ONTOLOGY, NATURALISM, AND THE QUINE-BARCAN MARCUS DEBATE

F.M. JANSSEN-LAURET

Published as two papers, (1) ‘Committing to an Individual: Ontological Commitment, Reference,

and Epistemology’, Synthese, 193(2), 2016, pp. 583-604. Available at: https://www.academia.
edu/12100665/Committing_to_an_Individual_Ontological_Commitment_Reference_and_Epistemology_
published_in_Synthese (2) ‘Meta-Ontology, Naturalism, and the Quine-Barcan Marcus Debate’,
in Quine and His Place in History (ed. G. Kemp and F. Janssen-Lauret), Palgrave Macmillan, 2015,
pp. 146-167. Available at: https://www.academia.edu/8096867/Meta-Ontology_Naturalism_
and_The_Quine-Barcan_Marcus_Debate.

1. Talking About Objects

What is the best strategy for working out what there is according to someone (’s theory)? Natural
language isn’t a good guide.
‘There is a net migration between Poland and the UK’ is true, but does not entail that there is
an object called a ‘migration’. It means more people move from Poland to the UK than from the
UK to Poland.
Using a formal language to settle ontological questions (regimentation [8, ch.5] is helpful because:
(1) it avoids ambiguity [2]
(2) formal languages, unlike natural languages, have determinate consequence relations so the
question whether some existential claims does or not follow from some theory is answerable
(3) it separates the categoremata (for Quine: variables) from the syncategoremata (punctua-
tion, truth functions, predicates)
(4) irenic argument: if you think that there is something which I think there is not, I don’t need
to contradict myself by saying, in the object language, ‘there are some things I don’t believe
in’ — I put quote marks around your existence claim, and pronounce it false, speaking only
of words.

2. Direct Reference and Ontological Commitment

Quine’s idiom of choice is first-order logic without individual constants. All ontological com-
mitments, of any theory, have the form p∃xϕxq. But others (Barcan Marcus, Russell) think that
proper names or logically proper names (‘I’, ‘this’, ‘that’) uttered in a true, affirmative theoretical
context also come with an ontological commitment to their referents [3, 15]. Two issues:
(1) Perhaps they are right, and direct reference in a true, affirmative theoretical context incurs
an ontological commitment just as much as an existence claim would.
(2) Perhaps they are wrong, but we nevertheless want to be able to say what they believe in
in our canonical language: that’s the irenic point of ontological commitment.
Add constants to the canonical language? Quine says no: change the name ‘a’ into a predicate
‘Ax’ to fit into a Russellian definite description. But Barcan Marcus is unhappy about this and
proposes an alternative canonical language.
1
2 F.M. JANSSEN-LAURET

Quine’s arguments for dispensing with names:

Philosophy of Logic: constants complicate the syntax. True, but not decisive: a more complex
syntax is justified if it’s useful.
Empty names: some expressions that look like names don’t refer. True, but not decisive either.
Some apparent quantified phrases don’t have a corresponding entity. We use regimentation to decide
which apparent referential phrases in natural language are actually ontologically committing.
Quine’s real reason: his holist epistemology. Thought always reaches out to objects via the
medium of a complete theory, never directly.

3. Two Conceptions of Ontological Commitment

Global holism: Quine Extreme foundationalism: Barcan Marcus

• Variables as ontologically committing
expressions
• Logical form: p∃xϕ(x)q with objectual
quantifiers
• Dispenses with names by paraphrasing
them as definite descriptions
• Epistemic contact with objects is always
mediated by our best theory [9]
• Names as ontologically committing ex-
pressions.
• Logical form: px = tq for directly refer-
ential tags t
• Dispenses with quantified commitments
via substitutional reading
• Objects are encounterable by direct epis-
temic contact: acquaintance [4]

My hybrid view: The canonical language should be neutral on epistemology. It should be

capable of representing both direct and indirect access to objects, so both kinds of ontologically
committing expressions should be included. We want to be able to say, irenically, what the foun-
dationalist, the holist, and everyone in between believes in [?].

4. Against the Dispensability of Direct Reference

4.1. Quine’s Dispensability Thesis. ‘F a’ is equivalent to ‘∃x(a = x ∧ F x)’, and so the latter is
always substitutable for it. So ‘a’ is dispensable except in the context ‘a =’, which can be written
more concisely as ‘A’. ‘F a’, then, is equivalent to ‘∃x(Ax ∧ F x)’. The typical use of a proper
name is to uniquely specify an object. But, by hypothesis, the new predicate ‘A’ does exactly that.
After all, it is just short for ‘a =’. This amounts to a general strategy for dispensing with names
efficiently and in all contexts [10, pp. 25-26].
This is the formal equivalent of the ‘pegasising’ argument: names can be banished from the
grammar altogether without loss of information [7, p. 27].

4.2. Is ‘The Pegasiser’ a Bona Fide Paraphrase? Or is there some loss of expressive power
in translating out constants?
‘If the notion of Pegasus had been so obscure or so basic a one that no pat translation
into a descriptive phrase had offered itself along familiar lines, we could still have
availed ourselves of the following artificial and trivial-seeming device: we could have
appealed to the ex hypothesi unanalyzable, irreducible attribute of being Pegasus,
adopting, for its expression, the verb “is–Pegasus” or “pegasizes”.’ [7, p. 27]
Two strategies for excising ‘Pegasus’: a) brute force/‘x = Pegasus’; 2) ‘pat translation’.
 ONTOLOGY, NATURALISM, AND THE QUINE-BARCAN MARCUS DEBATE 3

4.3. Dilemma: Name-Recycling Predicates vs. Regress. Quine thinks the ‘=’ in ‘x =
Pegasus’ is not a logical predicate, but a dummy predicate short for indiscernibility-within-the-
language [10, p. 63]; [8, pp. 114–118]. So ‘x = y’ is equivalent to ‘x satisfies all and only the
same open formulae as y’; ‘x pegasises’ means ‘x satisfies all and only the same open formulae
as Pegasus’. But what is ‘x is identical with Pegasus’ or ‘x satisfies all and only the same open
formulae as Pegasus’ to connote unless something has already been named ‘Pegasus’ ? ‘Such devices
do not eliminate the name; they recycle it’ [6, p. 211, emphasis in the original].
What about replacing names with non-name-recycling predicates, like ‘the most illustrious stu-
dent of Socrates’ for Plato? Then what about ‘Socrates’ — is that name replaced by ‘the philosophis-
ing son of Phaenarete’ ? What about ‘Phaenarete’ ? We keep having to reach for either another
proper name or a logically proper name like ‘I’, ‘this’, or ‘that’, so this ends up in a regress.

4.4. Loss of Expressive Power 1: Descriptivism by Any Other Name. Perhaps name-
recycling predicates are salvageable if some natural-language proper names are directly referential,
but translated into descriptive form in the regimented language. But this is inconsistent with
Quine’s views on theory formation, which move from observation sentences via assent and dissent
to truth functions and then to reification, by means of the variable, on significant intersections of
observations. Objects only put in an appearance as the referents of variables, qua ‘ideal nodes’ [13,
p. 24]. They are theory-laden [12, 11], not directly encounterable and nameable.
This means that the only option left to Quine is a form of descriptivism—and name-recycling
predicates, by their very structure (‘x = a’) need the name ‘a’ to have been assigned to a referent
before they themselves can be formulated. So they cannot be used by descriptivists.

4.5. Loss of Expressive Power 2: Identity vs. Indiscernibility. According to Barcan Mar-
cus and Russell, only directly referential expressions—variables and names—can legitimately be
concatenated with the identity predicate. It is true or false tout court that ‘a = b’; that a and b are
the same thing. Although in natural language, we think it makes sense to write ‘The last pharaoh
= the eldest daughter of Ptolemy XII Auletes’, its true logical form is ‘∃x((P x ∧ ∀y(P y → x =
y)) ∧ ∃z(Dz ∧ ∀w(Dw → z = w)) ∧ x = z)’.
Quine, though, renders ‘x = y’ as ‘x satisfies exactly the same open formulae as y’. And of
course, he combines this view of identity with the Dispensability Thesis, recommending that we
translate ‘a’ into the canonical language as ‘the A-er’, or ‘the x such that x = a’. Put the two
together, and Quine is compelled to translate ‘a = b’ as ‘the A-er satisfies exactly the same open
formulae as the B-er’. This is weaker than ‘a = b’. It is always logically possible that indiscernibles
are distinct, but not identicals [14, p. 31], [5, p. 197].
This loss of expressive power suggests that a canonical language with constants and (old-
fashioned, non-facsimile) identity is to be preferred.

5. Epistemology
5.1. Global Holism about Objects. Motivations: Thought reaches out to objects only via the
medium of a theory. Also, posits sometimes change, and we no longer have reason to believe in
some entity. To work out whether a posit is dispensable, we need to consider its role in the theory
as a whole.

5.2. Foundationalism about Objects. Motivations: classic nominalism (Barcan Marcus), thought
reaching out to objects directly, via acquaintance. Plausible for all objects? Plausible for some class
4 F.M. JANSSEN-LAURET

of objects (perhaps psychological entities and privileged access)? Even if not plausible, still logically
coherent, and incurs ontological commitments, so should be representable in canonical language.
5.3. Foundherentism about Objects. As there are object-holists and object-foundationalists,
there is also an analogue of Haack’s foundherentism [1]. Here we find ontological commitments via
both constants and variables.

References
[1] Susan Haack. A foundherentist theory of empirical justification. In Ernest Sosa and Jaegwon Kim, editors,
Epistemology: An Anthology, pages 226–236. Blackwell, Oxford, 2000.
[2] Peter Hylton. Quine. Routledge, London, 2007.
[3] Ruth Barcan Marcus. Modalities and intensional languages. Synthese, 13(4):302–322, 1961. Reprinted in her
Modalities.
[4] Ruth Barcan Marcus. Nominalism and the substitutional quantifier. Monist, 61(3):351–362, 1978. Reprinted in
her Modalities.
[5] Ruth Barcan Marcus. A backward look at Quine’s animadversions on modalities. In R. Barrett and R. Gibson,
editors, Perspectives on Quine, pages 230–243. Blackwell, Oxford and Cambridge, Mass., 1990. Reprinted in her
Modalities.
[6] Ruth Barcan Marcus. Modalities. Oxford University Press, Oxford, 1993.
[7] Willard Van Orman Quine. On what there is. Review of Metaphysics, 2:21–38, 1948.
[8] Willard Van Orman Quine. Word and Object. MIT Press, Cambridge, Mass., 1960.
[9] Willard Van Orman Quine. Existence and quantification. In Ontological Relativity and Other Essays. Columbia
University Press, New York, 1969.
[10] Willard Van Orman Quine. Philosophy of Logic. Prentice-Hall, Englewood Cliffs, first edition, 1970.
[11] Willard Van Orman Quine. Five milestones of empiricism. In Theories and Things. Harvard University Press,
Cambridge, Mass., 1981.
[12] Willard Van Orman Quine. Things and their place in theories. In Theories and Things. Harvard University
Press, Cambridge, Mass., 1981.
[13] Willard Van Orman Quine. Pursuit of Truth. Harvard University Press, Cambridge, Mass. and London, revised
edition, 1992.
[14] Frank Ramsey. The foundations of mathematics. In The Foundations of Mathematics and other Logical Essays.
Kegan Paul, London, 1931.
[15] Bertrand Russell. The Problems of Philosophy. Williams and Norgate, London, 1912.