Chris Wilcox - Thesis - Ch 3
—-
Introduction
As mentioned earlier, the indispensability argument as it is being discussed here has a valid argument form. Given this, a common approach to criticising the argument has been via the denial of one or more of its premises (for example, Field famously denies that mathematics is indispensable to our best scientific theories\citet{Field:???? and :???? for example.} — but {[lime suggest “and” ]} many other criticisms have been made). A path somewhat less travelled however, is the denial of Quinean naturalism.
In what follows I will argue that Quinean naturalism fails in two ways: firstly, the Quinean ontic thesis fails to provide a satisfactory, unambiguous account of what exactly the ontological commitments of a theory are; and secondly, the idea of a “semantic ladder” leading from the semantic to the ontological inherent in the Quinean naturalist’s ontological position is highly problematic. I believe that These {[lime typo capital T ]} two issues point to the need to reevaluate the context in which the ontological claims of analytic philosophy are to be evaluated {[lime repetition of “evaluate” is fine semantically but sounds clunky ]} . It will end up becoming apparent that there is a sense in which ontological claims should be viewed as phenomenological (in a broad sense) rather than fundamentally metaphysical.
I will begin with an examination of the {[lime omit “the” ]} just what a scientific theory commits us to, ontologically speaking.
The Ontological Commitments of Scientific Theories
In Chapter 1, the Quinean procedure for determining the ontological commitments of our best scientific theories was discussed {[lime switch between first-person and third-person language ]} : one first translates a theory into an ideal logical language and then determines which bound variables the theory needs - these variables are then held to be what that theory is ontologically committed to (i.e. if one holds the theory to be true, then one is committed to the existence of the entities represented by the bound variables). Yablo provides one good reason to question this picture of ontological commitment.
Yablo begins by assessing the debate between Carnap and Quine over ontological commitment with a view to building on/modifying Carnap’s position. Central to Carnap’s positionlaid out in is the idea of a “linguistic framework” (a set of rules for use and meaning of linguistic terms). According to Carnap, all meaningful discourse takes place within such frameworks {[lime or, more precisely, every piece of meaningful discourse takes place within some such framework ]} . Thus there are two ways in which one may interpret the claim “there exists an X” (where X is any entity): firstly, one may take the existence claim to be postulated internal to a particular linguistic framework, in which case the assertion is almost tautological (and certainly of no ontological significance) given that the rules of the framework are meant to be stipulating the meaning and usage of the term ‘X’ {[lime Need to expand this. It’s not yet obvious why that gets rid of ontology even locally. ]} . Secondly, the claim could be taken to be made external to any linguistic framework. However, given that it is via linguistic frameworks that assertions gain meaning, such an interpretation would render the existence claim meaningless. Thus, Carnap concluded, existence claims can be of no metaphysical import. Better stick to “ontological” here. Some people (e.g. Price sometimes) use “metaphysical” in a very (strangely) broad way so that even linguistic frameworks are metaphysical.
Quine responded by claiming that Carnap’s position is reliant upon the analytic/synthetic distinction. The distinction between framework-internal and framework-external questions, he argued, is just another form of this distinction. Given this, and the fact that he denied the coherence of the analytic/synthetic distinction, Quine denied the coherence of Carnap’s position.
Yablo attempts to develop a position along the lines of Carnap’s but also employing the distinction between metaphorical and literal content. He begins by arguing that Quine’s critique is flawed. One does not need to hold that rules making up a linguistic framework are analytic in order to be able to understand the need for a framework {[lime ??? ]} in order to understand the meaning of terms. Additionally, Yablo points out that Quine himself accepts that a statement can be asserted purely for practical advantage without the asserter actually holding that what it entails metaphysically is actually the case. This acceptance is at the heart of Yablo’s modified view.
Yablo’s proposed modified linguistic framework view involves frameworks being adopted as a kind of “game” {[lime should probably make it clear that you know this is Wittgensteinian language ]} in which the players assess the truth and falsity of statements the framework without any commitment to belief in the truth and/or falsity of those statements outside of the framework. Thus, Yablo argues, under his modified Carnapian view there are again two ways in which a statement may be interpreted: literally (i.e. external to a particular game/linguistic framework) or (i.e. internal to a particular game/linguistic framework). {[lime Expand, to make it clear to an examiner why internal statements are metaphorical. Note that it needs to be clear that Quine accepts this … unless you change the argument of the following paragraph slightly. ]}
Yablo goes on to argue that this literal/metaphorical distinction causes problems for the Quinean ontological regime. Given that Quine accepts that assertions may be made in a metaphorical sense {[lime YES … but cite something ]}, and that when this is the case no ontological implications may be drawn from such assertions, Quine needs to provide a clear demarcation criterion to differentiate between metaphorical and literal assertions if his ontology is to be viable. {[lime I would add: If Quine, or a Quinean, cannot do this, then their position remains reasonable; but it involves an unclear delineation between assertions with ontological import and those without. This is just what one might expect given the way a Quinean web of belief operates, with some statements relatively but not absolutely more sacrosanct than others, and some statements surely only very provisional. Be that as it may, a Quinean who cannot make a sharp distinction between the literal and the metaphorical is left without a clear sense of what in the web is ontological. I should say that such a Quinean has no ontology at all. ]}
However, Yablo points out, this demarcation problem has been much discussed without progress.
Indeed, Yablo makes the (plausible) claim that for the most part when we make statements, we are unsure as to whether they are strictly literally true or if they are at least in part to be taken metaphorically. {[lime Examples to make this especially plausible? I’ll ask Alison if you can’t think of any. Or look in, or cite, “Metaphors We Live By” by Lakoff and Johnson and/or “Metaphor And Religious Language” by Soskice — I can lend you the latter. ]} Thus the Quinean needs to argue that in time these metaphorical parts of our statements will be “eroded”, leaving only the literal interpretation. {[lime Change the following to say that it’s unclear how a Quinean can do this without falling into the following trap. ]} However, this reduces the Quinean position to the following: one should sympathise with the idea that Xs exist iff the literal part of theories involve their postulation and one should count the part of a theory that involves the postulation of Xs literal iff there turn out to be Xs. Thus there is a circularity.
Therefore, there seems to be a strong case for it being unlikely that one will ever able to distinguish between metaphorical and literal usage. Hence, it is unlikely that one can construct a certain ontology using Quine’s approach. At the very least the supporters of Quine owe us a story here. {[lime small fix there. ]} As yet one has not been provided.
Building a Semantic Ladder
Central to the Quinean ontological view (and to many other views popular in analytic philosophy) is the idea that we can somehow analyse language and/or thought {[lime assertoric language and /or thought ]} in order to draw metaphysical conclusions. In essence then, they hold that there is some fundamental connection (sometimes termed a “semantic ladder”) between meaning (as determined via thought/language) and existence. I believe that there are good reasons to doubt this assertion.
The second avenue of criticism I wish to explore is founded in the antirepresentationalist approach to naturalism discussed in Chapter 1 (i.e. the view that our language or our thoughts are not somehow “representing” a world that is “out there,” external to us). As intimated above, what proponents of this sort of view criticise is the Quinean naturalist’s inclination to build a semantic ladder between language/thought and the ontological realm (i.e. the realm in which objects are held to exist in some strong metaphysical sense). Although views of this sort have been held by Rorty, X Xx and Carnap, amongst others, the argument I shall be focussing on (primarily due to its focus on the Quinean as a target) is due to Huw Price.~\citet{Price:???? and :????.}
Price’s criticism again begins with an analysis of the debate between Quine and Carnap. He points that Quine actually agrees with Carnap on the status of metaphysics {[lime If you’ve got time, say here which things you’ve said earlier foreshadow this ]}. Similarly to Yablo, Price claims that Quine’s criticism on the basis of the invalidity of the analytic/synthetic divide is not actually a move in the direction of restoring the non-pragmatic, external perspective necessary for traditional ontological investigation. Furthermore, Quine’s criticism of Carnap’s logico-syntactic pluralism does not affect Carnap’s deeper pragmatic/functional existential pluralism. Quine’s point is that there is only one {[lime univocal ]} existential quantifier whereas Carnap’s point is that it is a category mistake to assimilate different types of existence claim. {[green Probably merge this folowing bit into the Carnap explication paragraphs above ]} Furthermore, Carnap claims, there are many fundamentally different, non-scientific uses of existential quantification. Thus there is no reason for scientific existential usage to be treated as the basis for general ontological claims (as the naturalist would want). Indeed, such claims (external to linguistic frameworks) are non-sensical as they are category mistakes.
Price goes on to mount an argument against representationalist naturalism, of which Quinean naturalism (as commonly interpreted at least) is one example. This argument is cashed out as an attack on “object naturalism,” the view that all there is is the world of science (with aim being to argue for “subject naturalism,” the view that philosophy must not conflict with what sciene tells us {[lime what science tells us about ourselves as subjects … or something ]}). Price’s claim is that what he terms “placement problems” arise for the object naturalist. Problems regarding how to place objects to do with morality, meaning, and, of particular concern here, mathematics in a “natural” way. One can approach these problems via a linguistic conception or a material one.
Conception Under the linguistic conception, Price claims, subject naturalism is given theoretical priority as {[lime reconsider “as” ]} the use of language is well within the realm of the natural subject. Thus under this conception subject naturalism is the perspective from which we must assess object naturalism.
Price observes that the linguistic starting point requires that the object naturalist provide some substantial means of moving from dealing with a linguistic term to its subject (i.e. they must provide what terms a {[lime omit “what terms” ]} “semantic ladder”). This is the characteristic representationalist move. However, this is in conflict with deflationary accounts of such moves (as proposed by Quine amongst others) in which such a move is viewed as a simple disquotational shift from talk of how we talk of object terms, to talk of the object terms themselves. On such accounts one never leaves the linguistic realm. lime. Hm. Can say: according to disquotationalism, your mode of access to the external world is always linguistic, simply because it can only come from disquoting a piece of language. Thus, the object naturalist requires a non-deflationist representational account.~, the critique provided in \citet{Stich:???? also seems to suggest that more than mere non-deflationism is needed. {[lime omit or say “for example” ]}}
However, Price points out that Boghossian’s criticism of the irrealist position {[lime “B. has a criticism of the irrealist position which Price builds on” … and cite ]} may be adapted here to highlight that the object naturalist cannot just legislate out the irrealist position, they must allow for the possibility that their semantic bridge terms do not have an empirical basis. To do otherwise would {[lime fix ]} But the object naturalist is not able to do this, as they would then not have a naturalist semantic ladder to talk about objects in the first place. {[lime The linguistic end of the ladder would be missig, or at least unexplicated. ]} Additionally, it should be noted that the subject naturalist is immune from any Boghossian-based critique — they may simply remain silent on the ontological question. Thus it is possible to coherently remain a naturalist while denying representationalism. However, it is not possible for one to be an object naturalist on the linguistic conception.
Conception When considering the {[lime add “pure” or something ]} material conception, the first thing to note is that some fairly mainstream philosophical positions (such as noncognitivism) do not even make sense under it. They are reliant on the linguistic conception. Thus, the object naturalist wishing to make use of the material conception as a starting point has already failed to provide an uncontroversial basis for their position. Additionally, starting from the material is anti-naturalistic in at least one sense: it removes the possibility of learning from linguistic use {[lime expand or omit ]}. Furthermore, {[lime flag that this is the main point ]} semantic notions are often used as tools in metaphysical investigation. If such tools are taken to be used in any “substantial” way (which, as Yablo has pointed out, is difficult to determine), then one must assume the linguistic conception and the object naturalist has failed. If one is using semantic notions in some way (for example, analysing them in order to make ontological commitments), then one is back is again faced with the central problem of the linguistic conception: providing a “ladder” from the semantic to the ontological. {[lime fix typos including space at end of sentence ]} However, even when one ignores the preceding difficulties, there are still major problems for the object naturalist under the material conception. {[lime omit ]}
One possible strategy may be to adopt some material conception based on Lewis’ causal description idea.~ However, this is problematic if one denies that all objects of interest are causal (as indeed many Platonists hold mathematical objects to {[lime not … or talk about noncausal instead of causal ]} be {[lime — see Chapter 2 ]}). Furthermore, if one tries to accommodate the above objection by allowing for many different roles to be used on a case by case basis, one runs into difficulties to do with justifying why each of these should be accepted in each case (similar to Lewis’ “all causation is physical causation” premise). Thus the material conception is just as problematic a starting point for the object naturalist as the lingustic conception. Additionally, the subject naturalist is again not faced with these problems. By removing the representationalist assumption, the linguistic conception becomes a viable starting point and placement problems are reduced to puzzles regarding the origins of linguistic use.
Again, Price’s argument doesn’t categorically rule out a satisfactory response from the object naturalist{[lime s ]} (who include Colyvan). However, it does much to call into question the prima facie plausibility of such a position. In doing this it again shifts the onus of explanation onto the object realist.
Towards Analytic Phenomenology? {[green (This should be the conclusion) ]}
In the preceding chapters we have seen there are major difficulties faced by the indispensability argument and indeed by the Quinean ontological machine more generally. In particular two challenges have been discussed in detail: in Chapter 2 I argued that it is questionable whether the QOM can deliver abstract entities in any uncontroversial sense; and in Chapter 3 I endorsed two arguments to the effect that, even if this problem is able to be overcome, there are good reasons to doubt the Quinean naturalist underpinnings of the QOM.criticisms are by no means exhaustive. As mentioned previously, see \citet{Field:????, :????, and :???? for some other major criticisms of the indispensability argument.}
However, it should be noted that none of the arguments offered above lays claim to delivering a “knockout blow” to Quinean ontology. At best what they achieve is a shift in the burden of proof from those that doubt the accuracy of Quine’s ontological position to those who advocate it.
Now, assuming {[lime assume ]} that the preceding argument has led one to the conclusion that the indispensability argument and its Quinean ontological underpinnings are fundamentally flawed. Does this mean that one must simply “throw away” {[lime “throw away” those quotation marks ]} the Quinean approach? I do not think this is necessarily the case. It may be that one can just reevaluate the context in which such observations are made. It may well be the case that certain observations made int he {[lime typo ]} course of the debate (by both sides) are generally correct at the level of the semantic. Really, all that is being called into question is the move to the ontological. But then what sort of significance could these observations have? One interesting suggestion is that the claims be construed with a view to assessing their phenomenological significance, broadly construed. Such a claim could be built upon a more meaty concept of the subject involved in subject naturalism. {[lime add: Wright points ]} However, detailed discussion of this is far beyond the scope of this thesis. All the discussion of this paragraph is meant to show is that throwing away Quinean ontology does not necessarily mean throwing away Quinean insight. {[green Need to rework this to be a little less hand wavey and a little more qualified. ]}
—-