Download Beliefs & Abilities: Understanding Self-Attribution and more Lecture notes English in PDF only on Docsity!
393
Notre Dame Journal of Formal Logic Volume 27, Number 3, July 1986
His Burning Pants
THOMAS McKAY*
/ John Perry has helped us to see more clearly the special nature of some beliefs about ourselves.^1 When I look over and see a man with his pants on fire, I am alarmed. I shout, "Drop and roll!" But my alarm becomes panic, shock, or dropping and rolling when I realize that I am looking in a mirror. When I first looked in the mirror, I formed a belief about someone —in fact, about me. But I acted as I might toward almost anyone but me. My sub- sequent discovery of the mirror led to a belief about myself that motivated a par- ticular kind of action —action that is appropriate for me to preserve myself rather than action of the sort that I might do in the attempt to help someone else. We shall call the second more intimate type of belief about myself "self- attributive" belief. This is not a particularly good name, because before my big realization I saw myself in the mirror and attributed the property of having burn- ing pants to me. I just didn't "realize" that it was me. Getting clearer about the nature of this realization will be the point of this paper. But meanwhile we need a name for the special beliefs about the self that result from such realizations, and although "self-attributive" is not very good, neither is anything else. Perry's discussion has made it clear that no purely descriptive belief, of the sort (B) The man with property F has burning pants, can be identified with self-attributive belief. Unless F is filled out by expressions involving " I " , "here", or "now", there is always the possibility that one can accept (B) and yet not accept (C) I have burning pants (at least for a short time). My self-attributive belief is not reducible to any descriptively characterizable belief lacking " I " , "here", or "now".
T h e writing of this paper has been supported by a grant from the Syracuse University Senate Research Fund.
Received October 15, 1985
394 THOMAS McKAY
In [1] Barwise and Perry have developed a semantical apparatus that they use for characterizing belief-states (pp. 83-85). In particular, they define an indi- vidual role /, intended to be used in characterizing self-attributive beliefs (pp. 85-87, 259-260). In this paper I will go a few steps beyond what is said in [1] to explore what is involved in self-attributive belief. It seems that Barwise and Perry have the material necessary for giving a sound and systematic account of self-attributive belief, but they never bring it together to do the job.
2 Many aspects of the phenomenon of self-attributive belief also occur when there is no self-reference.^2 Barwise and Perry give a fundamentally sound and simple account here. Consider the familiar case of Pierre, who has beliefs about a place he calls "Londres" and about a place he calls "London". In particular, he sincerely asserts (1) Londres is pretty. (2) London is not pretty. But since his uses of "Londres" and "London" refer to the same city, it appears that he has contradictory beliefs about that city. Yet (the story goes on) he has made no logical or conceptual error in having these contradictory beliefs. Barwise and Perry have introduced the systematics for representing Pierre's state of mind and saying clearly what goes wrong and what is all right in Pierre's belief. The relation Br relates a believer to a belief-schema. A belief-schema (an event-type or "disjunction" of event-types) is a construct of individual roles and properties. Thus Pierre's belief-state can be characterized in the following way ([1], P. 252): in e 0 at /: Br, Pierre, _Eχ{a)_ yes Br, Pierre, E 2 (b); yes of, a, London; yes of, b, London, yes where in Ex : at lx: pretty, a; yes in E 2 : at _12 _ pretty, b; no. Since event-types _E_ and E 2 are jointly possible, this is a cognitively coherent situation. Pierre could have been so situated that events of types Ex and E 2 both occurred. But the relation Br and the event-types it relates Pierre to are only the internal (cognitive) part of the belief-story. There is also an external component. For the belief to be correct, there must be an anchoring for the roles that makes it true that events of types Eλ and E 2 occur. But, in the case of Pierre's beliefs, the two roles a and b are anchored in the same individual, London. Events of types Eγ and E 2 with that anchoring cannot co-occur, thus the beliefs are exter- nally incoherent even though they are cognitively coherent. The doxastic con- ditions (the Br relations and associated event-types) are coherent, but they are
396 THOMAS McKAY
ment seems like an inappropriate metaphor. The Barwise and Perry concept of an individual role is better seen as a marker of the fact that the doxastic state characterized is linked to a particular network. Being linked to that network is what constitutes a belief's being properly characterized as involving that role. A property is a mode of recognition of an individual by being linked to such a network. The metaphor of containment can only conceal the mutuality of these relationships among roles, beliefs, and modes of recognition. Cases (such as Pierre's) of unmerged roles anchored in a single individual make the account of actions more complicated than it might otherwise be. One might ordinarily be inclined to generalizations something like this one connecting belief and action: If (i) individual S wants to V some F 9 (ii) S believes x to be F, then $ will try to V x (if S has nothing better to do).^4
But clearly this will not work. If Pierre is near London and he wants to be in a pretty city, this will not motivate him to act even though he believes it to be pretty (etc.). Pierre must connect his ability to act (his proximity to the city) with his belief that the city is pretty. But if his ability is recognized by him only in connection with his concept London and his belief that the city is pretty is asso- ciated only with his concept Londres, then Pierre will fail to act. This failure occurs even though Pierre believes it to be pretty and also believes of it that he is near. If Pierre merges his concepts he will not have these problems. But merg- ing is more than we really need. We need for him to apply the relevant concept to the action at hand ([1], pp. 248, 251). If his motivating desire is to be in a pretty city, then he must: (i) have a belief of some city that it is pretty, and (ii) he must be able to apply the concept that plays a role in the belief to the indi- vidual with respect to which he has the desire. Some single concept must play a role in both his belief and his desire.^5 This is less than merger. (For example, Pierre might apply his Londres concept if he has a French-language map of England and is near London, even if he fails to merge the concepts London and Londres.)
3 In applying this to the first-person case, we can say, as a first approxima- tion, that the same ideas should work. When John sees the man with his pants on fire, he first has a belief the character of which is formulated by the event-type in E: at /: pants on fire, a; yes.
When he "finds out that it is his own pants", he applies the belief. In this case, he merges a with his "self-concept" /. Having done so, he applies E by connect- ing it with a desiderative state E'
in _E'_ at /: avoids burning, i; yes and a doxastic state in _E"_ at /: best avoids burning (when pants are on fire) by dropping and rolling, /; yes.
HIS BURNING PANTS 397
This way of dealing with self-attribution has one immediate and obvious virtue. The fact that the individual role / figures in a belief does not require that there be some privileged mode of presentation associated with / that is a com- ponent of the belief representation. Like the individual roles London and Londres, the role / can be associated with a network of representations, requiring no privileged mode of presentation for any particular thought involving that role.^6 But we should still be a bit puzzled about the cognitive role /. What makes beliefs involving / especially motivating? Note that the answer just given at the opening of this section will not really do. Even if John had not merged a with i in this case, he might have had a desiderative state E* corresponding to E' and a doxastic state E** corresponding to E"
in E*: at /: avoids burning a: yes in E**: at /: best avoids burning by dropping and rolling (when pants are on fire), a; yes
But we have not answered the real demand to say precisely when E' and E", involving /, motivate John in a special way that E* at E** do not. To meet that challenge, we will do best to consider some simpler desires and actions. Suppose that I see a cookie on the table, and I want to eat it. I pick it up and eat it (if I have no special reason not to). I might do this because I am in a particular doxastic state:
in E: able to eat a cookie by picking it up and putting it in the mouth, etc., /'; yes.
But such a doxastic state explicitly entertained would be rare. And this doxas- tic state need not play even an implicit role in motivating my action. Much of what I do is an immediate response to beliefs about what is around me and desires concerning those individuals. No mediating belief about my abilities is needed. Rather, I just respond by picking up the cookie. It seems clear that at some time in our lives we are capable of such responses without the correspond- ing beliefs about our abilities. I suggest that this is a continuing feature of our abilities to act. The belief that I have the ability does not need to play a role in my exercise of the ability. Of course standards of belief attribution are notoriously slippery. One might be tempted to go as far as to say that the fact that I acted in the way I did in picking up the cookie is enough to show that I have the belief that I am able to do that. Young infants who reach out to pick up cookies simply have such beliefs "wired in". But even if we were to take this more liberal position about belief attribu- tion, we would still have to recognize an important distinction between self-attri- bution and attribution to others. Whether we say that I have an immediately available ability to pick up appropriately nearby objects when I want them (with- out requiring a belief that I have such an ability to satisfy the desire), or whether we say that I have a wired-in belief that I have such abilities to fulfill my desires, my relationship to my ability to fulfill my desires is different from my relation- ship to other individuals' abilities to fulfill their desires. My beliefs about what is around me are immediately available to serve my desires. I don't need to
HIS BURNING PANTS 399
Nonetheless, it seems that Barwise and Perry's general apparatus of indi- vidual roles can be employed to mark the way in which beliefs get applied by being linked to other beliefs. And we can extend that apparatus to recognize a special individual role / that is distinguished by its links with immediate abili- ties to act in response to desires. We should not think, however, that the special doxastic role of / corre- sponds to a special semantical role for " I " and "my". I come to an important realization in accepting as true "My pants are on fire" and not just "His pants are on fire". There is a major difference in my doxastic condition, because the property of having burning pants is tied in to a substantial network character- ized by the individual role / and directly linked with my ability to drop and roll. But this property is still playing a role in a belief about a particular person — the belief that the property having burning pants applies to him. If bystander Smith sees me and notices what he would express by His pants are on fire
and bystander Jones notices and warns
Your pants are on fire
their beliefs agree with mine both before and after my big realization.^6 It has sometimes been thought that "My pants are on fire" expresses a very different belief from "His pants are on fire" when I come to accept it. Given that both express a belief that agrees with a belief held in common by Smith and Jones, I see no reason to say that they express distinct beliefs. The English words " I " and "my" refer to me by a different mechanism than do the English words "he" and "his", but they simply refer and bring to the proposition expressed only the referent. Nevertheless, if I am a normal speaker of English, then the word " I " (or "my") will be used in expressing that belief only after it is tied in with the network of beliefs and special abilities constituting the individual role /. When I accept "My pants are on fire" I will come to have many new beliefs (e.g., that for the first time in my life my pants are on fire, that the pants I bought just yesterday are on fire), but those beliefs are not expressed by "My pants are on fire".
NOTES
- Most notably in [6]. In [1] the special character of such beliefs is considered in several places, e.g., pp. 86-87, 248-249, 259-260.
- Here I follow [2] and the parallel drawn there by Boer and Lycan between attribu- tions of properties to other things and self-attributions. Barwise and Perry draw a similar parallel in [1], p. 249.
- Peacocke, for example, introduces '[self]' for the mode of presentation that he assumes to be part of self-attributive beliefs. See [5], p. 108.
- The clause 'if S has nothing better to do' is obviously a feeble attempt to take account of the fact that S has to choose from among the things he wants to do. The problems in clarifying this are irrelevant to the points under discussion here.
400 THOMAS McKAY
- This is also suggested by Sosa and Pastin in [4]. Their suggestion is refined by me in [3].
- We are not forced into the Fregean view that sentences like "My pants are on fire" express a proposition that only I can entertain, i.e., a proposition containing my unique mode of self-presentation as an element.
REFERENCES
[1] Barwise, Jon and John Perry, Situations and Attitudes, MIT Press, Cambridge, Massachusetts, 1983.
[2] Boer, Steven and William Lycan, "Who, me?," Philosophical Review, vol. 89 (1980), pp. 427-466. [3] McKay, Thomas, "Actions and de re beliefs," Canadian Journal of Philosophy, vol. 14(1984), pp. 631-635. [4] Pastin, Mark and Ernest Sosa, "A rejoinder on actions and de re beliefs," Cana- dian Journal of Philosophy, vol. 11 (1981), pp. 735-739. [5] Peacocke, Christopher, Sense and Content, Oxford University Press, New York,
[6] Perry, John, "The problem of the essential indexical," Nous, vol. 13 (1979), pp. 3-21.
Department of Philosophy Syracuse University Syracuse, New York 13210
