Beruflich Dokumente
Kultur Dokumente
By Nathaniel J. Kan
In "On What There Is," W. V. Quine attempts to develop a method for evaluating the
value of an ontology. Ultimately, Quine decides that any ontology is sound so long as it obeys
certain rules: most importantly, accepts the existence of objects in the domain over which bound
variables in our statements quantify. I believe that this description of an ontology does not go far
enough, and will attempt to show that modifications to Quine's theory would improve it to better
Quine's theory relies primarily on Russell's Theory of Descriptions (RTD). RTD (or at
least, Quine's use of it) is a method of description quite similar to modern logic, which explains
objects by quantifying over a bound variable. For example, 'Fred is wearing a hat' is translated by
RTD into 'there is something that is Fred and is wearing a hat'. More generally, if there is a
specific object c to which the predicate P applies, letting C be the predicate which expresses the
set of properties that describe being c, then 'Pc' translates into 'There exists x such that Px and
Cx'.
And so Quine claims all statements can be translated by RTD into statements that do not
presuppose the existence of non-existent objects. 'Pegasus exists' becomes 'There exists x such
that x is-Pegasus (or pegasizes)' which is then a false statement. The only thing required here is
that we now accept an ontological commitment to the contents of the domain of the bound
variable x. Essentially, what this means is that we accept the existence of all possible values of x,
which in this case might be every individual animal, or all collections of particles in the universe.
We can imagine an objection to this. The statement Δ, 'The man is wearing a hat', and its
RTD translation Γ, 'There exists x such that x is a man and x is wearing a hat, and there does not
Kan 2
exist x such that x is not a man and x is wearing a hat'1, seem to be incorrect statements; there are
surely multiple things to which 'is a man' may be applied, and some of these may not be wearing
hats.2 But these seem to be perfectly reasonable statements. The error in reasoning comes from
the failure to remember that Δ implies we are referring to a specific man, while Γ does not. Thus
in order to correct this mistake, we might translate Δ into Γ', 'There exists x such that x is a man
and x is wearing a hat and x is being referred to by me, and there does not exist x such that x is
not a man and x is wearing a hat and x is being referred to by me'. More generally, if we take our
previous example of c and P, and we are referring to 'this' or 'that' c, then 'Pc' is really 'Pc and
Tc' where T is the predicate which expresses reference to, and this translates into 'There exists x
Now we can say that these statements do presuppose the existence of non-existent
objects. 'That I am referring to' seems to presuppose that the object of my reference exists, else
how can I refer to it? But it is important to look exactly at what is being stated. The statement Ψ,
'There exists x such that Px and Cx and Tx' only requires that there exist the objects in the
domain over which the bound variable quantifies, but the existence of 'Tx' is not required to exist
for us to make the statement Ψ. It might well be the case that 'There does not exist x such that Px
Thus we might still accept Quine's rules for ontology, by requiring the existence of the
objects within the domain from which bound variables are drawn. However, before we can
decide whether Quine's theory is valid, we must consider his denial of possible beings.
Possibility
Quine makes the claim that theory of the existence of possible beings is a fallacy, as it
1
The second clause appears because 'the' in 'The man is wearing a hat' implies uniqueness.
2
Inspiration for this line of reasoning comes from Ramachandran's "A Strawsonian Objection to Russell's Theory of
Descriptions", Analysis, 53 (1993) pp. 209-212.
Kan 3
had about them. Quine admits that a statement as a whole may be 'possible', but denies possible
entities. However let us consider the quantum mechanical case of the hydrogen atom (or any
atom, for that matter; this is just the simplest case). Before measurement, the precise location of
the electron cannot be known, only probabilities of its being found at certain locations. Thus we
might say there is a possible electron at Cartesian coordinates {a1, b1, c1} (we could give its exact
probability, but that would still make it only possible), and there is another possible electron at
coordinates {a2, b2, c2}, and so on for all the locations the electron might be at. So, it seems,
A counterargument would be to claim that all these possible electrons are the same one
electron, and because the electron has a total probability of existing somewhere of 1, it is not
possible, in fact it does fully exist. However we should be able to speak of the electron that is at
{a1, b1, c1}; this electron is different from the electron at {a2, b2, c2}, as it differs in its location
property. Surely two objects with two different sets of properties are not identical.
This does not defeat Quine, however, because of the very fact that he admits possible
statements. We can rephrase 'there is a possible electron at Cartesian coordinates {a1, b1, c1}' with
RTD into 'it is possible that A' where A is 'there exists x such that x is an electron and x is
located at {a1, b1, c1}'.3 In this way the possibility becomes a function of the statement. In fact, it
appears all such possible entities can be regarded this way as possible statements. For example, if
we consider a specific possible centaur, such as the one that may or may not live in my backyard,
we might say 'it is possible that B' where B is 'there exists x such that x is a centaur and x is
located in my backyard'. We then know that this statement is false, because we have looked and
not seen it, we have used infrared detection, etc. The only things that cannot be regarded this way
3
I use the multiple statement form 'it is possible that A' and A because it is important to remember that the
possibility applies to a statement, whereas 'it is possible that there exists x such that …' allows the possibility to
apply perhaps to the being.
Kan 4
This does not pose a problem for Quine, however, as it can be shown that universals do
not exist.
Existence
It is important to first clarify the idea of existence. To say something exists under RTD
means 'there is something that we accept in our ontology that P' where P is the predicate which
describes that thing. I would tighten the definition of exists, however, to 'there is such a thing in
the physical universe that P'. After all, when we discuss ontology and "What there is" it seems
less interesting to ask, "What ontology is best suited for speaking of things?" but rather "What
things exist in the physical world?" (assuming we believe in the existence of things independent
of the mind) or "What things exist in the mind?" (assuming we do not believe in the existence of
thing independent of the mind). Although here we are assuming the existence of things
phenomenalistic one.
This conflicts with Quine's previous definition of existence. "When we say that some
zoological species are cross-fertile, we are committing ourselves to recognizing as entities the
several species themselves, abstract though they be. (Quine 32)" Here we would be forced by the
previous definition to recognize non-physical beings. Under our new definition, statements
translated in RTD are only valid if the bound variables in them quantify over sets that contain
objects that exist physically. This, however, is compatible with most statements valid in the
previous definition of existence. Quine's 'Some dogs are white' quantifies over a set that includes,
Under this definition of 'exists' then, to say a property exists (take red) is to say, "Red is a
thing in the physical universe." But we will find no instances of the physical red, only the
property red. There can be a property that does not exist physically; in fact, all properties are of
this nature: no properties exist. Properties only occur as ideas. If we let Pegasus[idea] represent
Kan 5
'the idea of Pegasus in one's mind', then it is obvious that Pegasus[idea] exists. Analogously,
red[idea] does exist; what we call red[idea] is just a pattern of neural firings in the human mind.
Universals
Now it becomes clear that universals do not exist (physically), just as properties do not
exist. Universal ideas might be argued to exist, however, as a pattern of neural firings in the
human mind. And thus an idea like this might be said to be a universal, for example, if red[idea]
exists in everyone's mind (or simply more than one man's mind). However, the red[idea] that
exists in my mind (or rather, through my mind, as it exists physically as a pattern exhibited by
my brain) is not identical to the red[idea] that exists through your mind: perhaps your vision is
slightly different than mine due to differing amounts of cones and rods in our eyes, or perhaps
one of us is colorblind, or perhaps it is simply that I have different associations with the idea of
red.
Likewise, we might think of the universal justice. The justice[idea] that exists through my
mind is different from the justice[idea] that exists through your mind. Thus we can speak of a
justice[idea-Nathaniel] that is the idea of justice that exists through my mind, and likewise a
justice[idea-Quine] for the idea of justice that exists through Quine's mind. Thus universals do
not exist.
Conclusion
The incompleteness in Quine's argument is that he accepts the existence of the objects in
the domain over which a bound variable is quantified. So long as one accepts this criterion, he is
allowed to pick and choose his ontology, just as one would pick and choose a system of natural
ontological system is too closely tied to one's chosen physical system to be done independently.
Kan 6
A phenomenalistic view of experience still upholds a physical law structure for our experience.
If we consider Quine's theory (the use of Russell's theory and acceptance of possible
statements) altered so that the bound variables range only over the physically existent, then we
get a plausible ontology that denies universals (which fits within our physical law experience). I
believe Quine to be mostly correct, his theory just does not go far enough, for as Quine states, we
are asking "What is there?" and not simply "What things is it convenient to say there are?"