) If I say "The average American male has 1.37children" then I am not committed to a referent for"the average American male."
One reason for dwelling on such examples is that
historically a belief in nonexistent objects has been
linked with the view that "every grammaticallycorrect denoting phrase stands for an object,"
where denoting phrases are identified not by their
use or meaning but solely in terms of their gram?matical form.6 Such a view
puts"the
averageman"
on a par with both "Sherlock Holmes" and "the
detective in Alice in Wonderland,99 and fails to do
justice to scope ambiguities and the like. It also
leads to inconsistency, as Russell pointed out.7 We
need a more sophisticated view than this, and one
that hopefully can be articulated and tested in a
way that appeals to ordinary nonphilosophicalevidence.
My proposal is going to sound disappointinglyold-fashioned. I propose that a use of a sentence
involvinga
grammaticallycorrect
singularterm
commits the user to a referent for that term just in
case it commits the user to the existential
generalization of the original claim. That is, a use
of a sentence of the form "...i..." commits the user
to a referent for 7' just in case it commits its user
to a sentence of the form "There is something such
that ...it..." Now tests of this sort are often taken
to be appropriate if applied to sentences that have
already been paraphrased into some kind of
canonical idiom. This is not what I am suggesting.I am suggesting that the test be applied to real live
uses of English sentences. The advantage of this is
that the test depends mostly on what people mean
when they speak in ordinary language about or?
dinary matters, and is relatively independent of
philosophical presuppositions. The disadvantageis that the sentences in question are often
somewhat stilted, and this affects our judgment. (Iinsist on the test being exactly applied because
seemingly trivial rewordings can alter our
judgments.) Here are what I take to be results of
the test:
(1) A normal use of "My birthday present arrived
yesterday" does commit one to "There is something
such that it arrived yesterday."(2) A normal use of "The average American male has1.37 children" does not commit one to "Something issuch that it has 1.37 children."
(3) A normal use of "The king of France is not bald"
may or may not commit one to "Something is suchthat it is not bald" depending on how it is intended: as
This content downloaded from 205 .133.226.104 on Tue, 6 Aug 2 013 22:42:07 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/10/2019 Parsons (1982). Are There Nonexistent Objects
5/8
-
8/10/2019 Parsons (1982). Are There Nonexistent Objects
6/8
ARE THERE NONEXISTENT OBJECTS? 369
thickness" are not merely false but, on analysis, in
contradiction with something in that from which theyseemed to be logical consequences. The only solutionis to see that being a fictitious person is not to be a
person of a certain sort...10
Has Ryle said anything here that would cast doubt
on the view that "Mr. Pickwick" succeeds in refer?
ring to a nonexistent thing? That may depend on
additional assumptions about what fictional ob?
jects are like. The most natural view is that such
things are pretty much the way we conceive them
to be when we read the stories in which they
originate, so that Mr. Pickwick would actually be
a (nonexistent) person who actually took part in
certain escapades. Another possible view,
however, is that Mr. Pickwick is not a person atall, but ismerely a thing that has personhood attri?
buted to it in certain stories.11 On this latter view,
Ryle has said nothing at all that would cast doubt
on there being aMr. Pickwick. For this latter view
agrees with Ryle in holding that fictional personsare not persons of a certain sort. It disagrees with
Ryle in holding that there are fictional persons,but Ryle's argument has not touched this point.
Probably Ryle intended to attack what I have
called the more natural view, the view according
to which fictional objects actually have theordinary sorts of properties that they are
understood to have in their stories.12 But even here
I think his argument fails. Granted, on this view
Mr. Pickwick actually has the property of being a
person, as Ryle suggests, and even has the pro?
perty of "having been born sometime," for these
are both properties that he is understood to have
in the story. But Ryle is wrong to think that it
follows from this that there is some particualr time
such that Mr. Pickwick has the property of havingbeen born then. This does not follow on the view
in question, for the story does not specify a time at
which Mr. Pickwick was born, nor does it followon general logical principles. The form of the in?
ference in question is this:
From: x has the property of there being a time atwhich he was born,Infer: there is a time, t, such that x has the propertyof having been born at t.
This inference depends on a property abstraction
principle that is known to fail when nonexten
sional contexts are concerned, and according to
the view under discussion, it also fails in manycases
involving quantificationover nonexistent
things. Ryle has taken a principle of limited val?
idity and applied it to a domain in which it does
not hold.13
Let me turn briefly to Ryle's second test for
systematically misleading expressions. This is not
a test that he ever articulates, but it is one that he
uses quite a bit. Here is a passage in which he
applies the test to the phrase "the idea of taking a
holiday":
.. .the statement "the idea of taking a holiday has
just occured to me" seems grammatically to be
analogous to "that dog has just bitten me."...But the
appearance is a delusion. For while I could not restate
my complaint against the dog in any sentence not con?
taining a descriptive phrase referring to it, I can easilydo so with the statement about "the idea of taking a
holiday," e.g., in the statement "I have just been
thinking that Imight take a holiday."14
The claim seems to be that if a sentence containg a
definite description can be paraphrased by one
without a corresponding name or definite descrip?
tion, then we should not treat the original definite
descriptionas a
referring expression.Unfortunately, when Ryle attempts to apply
this test to the sentence "Mr. Pickwick is a fiction"
the test fails. Ryle attempts two different
paraphrases. The first one is "some subject of
attributes has the attributes of being called
Dickens and being a coiner of false propositionsand pseudo-proper names."15 This clearly fails as a
paraphrase since it fails to specify which of
Dickens' characters is being identified as fictional.
The second attempted paraphrase is "some subjectof attributes has the attribute of being a book or a
sentence which could only be true or false (/"some?one was called 'Mr. Pickwick'."16 This iswrong for
many reasons, a chief one being that it does not
make any link with fiction. Ryle's paraphrases do
not work, and no one has been able to improve on
them in the last fifty years. Perhaps that is because
"Mr. Pickwick" is not systematically misleading at
all.
This content downloaded from 205 .133.226.104 on Tue, 6 Aug 2 013 22:42:07 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/10/2019 Parsons (1982). Are There Nonexistent Objects
7/8
370 AMERICAN PHILOSOPHICAL QUARTERLY
IV. Are There Really Nonexistent Things?
No doubt many philosophers will agree with
almost everything I have said so far, insofar as it is
taken as a contribution to "descriptive meta?
physics." They will suggest that I have been
describing the everyday beliefs of everyday peoplewithout philosophical training, and pointing out
that these beliefs involve a commitment to nonex?
istent things. But I have failed to challenge this
naive point of view and address the real question,which is whether there actually are any nonexis?
tent things.
Probably very little of substance can be said at
this point. Bynow we
have learned that if webegin by rejecting all of our ordinary beliefs, then
no conclusions follow at all, whereas if we reject
only part of them, then our conclusions are deter?
mined by the part we have arbitrarily chosen to
keep. Because of this, Carnap thought that to posea question of the form "Are there really things of
such and such a sort" is to pose a pseudo-question.It is merely to express the fact that we are faced
with a choice, a choice of whether or not to adopta language that uses quantifier idioms in connec?
tionwith
common nounsthat describe things of
the sort in question.17 If Carnap is right, the only
thing that I can reasonably say is that many
philosophers are wrong in thinking they have goodreasons for rejecting a form of language that other
people use with comfort.
Quine thinks that more can be said on such
issues. He thinks that we can make our pragmaticchoices rationally, if we take into account notions
such as simplicity. Here is a passage in which he
explains why we should believe in physical objects:
Imagine... that we have devised the mosteconomical set of concepts adequate to the play-by
play reporting of immediate exerience. The entitiesunder this scheme?the values of bound variables?
are, let us suppose, individual subjective events of
sensation or reflection. We should still find, no
doubt, that a physicalistic conceptual scheme, pur?
porting to talk about external objects, offers great ad?
vantages in simplifying our over-all reports. By bring?ing together scattered sense events and treating themas perceptions of one object, we reduce the complexi?
ty of our stream of experience to a manageable con?
ceptual simplicity.18
The claim here is that it is legitimate to assume
that there are physical objects because, amongother things, this shared assumption greatly
simplifies reports of our experiences. I claim that
it is legitimate to assume that there are nonexistent
objects for exactly the same reason?it simplifiesreports of our experiences in exactly the same way.
Let me give an example. Suppose I tell you the
following:
I've dreamed about the same unicorn three nights ina row. It looks a lot like a dog I once owned, thoughthe way it talks reminds me of my grandmother. Ac?
tually I'm growing quite fond of it, and I'm hopingthat it will be back tonight.
This could be a serious, informative report of a
meaningful experience. The shared assumptionthat there is something that I have been dreamingabout tells you as much about my dream exper?ience as the assumption that there are physical
objects does when I report my waking experiences.It would be just as difficult to translate my dream
report into purely phenomenological terms as it
would be to translate perceptual reports of
physical objects into such terms.
Quine thinks that our belief in physical objectsis a convenient myth.19 Perhaps nonexistent
objects are no better off. But they are no worse
off.
University of Calif ornia, Irvine Received February 3,1982
NOTES
1. This paper is to be presented in a symposium of the same title at the Pacific Division APA meetings inMarch 1982. The other
participants will be Alan Code, Jaakko Hintikka and Howard Wettstein.
This content downloaded from 205 .133.226.104 on Tue, 6 Aug 2 013 22:42:07 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/10/2019 Parsons (1982). Are There Nonexistent Objects
8/8
ARE THERE NONEXISTENT OBJECTS? 371
2. Especially in "The Methodology of Nonexistence," Journal of Philosophy LXXVI (1979) 649-62; Nonexistent Objects (NewHaven: Yale University Press, 1980); "Fregean Theories of Fictional Objects." Topoi, forthcoming.
3. This example is patterned after one by Peter van Inwagen in a talk given to the American Society of Aesthetics in October 1981.
4. In logic such ranges are often called "domains of discourse," and they are taken to be limitations on the range of one's quan?
tifiers. In a language like English in which quantifier words always appear in conjunction with common nouns it makes no dif?
ference whether they are taken to limit the quantifiers or the nouns.
5. This example is patterned after one in Richard Routley, Exploring Meinong's Jungle and Beyond (Research School of Social
Sciences, Australian National University, Canberra, 1980), p. 423.
6. B. Russell, "On Denoting," Mind, vol. 14 1905.
7. The arguments that Russell explicitly states do not quite show this, but simple reformulations of them do; see Parsons (1980),
Chap. 2.
8. Examples (4) and (5) illustrate Frege's point that our commitment to numbers arises from our commitment to arithmetic, and
not from our ordinary uses of number words; see G. Frege, Foundations of Arithmetic (Oxford: Blackwells, 1950).
9. G. Ryle, "Systematically Misleading Expressions," Proceedings of the Aristotelian Society, 1931-32; reprinted in M. Weitz
(ed), Twentienth-Century Philosophy: The Analytic Tradition (New York: Free Press, 1966). Page references are to the reprinted
version.
10. Ryle, op. cit., p. 202.
11. The two views that are being contrasted here are called "Mod 1" and "Mod 2" in Parsons (1980) chapter 7. For a version of the
latter view, but one that takes fictional objects to actually exist, see P. van Inwagen, "Creatures of Fiction," American
Philosophical Quarterly vol. 14 (1977).
12. This view depends on distinguishing "ordinary" properties from others; this is discussed throughout Parsons (1980) where or?
dinary properties are called nuclear properties.
13. Some have held that the inference even fails in physics: a particle may be located someplace in a given region without there be?
ing any particular place at which it is located. The inference also fails for claims about the future, given an Aristotelian view of
future contingents. Much more needs to be said then this, of course; see Parsons (1980), especially chapters 4 and 7 for discussion.
In case the other symposiasts wish to criticize this account, I hereby incorporate the relevant portions of that text into this paper.
14. Ryle, op. cit., p. 198.
15. Ryle, op. cit., p. 189.
16. Ibid.
17. R. Carnap, "Empiricism, Semantics and Ontology," inMeaning and Necessity (Chicago: U. of Chicago Press, 1956).
18. W. V. Quine, "On What There Is," in From a Logical Point of View (New York: Harper & Row, 1961), p. 17.
19. Quine, op. cit., pp. 18-19.
This content downloaded from 205 .133.226.104 on Tue, 6 Aug 2 013 22:42:07 PM
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp