INTENTIONS AND DEMONSTRATIONS
MARGA REIMER  has forcefully challenged David Kaplan’s recent claim (, pp. 582-4) that demonstrative gestures, in connnection with uses of demonstrative expressions, are without semantic significance and function merely as ‘aids to communication’, and that speaker intentions are what determine the demonstratum. Against this Reimer argues that demonstrations can and do play an essential semantic role and that the role of intentions is marginal at best. That is, together with the linguistic meaning of the demonstrative phrase being used, an act of demonstration determines what is said. I will argue that Kaplan’s view is borne out if we consider the referential intentions specific to communication. Reimer may be correct about such intentions as she considers, but she overlooks specifically referential ones. When these are taken into account, we find that although demonstrations contribute in a way to what is said, this does not make them semantically significant. For what is said, to the extent that it is not fixed by linguistic meaning, is determined by speaker intention, which itself can include the intention to refer to what one is demonstrating.
Inasmuch as Kaplan produces no argument for why demonstrations are not semantically significant, Reimer attempts to reconstruct one. She describes two situations in which the phrase ‘that dog’ is used to refer to a certain dog. In each case a certain dog Fido is salient, by virtue of being either the only dog around or the only clearly conspicuous one, say by barking raucously. In both cases the speaker intends to refer to the very dog he is pointing at. Reimer’s guess is that Kaplan arrived at his view by considering ‘scenarios where the intended and the actual demonstratum converge—and would have done so even in the absence of the accompanying demonstrations’ (, p. 178). In the latter case, which she oddly describes as involving a demonstratum even though there is no act of demonstration (I can only assume that by ‘demonstratum’ she means anything referred to by means of a demonstrative expression), no act of demonstration is needed to secure reference. So, she imagines Kaplan concluding, acts of demonstration are not semantically significant even when they are performed. However, she thinks there is a better way of handling the above cases, a way which also works in cases where Kaplan’s approach goes astray.
Reimer begins with the truism that an expression of the form ‘that F’ refers to a certain F which has been discriminated in some way or another from all other F’s. Being the only F around is one way, being the only attention-getting F is another, and being pointed at is a third. In the first two cases no demonstration is necessary because the relevant F is already singled out. And since the speaker’s intention plays no role in singling it out, the mere fact that the actual demonstratum is the intended one does not establish the primacy of intentions over acts of demonstration. As for the third case, where being pointed at is what distinguishes the relevant F, Reimer finds it plain that it is the act of demonstration that determines the referent. The distinctive F is the one being pointed at.
Next Reimer considers a case which combines the second and third ways of being distinctive. Fido distinguishes himself from the other dogs around by barking raucuously, so that one could without pointing use ‘that dog’ to refer to Fido—and only to Fido; but one could point to another dog, Spot, and thereby be referring to him. Here the act of demonstrating Spot overrides the prior distinctiveness of Fido, strongly suggesting that what determines reference is the demonstration—the intention doesn’t really enter in. As Reimer argues, although uniqueness or distinctiveness makes the act of demonstration unnecessary for securing reference (for ‘determining a demonstratum’), when there is a gesture it determines the demonstratum—even if the demonstratum were already unique or otherwise distinctive. Since the gesture is controlling, it is semantically significant. It, not the speaker’s intention, determines the referent.
Reimer takes up two apparent counterexamples to her view. In the first case there is no gesture but the actual referent (‘demonstratum’) seems to be the intended demonstratum rather than what is clearly discriminated by the audience. For example, Spot might be barking raucously, thereby capturing the attention of everyone nearby, but the speaker, who without pointing intends to refer to a dog he sees in the distance, utters, ‘That dog looks just like Fido’. It would seem that the reference here is not to Spot but to the not-so-salient dog in the distance. In the second case the speaker inadvertently or carelessly fails to demonstrate the intended referent. For example, the speaker might intend to be pointing at Fido, but Spot and Fido might be frolicking about in such a way that it is unclear which dog is being pointed at. Reimer concedes that insofar as there is a demonstratum, it is the dog that the speaker is attending to, but, she argues, what is decisive is not the speaker’s intention but his attention. The same is true in the first case, though here we must distinguish between the dog the speaker is actually talking about and the one the audience takes him to be talking about. What matters in regard to what is actually said is the speaker’s point of view, not the audience’s, but again, according to Reimer, what matters is the speaker’s attention, not intention.
To clinch her case, Reimer alters the example in which the speaker is not clearly pointing at the intended dog to one in which he is clearly not pointing at the intended one: he intends to be pointing at Fido but is clearly pointing at Spot. He is in fact referring to Spot and if he said, ‘That dog is Fido’, he would be speaking falsely, saying of Spot that he is Fido. The explanation for this, according to Reimer, is that what is controlling is not the intention but the act of demonstration.
All of the above examples are meant to support Reimer’s contention that demonstrations accompanying utterances of the form ‘that F’ are semantically significant. Her position may fairly be summed up as follows. The demonstratum (referent if there is no act of demonstration) is the F that is distinctive under the circumstances. If there is no demonstration, this F is distinguished by being the only F around or by being the maximally salient one; if there is a demonstration, the relevant F is the one being demonstrated, even if there is another one that was maximally salient prior to the demonstration; if there is a demonstration but it is not clear what is being demonstrated, the relevant F is the one the speaker is attending to. In no case does the speaker’s intention play an essential much less decisive role in determining reference, and in no case does it override the act of demonstration; even when, as is usual, the actual demonstratum is the intended one, what makes it the demonstratum is not the fact that it is intended to be. Ergo, acts of demonstration, not intentions, are semantically significant.
I believe this conclusion to be hasty. The basis for it is that intentions never override demonstrations and demonstrations are never overridden by anything else, such as uniqueness or maximal salience, which could determine reference in lieu of a demonstration. As I hope to show, however, the trouble with Reimer’s argument—and with her analysis of her examples—is that she fixes on the wrong intention. She overlooks the specifically referential intention involved in demonstrative reference.
Consider Reimer’s last example. If you intend to demonstrate Fido but in fact demonstrate Spot, you end up referring not to Fido, as you intended, but to Spot. Have you thereby failed to refer to the dog you intended to refer to? Yes and no. Yes, with respect to the intention to demonstrate Fido, for Fido is the dog you intend to refer to; no, with respect to another intention that you have as well: to refer to the dog you are demonstrating. But the latter intention, I contend, is the specifically referential one, the one which you intend and expect your audience to recognize and rely on in order to identify a certain dog as the referent. The audience is to identify what you are using ‘that dog’ to refer to by thinking of Fido not as Fido but as the dog you are pointing at. Surely you do not intend your audience to recognize your intention to refer to Fido; if it were, then saying ‘That dog is Fido’ could not be informative. So it is unsurprising that this intention can be trumped by a demonstrative gesture. However, the relevant intention, your intention to refer to the dog you are demonstrating, is not overridden by your act of demonstration. Quite the contrary, it is consummated by that act. In respect to that intention, you say what you intend to say in uttering ‘That dog is Fido’, namely, that the dog you are pointing at is Fido. Of course you wouldn’t say that if you realized that you are pointing at Spot by mistake.
The relevant intention here, the specifically referential one, is part of a communicative intention, a reflexive intention whose distinctive feature is that ‘its fulfillment consists in its recognition’ by one’s audience, partly by supposing (in Gricean fashion) that the speaker intends his intention to be recognized (, p. 15). This supposition is licensed by a ‘Communicative Presumption’, that utterances are issued with a communicative intention that is recognizable in the context (, p. 7). This presumption constrains the audience’s search for the speaker’s communicative intention. As part of that intention, a referential intention isn’t just any intention to refer to something one has in mind but is the intention that one’s audience identify, and take themselves to be intended to identify, a certain item as the referent by means of thinking of it a certain identifiable way (, pp. 49-53). Such an intention is goes unfulfilled if the audience fails to identify the right individual in the right way, that is, the one intended in the way intended. This is what happens in the above example.
Now let us take up in order Reimer’s other examples and consider how, in connection with the referential intention, the act of demonstration or other sources of salience help determine the referent. As we will see, in each case the speaker intends the audience to think of, by way of taking the speaker to be intending and expecting them to think of, a certain dog in a certain identifiable way. Reference succeeds when the audience does think of the right dog in the right way. The fulfillment of the referential intention requires that in the circumstances there actually be a unique way of thinking of a particular dog and that the dog thus thought of be the intended one. The successful use of ‘that dog’ may require that a certain dog be pointed at, but being the only one or being the most conspicuous one (or even having just been mentioned) may suffice. Pointing at what you have in mind may be necessary to bring it to your audience’s attention—and to put them in a position to take you to be intending this—but often the intended referent is already an object of mutual attention.
In the examples where there is no act of demonstration and a certain dog stands out by virtue of being the only one or the only distinctively conspicuous one, the speaker uses ‘that dog’ with the intention of referring to the dog that thus stands out. In forming the intention to refer to a certain dog, the speaker exploits uniqueness or maximal salience, intending and expecting his audience to recognize his intention on that basis. In the examples where the act of demonstration singles out a certain dog, even when some other dog was antecedently the most salient, that dog becomes the salient one by virtue of being pointed at. This is the basis on which the speaker intends his audience to identify the referent. Given that it is commonly understood that we point at things to call them to one another’s attention, the audience cannot but take the speaker to be calling to their attention the object he is pointing at. When the intended referent is not already distinguished, either because a different F stands out or because several F’s do, pointing at it distinguishes it. The fact that a certain dog is being demonstrated is no different in kind from the fact that a certain dog is barking raucously, is much larger than the others, or was just mentioned. To be sure, the fact that a certain one is being pointed at overrides any antecedently distinguishing fact, but this is so only because pointing is an intentional act standardly performed to direct others’ attention to something. That makes it significant, but only pragmatically, not semantically. An act of demonstration does not determine reference in the sense of making it the case that a certain dog is being referred to. Rather, like any other source of salience, it enables the audience to determine (identify) the referent.
Now consider the two cases Reimer regards as merely apparent counterexamples to her view. In one case the speaker, who is not gesturing, refers not to the dog capturing the audience’s attention but to the not-so-salient dog in the distance. Reimer is correct about the lack of communicative success in this case but attributes it to the discrepancy between the speaker’s and the audience’s focus of attention. However, she does not take into account the reason this discrepancy is relevant, namely that the speaker, in using ‘that dog’ in the context of an act of communication, intends to refer to a certain dog presumed to be readily identifiable in the context. This intention is not fulfilled because a certain other dog is the focus of the audience’s attention and the speaker has done nothing to change that, such as point to the dog in the distance. In the second case, where it is simply unclear which of two dogs the speaker is pointing at, the speaker intends nonetheless to be referring to the dog he is pointing at. He does not succeed at this, precisely because there is no unique dog that he is pointing at. He may be intending to refer to Fido, but the relevant intention, his audience-directed intention, cannot be fulfilled, since nothing qualifies as the dog he is pointing at. The only reason it matters that his attention is on a certain dog is that he expects his audience to be attending to it and thereby to take him to be doing so as well. So imagine there were a third dog which only the speaker could see. Even if the speaker were attending to the third dog, this fact, assuming it had no bearing on what he intends his audience to take into account (he knows no one else can see that dog), would be irrelevant to which dog he is referring to.
Reference is not determined by acts of demonstration or any other features of the context of utterance. Rather, such features are exploited by the audience to ascertain the reference, partly on the basis of being so intended. Referential intentions determine reference, inasmuch as they are the component of communicative intentions that concerns what an utterance is about. Now to say that referential intentions determine reference may seem to suggest that they succeed by magic or are somehow self-fulfilling. However, such a suggestion misconstrues their role. You cannot utter any old thing and gesture in any old way and expect to be taken to be referring to whatever you have in mind. You do not say something and then, as though by an inner decree (an intention), determine what you are using the expression to refer to. You do not just have something ‘in mind’ and hope your audience is a good mind-reader. Rather, you decide to refer to something and try to select an expression whose utterance will enable your audience, under the circumstances, to identify that object. If you utter ‘that dog’ and the dog you intend to be referring to is the only one around or is maximally salient in some way, you won’t have to do anything more to enable your audience to identify it. Otherwise, you will need to point at it. In so doing, you will be intending to refer to the dog you are pointing at. But being pointed at is just one way of being salient and, like the other ways, is not semantically significant.
San Francisco State University
San Francisco, CA 94132, USA
 Kent Bach, Thought and Reference (Oxford: Oxford University Press, 1987).
 Kent Bach and Robert. M. Harnish, Linguistic Communication and Speech Acts (Cambridge, Mass.: MIT Press, 1979).
 David Kaplan, ‘Afterthoughts,’ in Themes from Kaplan, edited by Joseph Almog et al. (Oxford: Oxford University Press,1989), 567-614.
 Marga Reimer, ‘Do demonstrations have semantic significance?’, Analysis 51 (1991) 177-83.
I won’t dwell on this point, but just what is at issue here ultimately depends on what is meant by ‘semantically significant’. There is a whole literature on whether Donnellan’s referential/ attributive distinction, which applies to definite descriptions, is semantically significant (in my view, , ch. 6, it is not). All of Reimer’s examples involve utterances of sentences containing demonstrative descriptions, sentences of the form ‘That F is G’. In such cases the linguistic meaning of the sentence does not fully determine what is said or, in particular, which F ‘that F’ is being used to refer to (the F which is being said to be G).
I won’t quibble here (I do in , pp. 176-9) about the question of whether demonstrative and other so-called referring expressions themselves refer or whether just their users refer (‘use an expression to refer’ is ambiguous in this regard). However, it is my view that there is no fact of the matter of what a demonstrative phrase refers to beyond what the speaker is using it to refer to and what he could be reasonably expect to be taken to be using it to refer to. What the audience actually takes him to be using it to refer to is a matter concerning communicative success.
As we all know, small children (and even some adults) are too egocentric to take others’ perspectives into account. They do not realize that objects of their attention may not be objects of others’. So they often refer to things that others are in no position to identify.