The Debate over Universals
At this juncture, Quine recognizes the need to address universals because of the introduction of predicates like “pegasizes,” having now dealt with the issue of rejecting the presupposition that Pegasus must in some sense be if it is said not to be. McX begins his argument for universals by citing the pre-philosophical common sense of recognizing that there are red houses, red sunsets, red roses, etc. The houses, roses, and sunsets have something in common, and that this commonality is all McX is referring to when he speaks of an attribute. That there are attributes is as “obvious and trivial”[6] as the fact that there are red houses, red sunsets, and red roses; no less does Quine expect from McX’s or anyone else’s ontology, which is basic to one’s conceptual scheme. Under McX’s conceptual scheme, the statement “there is an attribute ‘redness’” must follow from “there are red houses, red sunsets, etc.”[7]
Under a conceptual scheme different to McX’s, argues Quine, it is possible to admit the existence of red houses, roses, and sunsets while simultaneously denying that they have anything in common. “Redness” can be true of each of them individually, but there is no requirement that there must be some entity called “redness”; it could be that the houses, roses, and sunsets are all red irreducibly. Thus, there is no comparative gain in the explanatory power of McX’s theory provided by all entities given under the name “redness.” Incidentally, Quine notes that a potential argument for McX’s ontology was pre-empted by the earlier discussion of the difference between names and descriptions, and how the latter can be significant without becoming the former. Because of this, McX is unable to argue that in order for predicates like “red” or “is-red” to be meaningful, they must be names with the objective reference of a single universal entity.
In response, McX grants the distinction between naming and meaning, and cedes that “is red” and “pegasizes” are not names of attributes. With that, he counters that “meanings” are still universals, perhaps even things similar to the attributes he posits, whether named or not. Quine acknowledges this objection, explaining that he can only satisfy it by refusing to ontologically admit meanings, but he also explains his lack of hesitation in doing so: refusing meanings does not entail the absence of meaningfulness of words and statements. This is evidenced by the fact that McX and Quine can agree perfectly upon classification of linguistic forms as the meaningless and the meaningful, though McX’s criteria for meaningfulness includes the “having” (in one sense) of an abstract entity he labels a “meaning.”
Quine’s criteria are different; his basis for claiming the significance[8] of a linguistic utterance either derives from treating it as an ultimate and irreducible matter of fact, or from analyzing people’s ordinary reactions to the utterance in question and similar utterances. He reduces the useful ways that people commonly speak of meanings to two: the having of meanings (significance) and the sameness of meanings (synonymy). One’s “giving” the meaning of an utterance is his utterance of a synonym in a more ordinary and clearer language than the original. If such an interpretation of meaning is unsatisfactory, then one can simply speak of an utterance as significant or insignificant, and in relationship to other utterances (in synonymy or heteronomy). Though Quine recognizes the difficulty and importance of handling this approach properly, he once more refers to the lack of any increase in explanatory power resulting from adopting McX’s ontology- in this case, the adoption of special and irreducible intermediary entities called “meanings.”
In light of the preceding arguments, McX is led to question whether any statements are possible that lead one to be committed to universals or other entities Quine finds unwelcome. Once again, Quine cites Russell’s theory of descriptions in tandem with quantifiers, explaining that the entities can be stated as bound variables, so long as it is said that “there is something [a bound variable] which red houses and red sunsets have in common.” As explained earlier, the only way to make ontological commitments is to use bound variables. If “to be is to be the value of a bound variable,” whatever is said by names can be spoken of without names; names can be converted to descriptions, and then eliminated by Russell’s theory of descriptions; the purported namehood of an utterance can be repudiated if no respective entity is affirmed by the proper use of bound variables. Variables of quantification have a range of reference over the whole of an ontology (regardless of the particular ontology), and an ontological presupposition is convincing if and only if it must be considered among entities in this range of reference in order to establish the truth of an affirmation.
Therefore, the utterance “some dogs are white” does not commit the speaker to recognizing doghood or whiteness as entities. Rephrased, it states, “some things that are dogs are white,” which only creates the requirement that the quantifier “something” has a range of reference that includes white dogs, but need not include whiteness or doghood. However, it is also recognized that the statement “some zoological species are cross fertile” entails a commitment to the abstract entities “zoological species” unless the subject of the statement is reducible to another entity. Generally, a commitment to any reference persists until some means of paraphrasing a statement can be devised to change (or properly delineate) its bound variable’s reference.
Choosing an Ontology
Bound variables alone do not commit one to any single ontology, but only describes the process by which one becomes committed to an ontology. One means of adjudicating among ontologies, relative to a particular theory, is by finding an ontology whose entities are required to be within the range of reference of the bound variables of the theory in order to render the affirmations of the theory true. Modern disagreement over the foundations of mathematics is divided almost exactly on the issue of which entities lie in bound variables’ permissible range of reference.
Quine suggests that Occam’s razor be fully applied as an adjudicator among ontologies and that any ontology should be accepted in the same way that scientific theories are accepted: one must seek the simplest theory that accounts for all of the evidence. In the case of ontology, one must seek the simplest conceptual scheme that can be created to account for all the elements of raw experience. Quine’s argument, by his implicit admission, refutes the realist position on universals only as much as he asserts that a physicalist ontology containing universals is a useful “myth,” specifically in the fields of the physical sciences and more so in mathematics; put differently, he refutes it only by undermining its necessity by emphasizing the marked difference between naming and meaning, untangling Plato’s beard in the process. In the end, he states that the question of which ontology to adopt remains unanswered, with only “tolerance and an experimental spirit” as advice and judgment to be reserved for each myth based on its quality relative to a particular point of view.
Critique of Quine
One potential shortcoming in Quine’s argument lies in his approach to singular terms- their elimination and replacement by quantifiers- as an application of Occam’s razor. As was explained, if singular terms can be done away with, then their supposed implications about existence vanish. Hence, by using a singular term, one need not acknowledge the existence of the entity described by the term in order to be speaking meaningfully. Yet, if quantifiers could be done away with in the same manner, would they not be also subject to ontological elimination? One such possible elimination arises from combinatory logic, which was initially intended as a means of clarifying the role of quantifiers in logic by their elimination, much as quantifiers were intended to clarify existential statements by a similar process. In her book Philosophy of Logics, Susan Haack notes, “Quine concedes that his criterion doesn’t apply directly to combinatory logic, but observes that it can be applied indirectly, via the translation of combinatory into quantified formula.”[9] This may only be an evasion of the demand that the elimination of quantifiers places on their ontological status (via Occam’s razor). Even without delving into deeper discussion, it is a possibility worth mentioning, as it questions the validity of one of Quine’s necessary steps in reasoning.
Assuming that this problem with Quine’s methodology is somehow irrelevant to his general reasoning or can be answered appropriately, Quine’s dismission of the necessity of universals, as part of a common trend in dismissing the imaginary problems of Plato’s beard, is quite effective. Indeed, something appears highly flawed about the presupposition that denying the existence of an entity somehow presupposes that entity in the same sense that affirming that entity’s existence does. Quine accurately assesses logical possibilities (though not in those exact words) as meaningful, but does not make the error of “stealing” the concept of existence by making it a predicate.
For the non-Quinean, how much can Quine’s reasoning be used to make a more decisive case against the realist position on universals? On an absolute basis, Quine seems hesitant to commit himself ontologically,[10] and does not rule out the possibility of an ontology containing universals; he merely rules out the possibility of a poorly-reasoned ontology containing universals. To utilize Quine’s argument from an objectivist standpoint, there is not much that can be meaningfully done in the discussion of universals besides Occamite elimination, as is true with any other unnecessary multiplication of entities. In communication and in action, often times a person consistently holding the realist view on universals and a person not holding the view are totally indistinguishable, except in their assertions about the nature of universals. Lacking positive proof of a position or falsification of its negatory position, an appeal to Occam’s razor is the only logical argument left to pursue.
[1] P. 135
[2] Here, Quine briefly introduces a subtler-minded pseudonymous philosopher- Wyman- who advances the argument that Pegasus is simply an unactualized possible. Hence, when it is said that “Pegasus is not,” what is really meant is that Pegasus does not possess the property of actuality; in other words, it is an entity that is already understood to be. Wyman’s definition of the word “existence” entails that “Pegasus” has spatio-temporal connotations if “Pegasus exists,” but that “exists” does not (it merely refers to actualization). Quine then moves on to discuss some problems with unactualized possibles. This discussion is not directly necessary for his discussion of universals, except in as much as unactualized possibles can be looked at as entities in a similar manner to universals.
[3] What Quine means by “ambiguity” in this instance is that the quantifiers are subject non-specific on their own and only necessitate the satisfaction of arbitrarily-stated conditions in a proposition, not that they are poorly defined in usage.
[4] P. 137
[5] P. 138
[6] P. 139
[7] Ibid.
[8] Quine uses “significant” as interchangeable with “meaningful.”
[9] Haack, Susan. Philosophy of Logics. Cambridge: Cambridge University Press, 1978.
[10] At the very least, his hesitation is reflected in this article.