Microsoft word - ssppr-final-corrected.rtf

I once stood staring at a map in a large US airport, looking for an ATM. Next to me a couple also stared at the map, trying to figure out where in the airport they were. “Sheesh!” said the male at last, pointing to the red dot and the words ‘You are here’ in the key beside the map: “We’re way over here, right off the map!” Jenann Ismael’s understanding of red dots lies very much at the other extreme, but self-location – the task that couple were engaged in, however haplessly – is the unifying theme of The Situated Self. In Ismael’s hands, it is a thread that turns out to tie together a remarkable range of topics. Imagine being taken on a tour of your own home town, weaving in and out through various neighbourhoods, familiar and less so, and ducking between them via surprising paths and shortcuts.
(“This next to that? Who knew!”) The Situated Self is that kind of book – an exhilarating and often demanding excursion, startling in its breadth and the ways it relates seemingly unconnected topics.
Although the route of Ismael's tour is often highly original, the vehicle itself has a familiar shape, at least in outline. Ismael builds a chassis to specifications due especially to John Perry, fits it out to her own requirements, and then shows us where it can take us. Mostly, I found the excursion convincing, as well as fascinating. In what follows, I shall take up one case in which I was not convinced, and then raise an issue about another destination – somewhere in the same vicinity, apparently – and ask whether Ismael’s bus takes us there, too. First, though, to the fundamentals, as Ismael develops them.
1 Articulation, architecture, and abuse of notation
One of the insights Ismael takes over from Perry is that in some circumstances, we may think about an object, event or property, without explicitly representing What is it to be able to think about an object? One way is to have a thought with a Fregean component that denotes it, but Perry’s suggestion is that there is another way, which he describes by saying that one can have it as an unarticulated constituent of one’s thought. An unarticulated constituent of a thought is an object, event or property reference to which needs to be made in specifications of its truth Ismael illustrates this point with Perry’s Z-landers, who talk about the weather in their location, without specifically introducing a term to refer to that location. The reference to their location is thus an unarticulated component. They think about their location, without specifically representing it. Linking Perry to the literature on situated cognition, Ismael emphasises the economical advantages of unarticulated reference. In general, it pays speakers to do as much as possible “in architecture” – to take advantage of fixed causal and associative relations to the environment, and only to represent explicitly when matters call for it.
Matters call for it, above all, when speakers face the challenge of variation, in the unarticulated component in question: either when either they themselves vary in this respect – e.g., by moving around, where the component is their spatial location – or when they encounter other speakers who already differ, in the relevant respect. At this stage, coordination and mutual comprehension require a viewpoint-neutral representation, or map, of the relevant space of possibilities.
While to keep track of their own positions on such maps – their own circumstances, in the relevant respect – speakers need indexicals, or “red dots”.
Indexical self-location is the theme at the core of Ismael’s book. Like Perry, she characterises indexicals as terms whose semantic values depend on their context of use, and one of her main projects is to reveal the kind of mistakes we are liable to make in philosophy, if we don’t properly understand the role of such There is a general lesson here …. Add functions from contexts to semantic values to your vocabulary, choose symbols of the same graphical or grammatical type that you use to represent objects, allow the abuse of notation that permits you to use them as singular terms in a fixed context, and you are inviting a mistake. If you forget about the abuse, and reason in ways that apply to singular terms, you will conclude that there is an object distinct from anything named or described by your other singular vocabulary, a degree of semantic and metaphysical freedom that it has failed to capture and that it peculiarly cannot be extended to include. The lesson here is whenever you have an object or property, something you assign as a semantic value to a term in a representational medium, that exhibits one or all of the following syndrome of properties, you should look for an abuse of notation. If it seems to slip irremediably through the net of objective representation, if there are contextual constraints on reference to it, or if reference to it is immune to failure, there is likely to be a hidden parameter that relates it reflexively to the context of use. [TSS, p. 153] If these functions from contexts to semantic value are liable to confuse us in this way, wouldn’t we be better off without them? As noted, Ismael has two kinds of answer to this challenge. The first reiterates the economical advantages of unarticulated reference. Indexicals treat their contexts of utterance as unarticulated components, thereby “buying ease of reference”, as Ismael [TSS, p.!168] puts it. The second appeals to the necessity of the indexicals for certain sorts of jobs. Here, the point is a generalised version of Perry’s famous ‘Essential Indexical’. Action needs to break the symmetry of the objective maps, distinguishing particular times, places and individuals as those relevant to the actor herself. Wherever our notation establishes a symmetry by transcending our own viewpoint – whenever language makes a ‘Copernican leap’, as we might say – we need indexicals to anchor the located matters of practical life to detached information provided by the newly transcendent map.
Ismael’s discussion of these topics is fascinating, penetrating, and wide- ranging. One of the main themes of the book is that these questions of indexicality and self-location – broadly construed, in each case – are central, pervasive and poorly understood. About all of this, Ismael is very convincing. But I want to take issue with one aspect of her application of these ideas – perhaps the central application, as she sees it – to the understanding of the self itself. While I agree with Ismael’s positive proposal in this case, I think she makes heavy weather of a challenge she takes to motivate the proposal. I’ll then raise a question about the relation between indexical abuses of notation, in Ismael’s sense, and some other cases in which it has often been maintained that language leads us astray.
2 Self-identity over time
Ismael situates her discussion of the problem of the identity of selves over time against an objection Anscombe raises to Descartes’ Cogito. Anscombe argues that Descartes’ position has “the intolerable difficulty of requiring an identification of the same referent in different “I”-thoughts.”1 As Ismael puts it, the challenge is this: “How do we get from the purely reflexive ‘I’ of the individual thought to the ‘I’ of the temporal continuant? ” [TSS, p. 167] Ismael notes that for other indexicals and demonstratives, there is always a possibility of errors of reidentification and substitution: I can easily be wrong that two ‘today’ thoughts refer to the same day or that a pair of ‘this-train’ pointings refer to the same train because 1G.!E.!M. Anscombe, ‘The First Person’, in Samuel Guttenplan, ed., Mind and Language(Oxford: Clarendon Press, 1975), 45–65, at p.!58.
reindentification and intersubstitution require judgement, and where there is judgement, there is the possibility of error. Indeed, we can think of indexicals as buying ease of reference at the price of difficulty in reindentification and intersubstitution. [TSS, p. 168] But ‘I’ seems different in this respect: But here ‘I’ parts company with other indexicals. One doesn’t need criteria of identity to reidentify oneself over time, or to intersubstitute the ‘I’ of one thought or impression with the ‘I’ of another. And not only are one’s occurrent references to oneself immune to error, one cannot make an error in reidentification. You cannot mistakenly intersubstitute two noncoreferring ‘I’-occurrences. No criteria of identity are employed, and no errors are possible at any stage in identification, reidentification, or substitution. These facts are wholly unique to ‘I’ and capture much of the very special mystery in thought A natural concern here is that this proposal seems to help itself to a solution of Anscombe’s problem. Certainly, one doesn’t need a criteria, and you cannot mistakenly use noncoreferring ‘I’ expressions, but that’s because one is a unified self. All occurrences of “here” at the same place are necessarily coreferring, too, but that does nothing to prevent errors in that case. But Ismael wants to run this example the other way. She notes that creatures fixed to one place would be immune from error in the case of “here”, and suggests that ‘I’ works the same way. Each of us is confined to the same “place” – i.e., roughly, to the same head.
I think that this isn’t entirely convincing. For a community of creatures who each remain fixed in the same place, immunity from error in the case of “here” must come either from the fact that they are all fixed in one place – so that all uses of “here”, across the community, are coreferential – or from having already solved the problem of self-identity over time, so that each of them can distinguish their own uses of “here”. But neither condition applies, surely, in the case of “I” itself? The attempt to appeal to the second condition seems to beg the question.
Actually, I think it is more subtle than this. The problem of reidentification is a diachronic problem, and so a problem that presupposes extension over time. If there were no temporally-extended selves, for whom could the problem possibly arise? This suggests that Ismael is right that ‘I’ differs from other indexicals concerning its vulnerability to this problem, but not for the reason she offers. The real reason is that the problem presupposes the diachronic self: it is the problem of reindentification between two ‘I’-thoughts of such an individual.2 To see this, imagine a case in which there are multiple “natural” standards for reidentification over time. Consider a school which has one classroom for the Grade 1 children, one for the Grade 2 children, and so on; the same classroom, in each case, every year. Each classroom contains mementos of the activities in that room in previous years, and plans of activities for future years. Applied diachronically, terms such as “this class” or “this group” now have two obvious disambiguations. Used in the Grade 3 classroom, for example, they can pick out a particular group of childen (those who are in Grade 3 this year, Grade 4 next year); or the sequence of Grade 3 classes, containing different children each year.
“We have big plans for this class next year” can easily be read in either way. Each disambiguation gives rise to a different form of the problem of diachronic reidentification – but on each, it is analytic that a class’s own uses of “this class” Although this example suggests a different diagnosis than Ismael’s of the dissolution of the problem of reidentification of the self over time, it is entirely in keeping with her view of the self itself: When I talk or think about myself, I talk or think about the connected, and more or less continuous, stream of mental life that includes this 2Note that this doesn’t help Descartes, who is trying to show that he cannot doubt theexistence of such a thing.
thought, expressing the tacit confidence that that is a uniquely identifying description (in the same way I might speak confidently of this river or this highway pointing at part of it, expressing the tacit assumption that it doesn’t branch or merge), but it need not be. There is no enduring subject, present on every occasion of ‘I’-use, encountered in toto in different temporal contexts. The impression of a single thing reencountered across cycles of self-presentation is a grammatical illusion of the sort exposed [earlier]. [TSS, p. 186] 3 The taxonomy of notational abuse
As I’ve noted, such grammatical illusions are one of the main concerns of The Situated Self. But the idea that grammar leads us into philosophical puzzles is a recurring theme in philosophy over the past century or so, and hence a question arises about the relation of Ismael’s notion to others. Can other familiar proposals of this kind be viewed as examples of indexical abuse of notation. Or is Ismael’s indexical “self-abuse” simply a species of a broader genus? One way to draw connections with some other uses of the notion of abuse of notation is to begin with the observation that indexicals behave like regular referring expressions, in fixed contexts – in other words, in communities who all share the same “location”, in the relevant respect. A dorm full of Sleeping Beauties,3 all woken together, might discuss whether “today” is Monday or Tuesday, just as if they were discussing the speed of light, or the mass of the Higgs boson. Take this fact, combine it with the practical importance of indexical beliefs, and one can’t help wondering whether other matters of practical importance are handled in the same way – coordinated across communities by being treated “as if” what was being discussed was a matter of fact.
From here, it is a short step to familiar views. Imagine speakers who face the challenge of coordinating their decisions under uncertainty. They want to reduce variation, across their community, in two elements of their cognitive architecture, 3A.!Elga,‘Self-locating belief and the Sleeping Beauty problem’, Analysis 60(2000),143–47.
viz., their credences and utilities. They adopt the practice of signalling their credences and utilities to each other, saying “That’s probable” about events to which they assign high credence, and “That would be good” about outcomes to which they assign high utility. These expressions have the grammatical form of ascriptions of properties, and behave accordingly in discourse. From the point of view of individual speakers, it is “as if” there is a factual matter under discussion, and the effect is to provide some pressure to resolve disagreements – in other words, to bring credences and utilities into line, across the community.
Proposed as a genealogy of our terms “probable” and “good”, this would be a version of the kind of expressivism we associate with Hume, Ramsey, and many later writers. One of the advertised attractions of such views – again a feature they share with Ismael’s indexical cases – is that they side-step the kind of metaphysical puzzles that arise if we take the language concerned at face value, as simply referring to aspects of our natural environments.
Expressivism thus provides a second kind of case in which abuse of notation seems to set the stage for productive disagreement about practical matters – at the cost, once again, of some philosophical perplexity, which a diagnosis of the grammatical illusion in question may then ameliorate. There may well be other cases, but this example serves to illustrate the issues I want to raise. What is the relation between the indexical case and others? Can classical expressivism be construed as appealing to a kind of indexicality, or are does it rely on distinct Let’s be clear, first, about what we mean by indexicality. Let’s construe it with Ismael, in terms of the task of coordinating first-person and third-person viewpoints – of locating oneself on a map insensitive to features that distinguish one’s own viewpoint from that of one’s fellows.
We can certainly find a place for such a task in the probability case, I think.
Imagine a map representing a range of possible sets of evidence, and the probability of a given event, relative to each set of evidence. This is a third-person map, in the sense that it abstracts away from the evidence actually possessed by any speaker. Accordingly, its use in guiding credence requires an act of self- location: we need to put a red dot on our actual evidence, as it were, and base our credence on the probability at that location. But this manifestation of indexicality doesn’t explain the distinctive cognitive role of credence (i.e., its architectural link to decision under uncertainty) – if anything, it presupposes it. Similarly, disagreeing about what evidence we have is not the same as disagreeing about what the probability is, given certain evidence.
Hence it appears that indexicality is more like a species than a genus. The genus is identified by the constitutive role of architecture in explaining the use of expressions of a certain kind – pseudo-referring expressions, as it were – and indexicality characterises one important subclass of architectural relations, those associated with the task of self-location on third-person maps.
4 Kicking the wheels away
There is another intriguing possibility in this vicinity, I think. Expressivism proposes that not all third-person maps need be construed simply as representations of the map-makers’ natural environment. There may be a non- referential role for maps, as an aid to coordination of particular kinds of architectural differences, across a community. By mapping probabilities and values, and then arguing about the true contours of those maps, we are encouraged to coordinate our credences and preferences. The proposal doesn’t require that we understand that that’s what we are doing – from the inside, as it were, we may see no difference between these maps and others. Their true character emerges only under philosophical scrutiny, when – puzzled by the task of coordinating them with other maps – we investigate their functions and genealogy. At this point, the expressivist claims, the best option is to regard them as sophisticated abuses of referential notation – abuses we may abuse even further, when we self-locate on The intriguing possibility is that we might push this a stage further, and regard the expressivist’s account of the (apparently) referential character of expressive maps as the only account we ever need of referential vocabulary. We would thus explain our referential idioms “internally”, in terms of their role in a certain sort of coordination task, the task for which we construct third-person maps – a task with many different actual and potential manifestations, in the coordination of many different aspects of the complex architectural relationships between ourselves and our our natural environments.
In a model of this kind, the need for indexicals will be much as before. After all, the fact on which indexicality depends – that wherever speakers map variation, they also need to keep track of their coordinates – is insensitive to the nature of the variation. What has changed is that we now explain the third-person viewpoint, as well as the first-person viewpoint, in the same pragmatic terms. In the process, we’ve changed our minds about what counts as abuse. We’ve stopped thinking of centered and expressive maps as useful but flawed imitations of genuinely representational maps, and dispensed with the pure case altogether, in favour of a plurality of “abusive” cases, each merely a solution to a particular coordination problem, differing only in scope.
As I noted at the beginning, the vehicle from which Ismael gets such impressive mileage rests on a chassis she borrows from John Perry. Perry’s chassis, in turn, rests on wheels we all inherit from Frege – wheels which do the job of anchoring language to the world, as it were, by taking reference and truth to be the fundamental notions of a theory of language. My prediction here is that the pragmatic project will show us how to retract Frege’s wheels, within the body of the craft itself. It will remain connected to the natural world, in so far as it needs to be, by its natural origins, as a craft that natural creatures build for themselves.
But unlike its earth-bound Fregean predecessors, it will take us anywhere our 4 I am very grateful to Jenann Ismael and to Paul Teller, for comments onprevious versions, and to the Australian Research Council, for research support.

Source: http://prce.hu/w/preprints/ssppr-final.pdf