Introduction
In this essay I will outline Perry's view of Essential Indexicality and his argument for this view. I will then explain how Cappelen and Dever oppose Essential Indexicality. I argue that their case against indexicals fails because they only throw inconclusive doubt on the idea that indexicals are necessary for action, and give us little in terms of an alternative explanation.
Perry’s view of Essential Indexicality
Perry presents some interesting thought experiments aimed to show that indexicals are essential to action. It seems like action is impossible without reference to indexicals like “I”, “here”, “now”, “this” etc. According to Perry it is impossible to replace indexicals with proper names because they are not essential for action in the same way. Another claim made by Lewis is that indexical reference is essential in gathering information - what he calls de se information cannot be gathered without the first person perspective (Cappellen & Dever, 2017: 75). I will call this Indexical Information. There is powerful evidence for both these claims and Perry's thought experiments will show us why.
In this first case he is showing the importance of the indexical “I” or “Me”. He uses himself in the story about a shopper who is walking around in a supermarket. He discovers that someone is leaving a trail of sugar on the floor and understands that they must have teared up a bag in their cart without noticing it. He is trying to find the one who is making a mess but notices that at each turn around a corner the trail of sugar becomes thicker. Suddenly he realizes that it is he who has a torn bag, when he see this he adjust the bag of sugar in his cart and clean up the mess. For Perry to start cleaning up he must know two things; both that Perry is making a mess and that he is Perry (Perry 1979:1).
In order to act on his knowledge, Perry needs two kinds of information, information about the fact of the situation and information about how he is connected to these facts. Only knowing that someone is making a mess is not sufficient to cause him to act, he must also see the connection this information has to himself. Even if he knows that it is Perry who is making a mess, he will not realise that the mess on the floor is caused by himself unless he knows that he is Perry. This first person reference seems necessary for agency to take place (Cappelen & Dever, 2017: 72).
Here is a second case where Perry uses the indexical “now”. A professor wants to go to a department meeting that starts at noon. He is sitting in his office without moving when the time arrives. Suddenly he starts moving to go to the meeting. What explains this action? It was a change in belief, he knew all along that the meeting starts at noon, but at some point he makes the connection between “noon” and “now”. Again as in the example above, knowing that the meeting starts at noon is not enough to cause the professor to act, the indexical information “it is now noon” must be connected with this knowledge. These powerful and vivid examples have lead many philosophers to believe that agents need information in an egocentric form in order to act on what they know (Cappelen and Dever, 2017: 72).
The Sceptical Reply
The sceptics about essential indexicality claims that the first person view or any other indexical attitude is unnecessary for agency. In other words it is possible act without any indexical information (Cappelen & Dever, 2017: 72).
According to Cappelen and Dever there is weak inductive support for the view that indexical reference is needed for action. They admit that Perry's examples show that indexical reference can lead to action, but deny the stronger claim that it is necessary for action. On this view Perry's examples only proves that we cannot make the universal claim that indexical reference never leads to action, but this do not prove the stronger claim about it being necessary for action. The task for the sceptic then is to show how action can happen without indexical reference (Cappelen & Dever, 2017: 73).
Perry’s fax machine example is an attempt to show that action can not happen without indexical reference. If I want to press the buttons on the machine it is not enough to know where the buttons are relative to each other, I must also know where they are relative to me and my fingers to be able to push them. To act in the world our bodies need to be in contact with other objects around us, therefore it is necessary that we are able to locate our bodies relative to the external objects. I think this is a strong argument in favor of indexical reference (Cappelen & Dever, 2017: 73).
However the sceptic has two objections. The first is that even though it is true that we need to bring our bodies in contact with external objects in order to act, we can do this without representing ourselves in relation to these objects. Somewhere in the structure of my physical architecture the relation between me and the buttons are established, but it does not have to be known in a representational way. On the cognitive level it is just all about the buttons and their objective position in space, then neurons fire in my brain and I end up pushing them in the right way. If there is no representation of the fingers, then there is no indexical representation of the fingers (Cappelen & Dever, 2017: 73).
This seems very strange to me, how can I possibly know that I can press the buttons if I have no knowledge of their location relative to me? In order to push the buttons it is not enough to know where in objective space the buttons are located, if I am unaware that my fingers belong to me or that they even exist, how could I use them to push the buttons? Can I use something that I am not even aware exists? Whether I consciously, unconsciously or sub-personally need to represent that knowledge in order to have it seems to be a separate issue Cappelen and Dever takes into the discussion. Perry merely talked about knowing.
The second reply from the sceptic is that even when I do locate my fingers relative to the buttons, I do not have to think about them in a first-person way - I do not have to refer to them as my fingers. I can just refer to my name in a third person way, Camilla's relation to the buttons, I do not even have to know that I am Camilla (Cappelen & Dever, 2017: 73).
This also seems strange, and relies on the possibility of representing oneself in a non-first-person way. Cappelen and Dever still need to show how to make such representation intelligible. And if we stick to the knowing instead of talking about representation, it is even less clear how it is possible to know oneself in a non-first-person way, which is closer to Perry’s original thought.
Cappelen and Dever sums up the skeptical argument by making two points: First, our bodies position in space need not be mentally represented in order to act, and second, if the body is represented relative to the environment it does not have to be indexically represented.
Both points rely on a discussion about how to interpret knowing in Perry’s original example, and even if the knowing is interpreted in representational terms, it will need to be fleshed out with a full-fledged representational theory of knowledge. Perry’s thought experiment is prior to a theory of knowledge, in the sense that it is exactly the kind of thing such a theory should account for, how we know ourselves in relation to the actions we can take in an environment. By excluding the information we have of ourselves in relation to the actions we can take in an environment, their theory of knowledge fails to account for our most obvious experience of knowing things from the first person perspective. The only way this strategy can show merit, and justify such exclusion of phenomena, is if their general theory of knowledge is spectacularly successful in accounting for knowledge more generally.
Skeptical Reduction of Essential Indexicals to the Problem of Opacity
Indexicals can bee seen as a kind of opacity because the use of the indexical “I” for example, cannot be replaced with my name. If someone else says “I” it does not refer to me anymore but to them. So indexical opacity means that indexicals cannot be substituted without changing the truth value in action-explanation contexts. The skeptic about indexical reference sees this as a general instantiation of generic opacity - that co-referential terms cannot be substituted without changing the truth value in action-explanation contexts (Cappelen & Dever, 2017: 74).
To illustrate this point Cappellen and Dever describes a similar case to Perry and the torn sack in the supermarket; I believe that Clark Kent is making a mess, I keep passing by Superman but I cannot find Clark Kent. At last I realize that Superman is Clark Kent. I believed all along that Clark Kent was making a mess but I did not believe that Superman was making a mess. When I start believing that Superman is Clark Kent I stop looking for Clark Kent and ask Superman to clean up the mess. My change in behaviour is explained by my change in belief.
The sceptics take this example as proof that when explaining actions in general, it is not possible to substitute co referential singular terms. We can see Indexical Opacity as a case of Generic Opacity and then there is nothing special about indexicals that the arguments against Generic Opacity cannot explain (Cappelen & Dever, 2017: 75).
It is difficult to reduce Indexical Opacity to Generic Opacity when generic opacity relies on the lack of connection between certain first-person beliefs. For instance, in order to realize that Superman is Clark Kent, I need to be able make the connection between my beliefs about Superman and my beliefs about Clark Kent. What information do I get when I learn that Superman is Clark Kent? As both refer to the same person, the best candidate is indexical information, i.e. information that cannot be gathered without the first person perspective.
For instance, someone can tell me that Superman is Clark Kent and I still might not believe this fact. I can say “I do not believe it unless I see Clark Kent flying”. It does not matter what the facts are, unless I am in the believing relationship with this fact I will not know that Superman in Clark Kent. I need Indexical Information of the sort “I believe that Superman is Clark Kent” before I can know this fact.
Indexical information
According to Lewis there is knowledge that can only be obtained from the first person perspective. He calls this information “self-locating” or de se information. We can only access this information through indexical reference (Cappelen & Dever, 2017: 75).
According to the sceptic about essential indexicals, this is also just a case of opacity - there is no information that can only be gained by indexical information. It is just another case of the Frege puzzles. The only problem is to explain opacity and that has nothing to do with indexicals or self-locating beliefs (Cappelen & Dever, 2017: 76).
I on the other hand think the the problem of Opacity exists because of the Kripkean denial of the first person view. The attempt to fight indexicals with the same weapon as used against opacity, seems impossible. You cannot say that indexicals are just an error in our reasoning because it is obvious to me that we always start with indexical information in all our reasoning. If you cannot trust indexical information than how can you trust any of your knowledge? A theory of reference that severs all connection with our most immediate experience makes no sense to me.
Cappelen and Devers explanation of behaviour
If new indexical information do not explain my change in behaviour when I realize that I am the one making a mess in the supermarket, than what is it that makes me suddenly start cleaning up after myself? Cappellen and Dever tries to explain this by the case of Nora and Clara. They are walking in the woods when Nora is attacked by a bear. Nora climbs up in a tree while Clara calls for help. As Cappellen and Dever puts it: “We can imagine that they have the same non-indexical beliefs and desires” Then why do Nora and Clara act differently? The tempting answer is to say that Nora thinks to herself “I'm being attacked by a bear” while Clara has no such thought (Cappelen & Dever, 2017: 77).
I think this is obviously the relevant difference that explain their difference in behaviour. Cappellen and Dever however tries to explain this by saying that even if in some cases indexical belief makes it rational for A to act in a way that is not rational for B, it does not follow that there is something general about the connection between indexicality and agency. All kinds of difference in belief can make a difference to what is rational to do. All such situations are opaque and we can see this as just another case of opacity. Again there is nothing special about indexicals (Cappelen & Dever, 2017: 757).
An agent have infinitely many third-person beliefs about the world that will not result in action so Cappellen and Dever now needs to explain which beliefs explain action and which does not. A sort of selection mechanism. The de se solution to the selection problem is that the first person status is the selecting mechanism. Only first person beliefs and desires give rise to the right sort of intentions and actions. Cappellen and Dever needs to solve the selection problem without appealing to de se states (Cappelen & Dever, 2017: 77).
Their proposal is the Action Inventory Model. An agent has a large variety of third-person beliefs and desires, this gives rise to intentions and actions. Not all of these intentions will give rise to action because the agent has an “action inventory” which is a variety of actions that can be executed. The agent seeks to match his intentions with his action inventory and when he finds a match, actions happen. Without a match the action will not happen.The selection problem is solved by appealing to psychological or physical limitations of the agent. The reason that some action happens is because it was available to the agent at the moment when it would fulfill his intention (Cappelen & Dever, 2017: 78).
I have trouble understanding how the concept of an “agent” is possible on this picture. The action inventory is available to an “agent” who “seeks to match his intentions with his action inventory” (Cappelen & Dever, 2017: 78). But how can the agent know that this is his beliefs and desires when he only have access to third person beliefs and desires? “Someone wants ice cream” for example. If this is true then the agent does not have any knowledge that this is his believes and desires, so he has no reason to act on them. This “agent” cannot possibly know that he exists, only that beliefs and desires exists. Why he would act on these third person beliefs and desires are still a mystery to me.
Ultimately Cappelen and Dever have to appeal to the notion of what is available to the agent at any given moment (to match actions to inventory). But what does it mean that something is available to the agent? What frame of reference decides availability, and what aspects of the agent himself and the environment is available to the agent? These are questions that deserve a rigorous answer, and must be confronted by Cappelen and Dever if they decide to abandon the use of indexical information as a way to explain and determine availability.
Why I Believe Indexicals are Essential
Indexicals seems to give us a centre, a firm point where all information is processed. In one sense indexical information is the most secure and undeniable information we have. There seems to be a special certainty in the utterance “I am”. “I am not” is a contradiction because in order to even say this sentence you must exist. In the same way with the indexical “now” I cannot possibly doubt the fact that it is always now. From a first person experience, indexical sentences seems to be the only knowledge that cannot be doubted when it is uttered. This was the discovery of Descartes when he tried to find out what knowledge in himself that was most trustworthy; “There is a deceiver of supreme power and cunning who is deliberately and constantly deceiving me. In that case I too undoubtedly exist, if he is deceiving me; and let him deceive me as much as he can, he will never bring it about that I am nothing so long as I think that I am something” (Descartes in Chalmers 2002: 10).
I find it peculiar how Cappelen & Dever seems so determined to deny that which cannot be denied, the existence of the first person view. I find it hard to understand the motivation behind this extreme insistence on explaining all cases of reference with “external” or “physical” information. How can they hope to explain reference by denying that we are part of it? This stance forces them to explain away our most obvious experience of being.
Conclusion
In this paper I have presented Perry's argument for Essential Indexicals and Cappellen and Devers attempts to explain them away. I conclude that their efforts only throw inconclusive doubt on the idea that indexicals are necessary for action, and give us very little in terms of a plausible alternative explanation.
References
Cappelen and Dever (2017). Puzzles about Reference: Contemporary Introductions to Philosophy of Language. Oxford University Press.
Chalmers, D. J. (2002). Philosophy of mind, classical and contemporary readings. Oxford: Oxford University Press.
Perry, J. (1979). The problem of the essential indexical. Noûs 13: 3-21. 2nd edition. Stanford: CSLI Publications, 2000.