Evans: The Causal Theory of Names

BACKGROUND Speakers Denotation vs. Name Denotation Evans distinguishes between these, and suggests that both Kripke and the description theorists he criticizes have failed to draw this distinction. Speakers Denotation What a name denotes upon a particular occasion of its use (p. 296). Name Denotation What conditions have to be satisfied by an expression [x] and an item [y] for [x] to be a name of [y] (p. 296). Two Kinds of Description Theories There will thus be a description theory of speakers denotation, and one of name denotation: Description theory of speakers denotation A name (on a particular occasion of its use by a speaker S) denotes whatever unique item satisfies most or all of the descriptions S would associate with that name. Description theory of name denotation A name (as used by a group of speakers) is associated with a description or set of descriptions (obtained from the beliefs of those speakers). The bearer of the name is the object that satisfies these descriptions. Kripkes Attack on the Description Theory According to Evans, Kripke attacks the theory of speakers denotation, not the theory of name denotation. Evans himself thinks that Kripkes criticism is effective against the first theory, but does not touch the second.

EVANS VS. THE DESCRIPTION THEORY The weaker version of the descriptive theory of speakers denotation: some descriptive identification is necessary for a speaker to denote something (p. 297, right). Evans thinks that even the weaker version is false. He sees it as a fusion of two thoughts, what we might call the intention requirement and its underlying philosophy of mind requirement: 1. Intention requirement In order to be saying something by uttering an expression one must utter the sentence with certain intentions i.e., one must be aiming at something with ones use of the name. (p. 297) 2. Philosophy of mind requirement To have an intention or belief concerning some item one must be in possession of a description uniquely true of it. (p. 297) Evans does not seem to have any objection to (1), the intention requirement. His objections are directed against (2), the Philosophy of Mind, which he says (p. 298) is held by anyone who holds that S believes that a is F if and only if: [(S believes x (x & (y)(y x = y) & Fx)) & a & (y)(y y = a)] What does this formula say? Roughly: there is a property that S believes to be uniquely instantiated by some F, and that is, in fact, uniquely instantiated by a. In other words, S has a certain description in mind that he takes to single out a certain F thing, and a alone fits that description. The condition is not sufficient Thats because it leaves out any connection between the believer (S) and the object of belief (a) other than the fit between a and a description () the believer has in mind. There is nothing else to actually connect that very object, a, to S. The general form of a counter example to this condition looks like this: Think of some property, , that S believes to be uniquely instantiated by some F, and let a be an object that (unknown to S) uniquely instantiates .

E.g., suppose that Tom believes that the property of being the oldest living European at midnight on July 1, 1999 is uniquely instantiated. [Who wouldnt hold such a belief? There has to be some European who is older than all the others at that time.] Further, suppose that Tom believes that the person who uniquely instantiates this property is Norwegian (perhaps he thinks that Norwegians are, on the whole, very long-lived). Finally, suppose that the oldest European at that time is an Italian named Giuseppe. Then the proposed account of belief holds, absurdly, that Tom believes that Giuseppe is Norwegian. But, clearly, Tom holds no beliefs about Giuseppe. Whats wrong with the proposed account is its omission of any causal relation between the believer and the object of belief. In order for it to be true that S believes that a is F (where a is a name), there has to be some causal relation involving a (or the name a) and S (or Ss use of the name a). Our counter example shows that the proposed condition for S believes that a is F is not a sufficient condition. But that may be irrelevant, for the formula Evans gives on p. 298 seems to have over-stated the Philosophy of Mind requirement as he stated it on p. 297. There, it demands only a necessary condition. The condition is not necessary We need a different kind of counter examplea case in which S believes that a is F, but S cannot provide a property that is uniquely instantiated by a and that S believes to be uniquely instantiated by some F. A counter example is suggested by Evans case on p 298 (top right): What makes it one rather than the other of a pair of identical twins that you are in love with? So let S be the man in this example, and a and b are the twins. S believes that a is the love of his life, but there is no that is uniquely instantiated by a and that is believed by S to be uniquely instantiated by the love of his life. (Presumably, any features by means of which S would try to pick out a would also be shared by b.) So what makes it possible for S to believe that a is F in the absence of a uniquely identifying description? Once again, a causal relation may take up the slack. We may suppose that S has never metand may not even know aboutb. But he has met a, perceived a, etc. Thats what makes it a, rather than b, that hes in love with. EVANS VS. THE CAUSAL THEORY Evans thus agrees with Kripke that there must be a causal component to a correct account of naming. But he thinks that the causal theory unamended is not adequate (301). 3

Dubbing and change of reference A problem with making the dubbing event an essential part of the causal chain: it does not take into account the fact that a name can change its reference, and become the name of something other than the object originally dubbed. An actual example of this: Madagascar. Originally, it named a portion of the African mainland. But, misunderstood by Marco Polo, it became attached instead to the great island off the coast of Africa (p. 301). Imaginary case: the switched babies (p. 301). Evans wants to sketch a theory which will enable Madagascar to be the name of the island yet which will not have the consequence that Gdel would become a name of Schmidt in the situation envisaged by Kripke . (p. 301) What are the relata in the causal relation? Let us begin with the case of Louis (p. 298-9). In a pub, S hears a conversation about a certain Louis, and joins in the conversation. His use of the name Louis thus acquires (according to the causal theory) whatever denotation it had when it was used by the other participants in the conversation. And their denotation, in turn, is traced back through a causal chain to an initial dubbing of the bearer of the name himself, say, King Louis XIII of France. So on Kripkes picture, Ss use of Louis denotes King Louis XIII. But Evans thinks this is the wrong result. For suppose that S has completely forgotten the conversation. Indeed, S may become thoroughly confused, and say something like, I think Louis was a basketball player. Still, for Kripke, S is talking about Louis XIII if the causal history of his acquisition of the name traces back to the dubbing of Louis XIII. Evanss objection is that this gives initial dubbings magical powers:
for [Kripke] an expression becomes a name just so long as someone has dubbed something with it and thereby caused it to be in common usage. This seems little short of magical.

Instead, Evans proposes that the relevant causal connection is not between the dubbing of Louis and Ss subsequent use of the name, but between Louis himself and the body of information that S associates with (a particular use of) the name Louis. According to Evans (p. 301, right), Kripke

has mislocated the causal relation; the important causal relation lies between that items states and doings and the speakers body of informationnot between the items being dubbed with a name and the speakers contemporary use of it.

Kripkes picture Dubbing Subsequent use of the name

Evanss picture Object named Speakers body of information associated with the name

The arrows here represent causality, not fit. So on Evans s theory, the reason that Madagascar names the island is not that contemporary speakers beliefs about what they call Madagascar fit the island better than the mainland, but because the island itself plays a dominant causal role in their acquisition of those beliefs. EVANSS POSITIVE THEORY Evanss aim is modest (p. 302): an account of what makes an expression a name, but an account that will allow for change of denotation. Further, Evans will make use of an unanalyzed notion of speakers reference. His theory combines elements of both the description theory and the causal theory. From the description theory: The denotation of a name is fixed by the bodies of information (something like the clusters of the description theorist). From the causal theory: The fixing of the denotation of a name is by causal origin, not by fit. The question is not which object satisfies most of the descriptions we associate with the name?, but which object is the dominant source of the descriptions we associate with the name? 5

The theory itself is stated (in a fairly complicated way) on p. 304, left. It amounts roughly to this: NN is a name of x if (and only if): 1. There is a community in which people use NN to refer to x; 2. It is common knowledge that NN is so used; 3. The reference in (1) relies on the knowledge in (2), and not on the knowledge that x satisfies some predicate embedded in NN. So worded, the account of naming seems to preserve little of either the description theory or the causal theory. But that is only partially correct. For although (3) explicitly disavows the description theorists fit, the causal theory gets incorporated in (1). The unanalyzed notion of using NN to refer to x implicitly relies on a causal notion people use NN to refer to x only if x is the dominant source of information they associate with NN. Evans summarizes his theory with the Turnip case (p. 306): A youth, A, (who has the nickname Turnip) leaves a town while still a youth. Fifty years later, a man B comes to the town and lives as a hermit. Falsely believing that A has returned, the surviving elders start calling B Turnip. (Evans claims that B is not Turnip, and that the elders are mistaken in thinking this.) The younger residents pick up the name from the elders, and begin to use Turnip to refer to B. Eventually, the elders die off, and the only remaining users of the name continue to use it to refer to B. At this point, which (if either) of the two is Turnip a name of? Evanss theory can accommodate either answerit depends on other, as yet unstated, facts. 1. If no further information about A gets passed on from the elders to the younger townspeople, then B will become the dominant source of the information they associate with the name Turnip, which will then transfer and become a name of B. 2. If there is a sufficiently rich body of information about A that the elders have passed on to the others, then A may well remain the dominant source of their information (p. 306). In this case, their use of Turnip will still denote A, and if they are apprised of all the relevant information, they too would acknowledge that man over the hill isnt Turnip after all.

Dubbings reconsidered Evans rejects the dubbing in favor of the object dubbed as the thing playing the initial causal role in a causal theory. (We must trace our use of the name back to the object, not to the dubbing of it.) But this leaves a residual problem: how can we use a causal theory to explain the naming of abstract objects (numbers, sets, etc.)? For example, the numeral 17 is a name of the number 17. But how can there be a causal relation between a number and subsequent uses of a numeral? In short, how can abstract objects stand in causal relations? It would seem that it is always something concrete that is a cause. In this kind of case, it makes more sense to trace the causal chain back to the dubbing, rather than to the object. For although 17 is an abstract object, the act of dubbing it (first performed by some mathematician, no doubt) was a concrete event that can stand in causal relations to subsequent events. SUMMARY 1. Evanss theory is more causal than anti-causal. For a name NN to be a name of an object, x, there must be a causal connection (not just a descriptive fit) between NN and x. 2. But the causal connection is not between an original dubbing of x with the name NN and a contemporary users use of NN; rather, the causal connection is between x itself and the body of descriptions that the contemporary user associates with NN. 3. Evanss theory, although causal, is not anti-intentional. That is, Evans does not propose to replace the notion of a names denoting an object (or a persons referring to an object with a name) with some purely causal relation between the user and the referent devoid of intentional content. There is no effort, e.g., to say that the relation of S having x is mind can be explicated as a purely causal (nonintentional) relation between S and x. 4. Evanss proposal makes use of many notions that are left vague or unspecified. E.g., intending to refer, community of language users, dominant source, etc. So, as with Kripke, we get more of a picture, rather than a theory, of naming.