On Routley's "On What There Is Not"
Comments on Richard Routley’s “On What There Is Not”
Richard Routley, “On What There Is Not,” Philosophy and Phenomenological Research, Vol. 43, No. 2 (December 1982), pp. 151-177.
Richard Routley (later Richard Sylvan) was never one to shy away from something because it is the unpopular opinion, and therefore it is not surprising that, at a time when Quine’s “On What There Is” was often treated as definitive, he jumped right into raising objections. His major article on the subject dives right into the fray without any hesitation:
Most things do not exist. For every thing that exists, for instance Three Mile Island nuclear reactor (an odd product of the null set), there are several things that do not exist, abstractions beginning with the null set and the property of being that reactor. And there are a great many abstractions other than those directly generated by the null set and things that exist. These truths we hold to be elementary, and where not self-evident they can be argued for. Quine, however, in a very bold stroke, has stolen much of the terminology we ordinarily use to state, and argue, these elementary facts - and as far as most philosophers are concerned, he has got away with it. (p. 151)
One of the things we noted when commenting on Quine’s article was that good for goose is good for gander; that is, Quine repeatedly uses moves in his arguments that his fictional McX and Wyman could just as easily use against him, and he would have no right to complain if they did. Quine had claimed that the two fictional philosophers, particularly Wyman, twisted the language to fit their (assumed to be) strange views; he can’t really complain if Routley jumps in and points out that he is doing the same. Quine had claimed that it was obvious that the answer to “What is there?” was “Everything”; as we noted there, this is not at all obvious, and that we often don’t answer that sort of question with that sort of answer. So Routley starts on strong ground by noting that we can make perfect sense of someone denying that “Everything” is the correct answer.
To make the matter more specific, Routley holds that we often use ontologically relevant terms in natural language, in ontologically relevant contexts, in ways that don’t have any existential import of the sort Quine assumes: thing, what, ‘there is’, everything, and so forth. We can (and often do) answer questions like, “Are there any fictional characters?” or “Are there any impossible things?” or “Are there things that are merely possible?” or “Is there anything that doesn’t ever happen?”, with the answer, “Yes,” and we can even give specific examples. Somewhat more controversially, Routley holds that when we return these “stolen goods” (p. 152), a number of philosophical problems begin resolving themselves. This includes the Platonic puzzle of non-being with which Quine had begun his discussion (that it seems that non-being must be if we are to think or say anything about it). Quine had been worried about multiplication of entities, but if we recognize that saying that it makes perfect sense to say that (e.g.) a round circle is an impossible thing, and that ‘thing’ here doesn’t imply any kind of existence of the sort Quine has in mind, there are no entities multiplied at all. If you want to simplify the situation and reduce the number of entities that are assumed in our explanations, the most straightforward way to do it is just to recognize that many of the things that occur in our explanations are not real entities at all. Unlike Quine’s radical surgery, this will at least often leave the explanations exactly as they were.
Thus Routley rejects not only Quine’s view, but also the views of both of Quine’s nonexistent philosophers, McX and Wyman. ‘Cerberus’ does refer to something, namely, Cerberus; and simply referring to Cerberus does not commit us to any particular account of whether Cerberus is or is not in any way. You don’t have to stuff metaphysics into your logical quantifiers; you can just let them be logical:
…there need be no confusion of meaning with naming (though meaning can be explicated through interpretation which is wider than naming – in worlds); there need be no confusion of meaning with things talked about (but naming is a sub-species of being about); there need be no appeal to attempts (inspired by the Ontological Assumption) to make nonentities exist somehow, for example, as shadows of entities, or as names, or somewhere, for example, in the mind, or in myth or in fiction and so on. (p. 154)
This might be considered the core idea of Routley’s argument. It is not only the primary component of it, but (we think) entirely correct and definitively established by his argument. Our form of logical reasoning does not itself precommit us to a metaphysics unless, perhaps, we are only using a logical system specifically rigged to conform to such a metaphysics (and even then it would not do so as a logical system). It is irrelevant to a purely logical analysis whether ‘Pegasus’ actually exists in the real world or not, and logic makes no assessment of how we should think of ‘round square’ beyond any logical contradiction involved. But Routley goes beyond this, in that he thinks we can generally do without these kinds of further accounts at all; this is his particular position, noneism. On this point he is sometimes on shakier, or at least less obviously stable, ground.
A significant component of Quine’s argument was that possible objects, etc., lack identity conditions. However, as Routley notes, we regularly make judgments of identity and difference with respect to such things. This is not an accident: “For the very same notions of identity – most importantly, extensional identity – and distinctness that apply to entities apply likewise to nonentities, objects that do not exist” (p. 156). We can make perfect sense of saying that Heracles and Hercules are the same, or that Pegasus and Thunderhead are different, and we can give exactly the same sort of explanation for those judgments whether or not Heracles, Hercules, Pegasus, or Thunderhead exist or not. Nonentities have properties ascribed to them (Pegasus is winged) and therefore there is no problem with any of this. We can therefore provide criteria of identity for Pegasus even if there is no Pegasus. Likewise, we can say that a real horse is more like Pegasus than it is like Medusa, or that dryads and naiads are similar. (And as we noted in our comments on Quine, Quine himself depends on his fictional McX and Wyman having such criteria of identity, since his comments on McX and Wyman repeatedly presuppose that he is talking about the same McX and the same Wyman throughout.)
On similar grounds Routley will answer Quine’s objection about possible men in the doorway:
Consider now a nearest open doorway and an arbitrary fat man who never has existed, e.g. Mr. Pickwick. Ask whether Mr. Pickwick is in that doorway. The answer is, as a matter of observation, No. In literal contexts the answer is the same in the case of every other merely possible fat man. Hence, the answer to the question ‘How many (merely) possible fat men are in that doorway?’ is: Zero….There are zero possible men in that doorway, zero possible fat ones, and exactly the same number of merely possible thin ones. (p. 159)
Routley suggests that when people are impressed by Quine’s possible men example, they are confusing modalities, and in particular, “There is a possible man in that doorway” and “Possibly, there is a man in that doorway.” (We are not so sure that this is the case, and Routley moves very quickly over this part of the argument. Our inclination is to think that Routley has slipped up by doing what he already knows he should not be doing: he is trying to make a logical distinction bear more metaphysical weight than it can.)
People tend to assume that there would be a significant differences between entities and nonentities; Routley wants to say that, at a logical level, they are very similar. They are not any more indeterminate or determinate than each other, and almost everything that Quine says about nonentities could be said about clouds, for exactly the same reasons. His cloud parody of Quine is quite amusing and a good example of how correctly to use a parody argument to make a parity argument, which seems to be a lost art these days. Routley is right that Quine gets his results by deliberately imposing a double standard; he exempts his preferred entities from the standards to which he holds the entities and nonentities he does not like. (We ourselves noted, in our comments on Quine, that his objections to the possible men in the doorway would apply to actual men in the parking lot, given only the same information.) Routley’s diagnosis of this is interesting; he suggests that people like Quine do this because they over-assimilate entities in general to artifacts, which can have sharp boundaries and very determinate status. But we ourselves think that this is probably a more general impulse, and that people often attribute sharp boundaries and very determinate status to artifacts not because of the artifacts themselves but because of the impulse, whatever its real reason may be. It’s harder to get out of saying that there are tables than that there are fictional detectives, so people try to pretend that tables are extremely well-defined things with no indeterminacy. When we press the matter, we find that it isn’t so; physical artifacts often have indeterminacies similar to those of fictional entities.
There is one set of worries that are somewhat trickier, however, namely worries that are concerned with reference:
It is thought that one cannot have contingent identities between nonentities because this is identity of reference and in the case of nonentities there is no reference to be identical. This problem is removed (in the theory of identity) by distinguishing identity of reference and extensional identity, that is identity under extensional properties. Referential identity, which can only apply truly to existing items, is defined in terms of coincidence of entities in extensional respects: it is extensional identity of entities. Thus if a and b are referentially identical ‘a’ and ‘b’ have interchangeable referential occurrences. Since expressions about nonentities have no referential occurrences in true statements, nonentities cannot have identity of reference. But they can still be extensionally (or contingently) identical, since they have extensional properties, and extensional identity of nonentities is coincidence of extensional properties. (pp. 165-166)
Here is a point at which we think Routley is on shakier ground, and we’re not quite clear why, having gone so far on other things, Routley balks at denying that reference can only apply truly to existing items. It’s not inexplicable why someone (like ourselves) would do this. It does at least at first glance seem that, if we refer to something, there must be something to which we refer; this does seem to require that that something be at least a logical object of some kind, and therefore an entity of some kind. But if we take Routley’s route, extensional properties do everything that is ever attributed on a purely logical basis to reference; ‘reference’ will always presuppose a metaphysics and will never in itself be relevant to a purely logical analysis.
Quine, of course, built his attempts at paraphrase on a version of Russell’s theory of descriptions, but Routley responds that this theory of descriptions keeps getting things incorrect. A Meinongian can recognize that a round square is square; this is a reasonable ground for thinking that it is contradictory. But on Russell’s theory of description, “A round square is square” is just false, which is certainly wrong. To take one possible way to develop this, it would also be the case that “A round square is round” is just false. But if it is just false that a round square is square and it is just false that a round square is round, then it follows that a round square is not contradictory for being square and round; we’ve just said it is neither of these things at all. This is far from being a unique case.
As a ‘noneist’, Routley wants to say that we think and speak about many things that are just don’t exist at all: attributes, properties, propositions, numbers, relations, classes, sets. But this does not eliminate them from our discourse; we have good reason to use them all. The number 1 is a number; it doesn’t exist, nor does any number, but we can explain things with it. Here Routley, at least in this paper, goes quite vague. It is in fact unclear how numbers can explain anything if there are no numbers, and we don’t get any account here. (Routley does attempt to expand on this in later works; we don’t think the attempts wholly succeed, but this gets into other arguments.) Perhaps, though, we can say that Routley has good reason to think that none of Quine’s arguments give a reason to say otherwise – Quine just assumes that if we must use something in explanation that it must exist, and there doesn’t seem to be a straightforward way, given his account of conceptual schemes, that he could argue this. To be is not to be the value of a bound variable, Routley wants to say; lots of bound variables are not, and things can be without being the value of a bound variable. (In this part of the argument, we think Routley is taking Quine to be making a somewhat stronger claim than he actually is. However, he is not the only one to do so, because some of Quine’s claims are stated as if there were no conditions or limitations to them, when in reality there seems to be some conditions or limitations built into the context in which they are said.)
Quine, of course, puts a great deal of emphasis on the importance of the ontological commitments of the ‘language of science’. Routley does not give Quine any quarter here. The language of science is not regimented the way Quine wants. The language of science is not the only one that has evolved specifically to raise ontological commitments. (We might add: Scientists often do not care about the ontological commitments of their theories, often treat them as having none, and often propose theories in which there are components for which they do not know whether they should be treated as entities or just as calculation devices. Likewise, when scientists are really concerned with ontological commitments, they often switch to ordinary language or ordinary language with a supplemental technical vocabulary, because this is often the language in which we talk about experimental results and what we should conclude from them. Cases like interpretations of quantum mechanics show that scientists can agree on the mathematical theory and not agree about its ontological commitments. And so forth.) Perhaps more directly important, conceptual schemes regularly include things that are taken in that conceptual scheme not to exist – and, of course, this would certainly be true of the noneist conceptual scheme Routley proposes as a possibility. Quine claims that ontologies are at the basis of conceptual schemes; Routley denies it. Nothing prevents a conceptual scheme from recognizing things that don’t exist; on the other side, committing to a conceptual scheme does not always involve committing to a particular ontology.
Because of his emphasis on the language of science, which is heavily mathematical, Quine, despite being a nominalist, makes some allowances for numbers and classes as a useful ‘myth’.Routley notes that at this point we have simply stopped talking about what is true, and as a result it becomes impossible to say what is going on in such a ‘myth’, or what it even means to call it a ‘myth’. At this point in his article, in fact, Routley is beginning to be reduced to going through each of Quine’s major claims toward the end of his article and pointing out that Quine does not have an adequate argument for this claim and is largely trying to slide things by the reader by rhetorical means.
Of course, as we noted when discussing Quine’s paper, Quine later abandoned or modified several key components of his article. Routley recognizes this, but thinks that all of Quine’s attempts to provide a more correct account than “On What There Is” provides, fail. We will not go through all this part of the argument. What is particularly to be noted is that Routley argues that we get a recurring problem about what ‘ontological reduction’ is, namely, that we get the transference of “a certain softness we can find in what we have good reason to accept as existing (and, similarly, as true) to a softness in what exists (and what is true)” (p. 174). All of the evaluation is in terms of things like what is ‘useful’, and as a result some terms are ‘deleted from’ (or, as it is often put today, ‘not allowed into’) our ontology. This gives us a smaller set of primitive subject terms that are most useful for a given theory, but why should we consider this as affecting the question of what exists? There are other theories, other standards of usefulness, and we might be able to work just fine with a different set of primitive subject terms. When we have done our ‘ontological reductions’, what have we actually done with respect to ontology? Nothing at all. What exists is not affected by whether it is useful for the particular theories we happen to be looking at.
In this, we think Routley is exactly right, and it is a problem that always plagues paraphrastic approaches to fictional characters, universals, possibilities, or anything else. We don’t actually have any direct power to use our theories to determine what is, no matter how useful they are, and we have generally no authority over the question of whether something should be ‘allowed’ into our ontology. What is, is, whether we end up making room for it in our theories or not. When we do all our Quinean paraphrases, it seems like the question of what exists or not, what is or not, is still left entirely untouched; all we’ve done is change our way of talking to something that, perhaps, we like better for some reason. But our ways of talking are not determinative. Quine has not shown that McX or Wyman or Meinong or a noneist like Routley are wrong; he has just stated his aesthetic preference for talking a different way. But Quine, or anyone else who takes a Quinean approach, can talk any way he wants, and play up the usefulness of talking so, and it will still be true that none of the questions of ontology are thereby answered.
The argument of Routley’s article sprawls more than that of Quine’s very rhetorically punchy article, and it becomes very noticeable that there are a few pages in which Routley is doing very little more than taking a series of things Quine says and responding, ‘No, that’s not right. That’s not right, either. No, that’s wrong, too.’ Indeed, some active impatience and anger sometimes seems to break through in places. But Routley was a logician in Quine’s league, and from his perspective Quine was trying to slip falsehoods about logic and its relation to ontology past the naive reader, so he responds as anyone might in such a situation. Nor was he entirely wrong. Quine had a very narrow, very controversial, view of logic, one that most logicians then and now have rejected, and the argument of “On What There Is”, important as it was for kickstarting interest in ontology, was always less a serious reflection on ontology than a sales pitch for his particular view of logic. You don’t have to agree with Routley’s full noneist view to see that many of his points hit, or that Quine says many things in that article for which he never provides any adequate argument.