REFLECTIONS ON THE INDEXICAL POINT OF VIEW: ON COGNITIVE SIGNIFICANCE AND COGNITIVE DYNAMICS, BY BOJISLAV BOZICKOVIC

Abstract

In accounts of indexicals, we encounter two problems: the problem of cognitive significance and the problem of cognitive dynamics. The problem of cognitive significance leads us to posit finer-grained sense content to account for the explanation of our actions and emotions. Meanwhile the problem of cognitive dynamics calls us to show how two episodes of thought can have the same fine-grained sense content even though they are expressed in different ways in different times and places. Bojislav Bozickovic offers a solution to the problem of cognitive dynamics by using objects to tie together different ways in which a sense content might be expressed, and arguing that the sense expressions might become untied if we come to wonder if we are talking about the same object. I argue that this strategy raises a number of questions, and may lead to destabilization of our thought contents based on past and future concerns about object identity.

Keywords:
Cognitive; Indexicals; Point of View; Significances; Dynamics

I recently wrote a book on the topic of indexical content, learning only after its publication about this excellent then book-in-progress by Bojislav Bozickovic. I wish it had been available earlier, as I would have profited from studying the arguments and ideas herein. Bozickovic’s book is not only a clinic in the application of “close reading” scholarship to analytic philosophy, it is the most complete survey of the literature that I am aware of, and it offers a fresh, well-argued, perspective that deserves to be addressed by the profession.

While Bozickovic and I worked on these topics ignorant of each other’s work, we were not exactly two ships passing in the night so much as we were two ships moving in the same direction while travelling in the night. And while we ultimately arrived at different destinations, it seems to me that we were following the same compass direction throughout. My plan in this review is thus to offer some contrast with my own views, not because I want to usurp some of the attention that Bozickovic’s book deserves, but because I think contrasting the two positions can help illuminate Bozickovic’s project. Or in any case, it is the easiest way for me to provide that illumination.

Before we get to the compare and contrast part, however, we need to review some philosophical debates that are going on in the background here. There are well over a hundred years of philosophy that set the stage for the concerns raised in this book, so we need to review that stage-setting in at least some detail.

The issue on the table, of course, is the topic of indexicals, and the discussion begins with a puzzle about indexicals that was first articulated by Frege over a century ago. On the one hand it appears that indexicals have senses and that the appeal to these senses is necessary for us to explain things involving human action and emotions. This puzzle has been animated by Perry (1979) in a series of series of example. So, for example (and keying off an autobiographical story by Ernst Mach), there is a thought I express with the words ‘I am a shabby pedagogue’ that I do not express with the words ‘he is a shabby pedagogue’, even though I might be looking at a reflection of myself in the mirror. There is a kind of first-person perspectival sense in the former utterance that is not found in the latter utterance. It is only the thought expressed by the former utterance that gets me to buy new clothes and go to the barber. We can say that the two utterances express thought contents that differ in their cognitive significance. Or alternatively, we can say that the utterances themselves differ in cognitive significance.

Perry notes that there are temporal cases of a similar nature. There is a thought that I express when I utter ‘I have an appointment with the dentist today’ that I do not express with the worlds ‘I have an appointment with the dentist Tuesday’, if, for example, I have lost track of what day of the week it is on Tuesday and I think it is only Monday. In the former case, the thought expressed motivates me to get up and go visit the dentist that very day. In the second case I may well spend the day in my house.

On the other hand, there are cases where this indexical content leads to puzzles. Consider the case, discussed by Frege, where there is a thought that I express with the words ‘Today is a fine day’ and then express that very same thought the following day with the worlds ‘Yesterday was a fine day’. We want to say that we are expressing the same thought in these two cases, but are we not deploying different senses by using the words ‘today’ and ‘yesterday’? How can we be expressing the same thought?

These two problems can be described as the problem of cognitive significance and the problem of cognitive dynamics. The problem of cognitive significance leads us to posit finer-grained sense content to account for the explanation of our actions and emotions. Meanwhile the problem of cognitive dynamics calls us to show how two episodes of thought can have the same fine-grained sense content even though they are expressed in different ways in different times and places. The first problem calls us to generate more sense contents. The second problem calls us to collapse sense contents. It is a recipe for philosophical puzzles.

When we take on the problem of cognitive significance and the problem of cognitive dynamics simultaneously, there initially seems to be a limited slate of options for us. One idea, offered by Richard Heck (2002) is to rethink the cognitive dynamics part. The idea would be that we aren’t expressing the same thought across these diverse perspectival positions. We are successfully communicating, but that doesn’t mean we end up having the same thoughts. Successful communication is weaker than thought-sharing. Another option is to give up on the idea of cognitive significance - or at least give up on supposing that senses and cognitive significance have a place in the theory of meaning. This is the idea associated with Almog (2008) and Wettstein (1986). These two different approaches (abandoning either cognitive significance or cognitive dynamics) are philosophically stable, but obviously concessive. Bozickovic’s project is more ambitious, in that it aims to carry through Frege’s leading idea.

Frege’s leading idea, in this case, is to say that the senses are not tied to linguistic expressions on a one-to-one basis, but rather, the same sense is expressed in different ways at different times. In Ludlow (2019), I put the idea this way: Interperspectival content is expressed in different ways from different perspectival positions. The content is interperspectival because it is shared, but it is fine-grained sense content because it explains human actions and emotions. So, for example, it explains why I get up and go to the dentist instead of sitting in my house and it explains why I buy new clothes and go to the barber instead of shaking my head and frowning at the shabby person in the mirror.

Now, the tricky part in this is the cognitive dynamics part, and the question is, how is it that we can have the very same thoughts at different times and at different places, given the fine-grained (and perspectival) nature of indexical senses? One standard answer - the received view in philosophy, I believe - is to say that the shared stuff (the interperspectival stuff) is largely referential content (Russellian propositions, for example), but to allow for some extra ingredient X, character in Kapan’s case, role in Perry’s case, that allows us to capture, to some degree, the shared cognitive significance. In Kaplan’s terminology this is the “linguistic meaning” and it involves understanding expressions like ‘I’ in terms of their roles in a context. Thus, crudely, the linguistic meaning of ‘I’ might be something like “the speaker s, in context c.” A term like ‘yesterday’ might mean something like “the day before d, the day in context c.” This linguistic meaning is then ignored when we evaluate the same utterance in counterfactual situations (in other possible worlds). Alternatively, Perry runs the distinction on the difference between what is believed/said and how it is believed/said. Finally, as Bozickovic notes, there are other accounts of this secret ingredient X, including for example, Perry’s (2001/2012) idea of reflexive content.

Bozickovic believes, and I share the belief with him, that characters and roles and related strategies are not sufficient to capture the cognitive significance of our uses of indexicals. One of Bozickovic’s concerns about Kaplan-Perry accounts is that they break down in cases like the Rip van Winkle case in which we lose track of the day. Bozickovic here draws on an example from Branquinho (1990). Let’s say that shortly before midnight on day 1, Joe thinks a thought which he would express using the words ‘Today is a fine day’. Two minutes after midnight he is asked whether “yesterday was a fine day.” Unsure whether midnight has passed, Joe withholds his assent. What this seems to suggest is that whatever the content of his earlier thought was, ‘Yesterday was a fine day’ is not going to capture it, although by rights, it ought to. The cognitive significance must be come apart from standard automatic realignment strategies involving character and roles.

Bozickovic does, however, make a concession to Kaplan and Perry which I would not, and that is the concession that what is properly understood as “linguistic meaning” ends with characters or roles. It seems to me better to say that linguistic meaning is substantially richer than what characters and roles (and related machinery like reflexive contents) can afford us. Following McDowell (1980), I have spoken of the truth theory for a language as “displaying” different meanings by using different expressions on the right-hand side of the truth conditions. Thus, I would say that the semantic theory is perfectly capable expressing sense contents or interperspectival contents, and that an adequate theory of meaning will explain how it accomplishes this.

Perhaps this amounts to some bookkeeping over what gets to be called “linguistic meaning,” but it does seem to me that the question is more than taxonomic, because we want to know if a semantic theory can successfully deliver a theory of sense. I think that it can, but I don’t believe Bozickovic thinks this. If I understand his proposal correctly, the cognitive significance must lie at least in part outside the boundaries of linguistic semantics - in another chapter of cognitive psychology perhaps. If this is so, then one wonders again if the position doesn’t ultimately collapse into the Almog-Wettstein view.

Apart from details, there is a problem - the previously mentioned cognitive dynamics problem - that stems from having finely grained sense contents of the sort that Bozickovic and I have proposed (this problem is shared by both linguistic and extra-linguistic accounts of cognitive significance). The problem is that if we hang on to fine-grained sense contents, the situation threatens to generate what Bozickovic calls a “proliferation of thoughts,” or if you prefer, it could be called a “proliferation of sense contents.” This is to say that there is a temptation to build a new sense content for each form of sense expression. But this isn’t what Frege wanted, as his passage about ‘today’ and ‘yesterday’ showed, and it isn’t what we want if we want to capture the notion of thinking the same thing at different times and places. What is needed is a way to bundle these multiple sense expressions into a single interperspectival content (my terminology), and the problem we face is how best to do this.

We can think of this problem with the metaphor of a three-ring binder. Each page represents the expression of a single sense content - e.g. one page representing its expression using ‘today is a fine day’ and another page representing its expression using ‘yesterday was a fine day’ - and the question is, what is the three-ring binder part of this metaphor representing? What is holding everything together so that we avoid the proliferation of senses?

My idea was that what holds together a thought expressed on day 1 with ‘today is a fine day’ and day 2 with ‘yesterday was a fine day’ (or alternatively, in the Branquinho case, ‘that day which was either yesterday or is about to be’) is a complex theory which explains why those expressions can sometimes express the same thing. And the reason they express the same thing is they are both part of a theory that explains how we coordinate and use linguistic expressions to explain the actions and emotions of person across different perspectival positions. That is, the thing that is really holding everything together is a theory that explains (on day 1) why I am doing what I am doing or feeling what I am feeling, and similarly explains (on day 2) why I was doing what I was doing or feeling what I was feeling the previous day. So, the theory tracks things by keeping track of the explanations for my actions and emotions and affords a sub-theory about how we articulate perspectivally sensitive theories of actions and emotions for each other.

This is where the ships now diverge paths and proceed to different destinations. Bozickovic wants to bind together the different ways of expressing the thought as well, but he has a different kind of binder in mind. What is his binder? The idea is that we express the same thought about some object o from different perspectival positions, just in case we unreflectively assume that o is held constant across these positions. There is a lot to unpack here, and there are concerns.

The first concern revolves around the idea that an object o is deployed to bind the different expressions of sense. One can see the position potentially collapsing into a direct reference theory of some form. That is, if what holds everything together is the object under discussion, why not use that object as the interperspectival content and forget about the sense content? The answer to that objection is that we still need the sense contents for all the usual reasons - explaining actions and emotions, etc - so we need sense contents. We just need to tie expressions of sense content together somehow, and the object under discussion is the way to do that. We will come back to this.

The second issue is whether this view doesn’t turn the whole “sense determines reference” dictum on its head. It seems we need a prior grasp of our reference to o before we can construct the theory of sense for thoughts about o. This is because the only thing we have to glue expressions of sense content together is the unquestioned identity of the object under discussion. That isn’t really fatal, but it does seem to be giving up a part of the Fregean project.

Something should also be said about the theory of reference necessary to get this idea to work. Reference needs to come pretty cheap. It can’t involve acquaintance, because there isn’t enough acquaintance to go around. We can have sense-mediated thoughts about future events, for example. The thought expressed by ‘Tomorrow is going to be awesome’ depends upon a day with which we have no acquaintance or causal connection yet. To avoid this problem Bozickovic adopts an approach to reference from Hawthorne and Manley (2012) according to which reference basically comes for free, in that no acquaintance or causal connection is required. Your mileage may vary on this approach. My only point here is to note that Bozickovic’s position does appear to be tied to the idea of easy reference - necessarily so, since tying together multiple sense expressions into a single sense content often requires reference to an object that is outside our sphere of acquaintance.

Still, even if reference comes for free (or nearly free), one wonders if this is enough. What about cases in which there is clearly no reference (cheap or otherwise) to anchor our thoughts? Consider the Geach (1967) example of the imagined witch that has Hob, Nob, and Cob distressed, and let’s update that example. Suppose, for example Hob, Nob, and Cob all believe in a witch and they all believe that they are thinking and talking about the same witch. Hob says “I saw the witch today.” The next day, Nob reports this to Cob as follows: “Hob said that he saw the witch yesterday.” Surely Nob is faithfully reporting Hob’s statement, expressing the relevant sense content. Obviously, Nob is expressing that sense content in a different way than Nob did because it is now a new day, but what is binding together these different expressions into a single sense content? It cannot be the object doing this work for us, because there is no witch. This would seem to be one of the dangers relying on an object to bind together the sense content.

There are also concerns with the business about unreflectively assuming that o is constant across cases. To use an example from Bozickovic, let’s say I put a bottle of wine on the top shelf of my refrigerator, saying to myself “this wine is supposed to be delicious.” The next day I open the fridge and look at the bottle and again say “this wine is supposed to be delicious,” without wondering if it is the same bottle or if a trickster has moved it or if I have forgotten replacing the wine in the fridge. In that case, an utterance of ‘This wine is supposed to be delicious’ on Monday and an utterance of ‘This wine is supposed to be delicious’ on Tuesday can express the same thought, but just in case it is the same bottle and I don’t start to wonder whether if isn’t the same bottle. But as soon as I wonder if the bottle has been switched, the sense contents must come apart.

For example, suppose that on Monday I put a bottle of wine in the fridge and say “that wine is supposed to be delicious.” On Tuesday I remind myself what I said and say “Yesterday I thought that wine is supposed to be delicious.” I can do that on Bozickovic’s proposal because it has never occurred to me that there has been a change in bottles. But now suppose that on Friday I learn that my roommate has been changing things in my fridge, drinking things in the evenings and replacing them, so that I come to wonder if it was indeed the same bottle. Do the senses of my earlier two thoughts events now come apart? It would seem they have to because I am no longer in a position to express both thoughts in the same way, as I have doubts about the stability of the underlying referential contents. That is to say, whatever sense content I express on Friday with ‘I thought that wine is supposed to be delicious’, it cannot be the same as both the Monday thought and the Tuesday thought. On Friday I can express the Monday thought, and I can express the Tuesday thought, but I cannot express both at the same time, because the sense contents have retroactively come apart.

There is a corollary to the problem just raised: Are my current thoughts susceptible to future undermining? It would seem so. If the stability of a thought depends not merely on an object, but the unreflective assumption that it is the same object, future reflections may undermine that assumption and thus force the conclusion that I am currently entertaining different thoughts. In other words, if it will be the case that I question whether it was the same object o (e.g., that bottle of wine), it must be the case that I am now entertaining different thoughts, even though it has never occurred to me that a switch has taken place, and indeed, no switch has taken place.

There is an inverse version of this puzzle as well. Suppose that on Wednesday I question whether it is the same bottle in the fridge, but then later completely forget my Wednesday doubts about the identity of the bottle and just unreflectively believe it is the same bottle, and on Friday (recalling only my Monday and Tuesday thoughts) I utter ‘For the third time this week I am thinking this wine is supposed to be delicious’. Is my current belief the same as either of the earlier beliefs? It is hard to see how, given that I was entertaining separate beliefs earlier in the week. Perhaps it requires some mereological union of the two earlier beliefs? If so, does that mean my count of belief events was off? One needs to see details here.

In the preceding concerns I have assumed that the notion of unreflective assumptions about identity is an unproblematic notion, but we can problematize the idea a bit. The general form of the problem is something like this: Reflectiveness is not a stable psychological state. We drift from unreflective states to reflective states and back. Can it really be that sense contents pop in and out of existence as we do so? And since our current states seem to play a role in determining the content of our past state, or at least how they can be described, it seems that the reflective/unreflective drift destabilizes our entire mental histories. I don’t mean to suggest that there is no answer to these concerns. I merely mean to point out that there are interesting puzzles to be resolved here.

Remaining puzzles aside, I want to close by revisiting my earlier remarks about the scope and profundity of this project. Obviously, I have pursued an alternative solution to this problem, but I wish I had had access to this work when I was building out my proposal. My own work would have been substantially better by having engaged the ideas in this wonderful book. I look forward to addressing these topics in more detail in the future, albeit (thanks to this book) from a much deeper and much more informed perspective.

References

• Almog, J. 2008. “Frege puzzles?” Journal of Philosophical Logic 37, 549-574.
• Branquinho, J. 1999. “The problem of cognitive dynamics.” Grazer Philosophische Studien 56, 29-56.
• Geach, P. 1967. “Intentional Identity.” Journal of Philosophy 74.
• Hawthorne, J. and Manley, D. 2012. The Reference Book Oxford: Oxford University Press.
• Heck, R., 2002. “Do Demonstratives have Senses?” Philosophers’ Imprint 2, #2, <www.philosophersimprint.org/002002/>.
  » www.philosophersimprint.org/002002/
• Kaplan, D., 1977. “Demonstratives.” manuscript UCLA. Reprinted in J. Almog et. al. (eds.), Themes from Kaplan Ithaca: Cornell University Press, 1989.
• _____. 1979. “On the Logic of Demonstratives.” The Journal of Philosophical Logic 8, 81-98.
• _____. 1990. “Thoughts on Demonstratives.” In Yourgrau (1990), 34-49.
• Ludlow, P., 2019. Interperspectival Content Oxford: Oxford University Press.
• McDowell, J., 1980. “On the Sense and Reference of a Proper Name.” In M. Platts (ed.) Truth Reality and Reference London: Routledge and Keegan Paul, pp. 111-130.
• Perry, J., 1977. “Frege on Demonstratives.” Philosophical Review 86, 474-497.
• _____. 1979. “The Problem of the Essential Indexical.” Nous 13, 3-21.
• _____. 2001. Reference and Reflexivity Stanford, CA: CSLI Publications. Second edition, 2012.
• Wettstein, H., 1986. “Has Semantics Rested on a Mistake?” Journal of Philosophy 83, 185-209.

• 1
  CDD: 153.4.

Publication Dates

• Publication in this collection
  02 Sept 2022
• Date of issue
  Jul-Sep 2022

History

• Received
  06 Apr 2022
• Reviewed
  12 May 2022
• Accepted
  23 May 2022