Disquotation and Cause in the Theory of Reference

Citation preview

Ridgeview Publishing Company

Disquotation and Cause in the Theory of Reference Author(s): Paul Horwich Source: Philosophical Issues, Vol. 6, Content (1995), pp. 73-78 Published by: Ridgeview Publishing Company Stable URL: http://www.jstor.org/stable/1523030 Accessed: 05/12/2008 06:54 Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=rpc. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that promotes the discovery and use of these resources. For more information about JSTOR, please contact [email protected].

Ridgeview Publishing Company is collaborating with JSTOR to digitize, preserve and extend access to Philosophical Issues.

http://www.jstor.org

n~~

ISSUES, 6

~~~~~6 ~PHILOSOPHICAL

Disquotation and of Reference Theory

Cause

Content, 1995

in

the

Paul Horwich

In the last twenty-five years or so discussion of reference has been focussed on the relative merits of three alternative models. First there is the description theory according to which x refers to y when x is associated with a description of y. (This is the Frege-Russell view, championed these days by Searle and Katz). Second there is the causal theory according to which x refers to y when there is a certain sort of causal chain relating x and y to one another. (This sort of idea has been promoted by Kripke, Putnam, Evans, early Field, Stampe, Dretske and Fodor; it deserves to be called the mainstream). And third, there is a relatively recent arrival on the scene, the deflationary theory (also known as 'minimalism' or 'disquotationalism') according to which x referring to y is roughly a matter of x being the word "N" (in quotes) and y being the thing N (out of quotes). (This point of view is advocated by the present author and in recent work by Field). Where does Loar stand in relation to these alternatives? As I see it, what he is proposing is a predominantly causal theory, laced with a dash of disquotationalism. More specifically, his view is that we have a pre-theoretical conception of reference which is tied to the

74

PAULHORWICH

disquotation schema, and that this conception points us towards the true nature of reference, which turns out to be a collection of causal relations. Let me try to improve this oversimplified characterization of Loar's view by explaining how he arrives at it. He begins with the intuition that reference relations are objective, external (typically causal) relations betweem mental terms (i.e. concepts) and the things to which they refer. However his view is more liberal, in two respects, than the standard causal theory. For Loar does not maintain that all terms (not even all names) refer in this way -only terms that belong certain semantic categories (for example, socially-deferential names and recognitional concepts). Moreover he does not maintain that there is a single causal relation constituting reference: he supposes that terms falling in different semantic categories will exhibit different objective reference relations. But this flexibility leaves him with two difficulties. First, for a given category of term, what singles out some particular causal relation as the reference relation for the terms in that category? For example, why should we identify reference with the causal relation between the word "tree" and trees, rather than the different one connecting the word "tree" and tree-surfaces? (This, he takes to be Quine's issue of the indeterminacy, or inscrutability, of reference). And second, what shared feature of the various causal reference rethat constitute reference for the various semantic lations -those marks each of them as relations of reference? How can categoriesthis bunch of different objective relations all count as reference? Loar's simultaneous solution to these two problems is to invoke disquotation. It is in virtue of our possessing a general disquotational conception of reference, and prior to our having identified which causal relations are the reference relations, that we know, for example, that our word "tree" refers to trees and not to tree-surfaces. Thus our pre-theoretical disquotational conception of reference provides us, firstly, with constraints that allow us to determine, for a given semantic category, which causal relation is its reference relation; and secondly, with an explanation for our grouping together, under the same heading ("reference relations"), the various causal relations that are identified in this way. Thus we see the ingenious combination of causal and disquotational elements in Loar's position: reference is a causal relation that is identified by means of a disquotational conception. But there remains, he thinks, a further difficulty. Having said that our conception of reference somehow engenders the knowledge that "tree" refers to trees, "London" to London, and so on, it remains to understand just how this is done. We need a characterization of our conception

6.

DISQUOTATIONAND CAUSE IN THE THEORY OF REFERENCE

75

of reference that will explain exactly how it could be that we are led by that conception to an appreciation of what refers to what, and thereby to the knowledge of what the reference relations are. This is especially problematic insofar as we are tempted to think of our conception of reference as a descriptive characterization of that relation. For it proves to be impossible to find such a thing. It can't simply be Reference is the relation between disquotational pairs since disquotation relates pairs of expressions, not expressions and things. (As Loar says, we don't arrive at an object by taking quotation marks away from a word). Nor can our characterization be Reference is the relation between the referents of disquotational pairs which is circular. And nor can it be Reference is the relation between the word "London"and London, the word "tree" and trees, and so on because that would make our conception implausibly dependent upon the particular words we happened to have in mind. Loar's answer is that our concept of reference is demonstrative rather than descriptive. When a term/object pair (x, y) is conceptualized as an instance of 'mention-use', or disquotation, there is, he says, a particular phenomenology, a particular feel, which the pair induces. What is going on in my mind when I identify the pair ("London", London) as a case of reference is similar to what is going on when I so identify the pair ("Aristotle", Aristotle) -in both cases I have in mind a pair of terms that are related disquotationally. Thus Loar arrives at the view that reference is what he calls "a recognitional-type-demonstrative"- a concept of the form 'that property again', which is triggered by the peculiar phenomenology associated with disquotation. Once again, the pair of entities, x and y, may be identified in a pair of ways, C and D, that are related disquotationally; this has a special feel to it which triggers a certain concept, 'that relation again'; and this is our concept of reference. So much for exposition. Now let me say briefly what seems to me to be correct, and what incorrect, in Loar's position. As a deflationist, I like the emphasis on disquotation. I agree that our conception of reference is intimately tied to the disquotation schema. Indeed I would say that our possession of this concept consists in nothing more or less than our disposition to instantiate that schema. I am

76

PAUL HORWICH

not convinced that Loar's phenomenological remarks add anything important to this idea; but nor do I see any particular harm in them. On the negative side, however, I doubt that reference has any sort of underlying nature. So I don't agree with Loar's view that our disquotational concept points us toward the true causal nature of reference. Let me end by making three points which seem to me to support a sceptical attitude towads the idea that reference is constituted by some underlying relation or relations. First: our conception of reference presupposes no such a thing and provides us with no reason to expect it. Suppose our conception does consist in a disposition to accept instances of the disquotation schema. Then this is the fact about the term "refers"that provides it (whether in English, Spanish, or mentalese) with its meaning. And insofar as we are prepared to deploy a liberal notion of 'property' whereby all meaningful predicates express properties, then we have every right to suppose that reference is a genuine relation. However, the question of whether or not this relation is constituted by some causal relation (or by any other objective, external relation) is an entirely separate issue. And the disquotational account of our concept -even given Loar's characterization of it as a recognitionaltype-demonstrative- gives no reason to expect that there are any such constituting relations. Second: deploying our disquotational conception, we can identify an unlimited number of term/object pairs, (x, y), such that x refers to y. In this way, as Loar points out, we are able to put ourselves in a position to discover a causal relation, R, such that, for all terms within a given semantic category, x refers to y if and only if xRy And on the basis of such a discovery, he says, we would be justified in concluding that reference (for that category) is identical to, or constituted by, the relation R. However, I think this final step is mistaken. We indeed might (though there is no reason to think we will) find some causal relation that is co-extensive with reference. But without a great deal of further argument it could not be concluded that reference is constituted by R. Consider another example. The basis for our supposing that the property of 'being water' is constituted by the property of 'being composed of H20 molecules' is not merely that every sample of water is made of H20. It is, in addition, that there exists a certain explanatory relationship between these properties: namely that the characteristics in virtue of which a certain sample is recognized as water are explained by the fact that it is composed of H20 molecules. In general, our grounds for supposing

6.

DISQUOTATION AND CAUSE IN THE THEORY OF REFERENCE

77

that superficial property S is constituted by underlying property U is that the possession of U explains the characteristics symptomatic of the possession of S. Thus a case for concluding that reference is constituted by R would have to include a specification of other properties that are correlated with and indicators of reference, plus an argument to the effect that the possession of those characteristic symptoms of reference is best explained by the holding of relation R. Therefore the burden of argument that must be met before we can suppose that R underlies reference is more severe that Loar suggests. Not only do we have no reason to believe we will ever discover a relation R that is so much as co-extensive with reference (for a given semantic category); but even that unlikely discovery would be insufficient for the conclusion that reference is constituted by R. Thirdly: it isn't just a matter of there being no reason to believe that such a case can be made: there is positive reason to believe that it cannot be made -there is positive reason to think that the characteristics of reference could not be explained in terms of an underlying causal relation. Therefeore, given the previous point, we have a positive reason to think that reference has no underlying nature. This reason (which I'm afraid I can only sketch here) is based on a plausible assumption about the function of our concept of reference: namely, that it is a device of semantic ascent -a device enabling us to formulate generalizations of a very special kind that would otherwise call for substitutional quantification. Here is an example. Suppose someone, speaking a language we don't fully understand, attributes a property -say, redness- to some object or other; but we don't know which is the thing to which redness is being attributed because we don't understand the singular term, "#", that is being used. How can we report what is said? With no concept of reference, we would have to put the matter as follows: If "#" means "a", then a is red, and

if "#" means "b",then b is red, and if ... where "a", "b",... are the singular terms of (extended) English. But this is an unstatable infinite list; there is an item for each thing "#" might mean. If, however, we have a term, "refers", satisfying the schematic disquotation principle If x means "*", then the referent of x is identical to * then that infinite list can be captured in a single statement. For, given the instances of that schema,

78

PAULHORWICH If "#" means "a", then the referent of "#" = a, and if "#" means "b", then the referent of "#" = b, and if ...

the original list is equivalent to If "#" means "a", then the referent of "#" is red, and if "#" means "b", then the referent of "#" is red, and if...

whose content (given that one of the antecedents is true) is The referent of "#" is red

Thus the notion of reference, insofar as it satisfies the disquotational principle, enables to capture certain generalizations that cannot be captured merely by using the usual devices (that is, "all", "every"or the universal objectual quantifier). If this is right, then facts articulated with the concept of reference (including those that correlate reference with other properties) cannot be deduced from ordinary non-semantic generalizations and, therefore, cannot be explained by them. Thus there could be no explanatory gain in supposing that reference has a non-semantic underlying nature. So the conditions for supposing that something constitutes the reference relation cannot be met. In summary, I see Loar as an advocate of the causal theory who has found that certain problems can be handled by treating it with a dose of disquotationalism. But this could be one of those cases in which the cure is more dangerous than the disease. For the disquotational account of our conception of reference, not only solves the indeterminacy problems, but threatens to leave nothing for a causal theory to explain.