Thursday, January 31, 2008

Quick Note on Autism and the First-Person

I recently read Eros Corazza's excellent book Reflecting the Mind which, despite the title, was basically a phil language book. I was struck, however, with one particular passage where he spoke of autistic people (I'm pretty sure he was talking about only some autistic people as I personally know at least one who doesn't have this problem) as having a hard time with personal pronouns like 'I' and 'you'. Corazza seemed to suggest that because they were not fully competent with these words that they had no first-person thoughts or concepts, etc. But that seems to me to tie the first-person too closely to language. Without first-person thoughts of any sort, there would be no genuine action at all - but autistic people can perfectly well perform actions (even if not always in the same ways or with the same skill as others). Their bodies don't just happen to move the way they do.

My counter-diagnosis would be that those who have a hard time with first-person or similar linguistic terms have such a hard time not because they lack no first-person representation (after all, some may still be able to use their own names in the way a non-sufferer would use 'I'), but rather because becoming competent with such terms requires knowing and applying a context-sensitive linguistic rule which requires a good grasp of understanding other perspectives and how the references of these terms change accordingly from one to another. It's a difficulty with "linguistic empathy", not thinking in the first-person.

Monday, January 21, 2008

Indexicals and Phenomenal Concepts

In this post, I want to connect up the contents of my previous two posts (found here and here). Indexicals connect up to non-indexical representations by showing up in non-indexical places in thought (for instance, 'now' shows up where the temporal parameter should be, 'I' where the personal parameter should be, etc.). Since it takes such a place in thought, it is thus is subject to a minimal amount of conceptualization and is conceptually connected to non-indexical concepts concerning the appropriate parameter it falls under. For instance, 'now' is conceptually connected to temporal concepts since it takes over the role of the temporal parameter at the level of explicit representation. Hence, we are aware of or at least attuned to the general function of 'now' as a representation with a certain kind of usage. And this is at least part of why indexical information "disappears" from the objective, third-person perspective - precisely because such information takes on an explicit role of a certain kind without explicitly assigning any particular value to the parameter it stands in for, and all of this is something of which we can be consciously aware of in the first person. This is why the 1- and 2-intensions of indexicals end up differing and why many of us are not tempted to think that there is any here or I or now in any kind of context-free sense.

With phenomenal concepts, however, we don't have all of this - hence, there is no "dissolution" of phenomenal information from the objective point of view. So even if phenomenal concepts end up being indexical-like in some important way (or recognitional or whatever), this important difference - that it doesn't take the place of a non-indexical parameter - helps make the difference between them and indexical concepts or other perspectival concepts. They simply do not have the conceptual role that normal indexical concepts do, even if their semantics may in some ways be similar.

Linguistic indexicals, for instance, "disappear" since we need to know third person truth conditions to interpret others' usage - and sometimes even our own. And this may sometimes be necessary even in our own thoughts - to know the links between our indexical concepts and the concepts relating to whatever these indexicals are standing in for. But with pure phenomenal concepts, things are a bit different - there is no need for a separate parameter cashed out using non-phenomenal concepts for which phenomenal concepts fill in. And hence there isn't such need to reinterpret what others say in phenomenal terms into non-phenomenal terms. If phenomenal concepts refer to features which can be reduced to the non-phenomenal, since these features would be so complex and outside the ability of most of the human race to get an accurate, specific and non-deferential cognitive fix on in non-phenomenal terms we would ordinarily have no need to have separate explicitly non-phenomenal parameters which phenomenal concepts fill in for - and, indeed, if phenomenal features are anything like brain states of a certain sort then not only would we ordinarily have no need but none of us at the present day (as far as I know) are in a position to even have such parameters. So we do not have such parameters and hence, even if the phenomenal is subsumed under the physical, there will not be much of anything in the way of conceptual connections between phenomenal and physical or functional concepts - even if phenomenal concepts are in some ways like indexicals.

Saturday, January 19, 2008

Tensed Thoughts 2.0

Assume all my representations are tenseless representations and I need to use them to act. Are these enough, representation-wise? Since we are talking about creatures which act at a time, for action, we need to coordinate the proposed action with time or else any success we meet in our endeavors will be a wild accident at best. So given tenseless information about a time at which I am located and need to act at, I need to be able to act reliably at that very time. This requires keeping track of the time in some way. To be reliable in coordinating tenseless temporal information with action, then, we need to make certain temporal information relevant in the proper way or action - or intention-in-action-producing at the relevant times. So, in other words, we need at least one clock-like-functioning system which somewhat reliably coordinates action with which times are appropriate to act at and so which tenseless representations are appropriate to act on. This system could take on multiple forms and, indeed, we may need more than one of these. Clock-like systems could include systems that act as timers, oscillators, accumulators, digital clocks, etc.

Now consider a speedometer. A speedometer represents a lot of different speeds. But that by itself is not its function - its function is to indicate (via the pointer) what the speed of the vehicle currently is. So lots of speeds are represented but only one in particular is represented as the speed of the vehicle - and this is done without explicitly representing anything but the speed itself. Now consider a carbon monoxide detector where a light labeled "Carbon Monoxide" lights up whenever the compound is present. Clearly, this detector represents carbon monoxide but its function is to detect the presence of carbon monoxide. So when the light is lit up, it represents carbon monoxide as present - even though its presence is not itself explicitly represented. This works even if the detector tokens a full representation which doesn't mention the presence of carbon monoxide, so long as the function of the system is still the same. So if the light is labeled "Carbon Monoxide is Very, Very Bad", the lighting up of the light and thus the system's coordination of the presence of the compound with the representation is still representing carbon monoxide as present despite this fact being extrinsic to the representation itself.

What we can learn from looking at these few examples is that if it is the function of a system to detect, indicate, or otherwise track that some F is G, and it does this by tokening a representation of F, it thereby represents F as G. Now apply this to clock systems. Clock-like systems can represent a lot of different times or be involved in coordinating times with representations which explicitly refer to a lot of different times. It is the function of the clock-like system to track the current time. And by doing so, it represents that time as being present or being now. So a clock-like system is essentially a tensed system - simply put, just a system for keeping track of the time. That is, we really do need tense after all - not simply tenseless representations.

What a tensed system like a clock in effect does is to attach temporal representations to the appropriate time in such a way as to be in a certain way infallible. This is especially apparent in the case of representations where time is not explicitly represented at all and yet which are still only about temporal matters (and are hence in this sense tensed) - the time enters into the represent implicitly or is represented by itself. An explicit NOW concept is perhaps at least partly a placeholder making explicit the implicit presence of the current time. Tensed systems and representations, then, constitute a kind of direct access to time that we need as agents to act.

If we have a system for keeping track of the time without necessarily requiring an explicit representation of time, it may be more economical for us to token representations which leave reference to the present time implicit. In that sort of case, an ordinary thought about the present time and one simply about how things are simpliciter or tenselessly may very well take the same explicit surface representational form – that is, there may be no syntactically present-tensed verbs at least at the explicit surface level. So we get things like 'Fred is cold' and 'Fred is human' where the first is to be interpreted tensedly whereas the second (arguably) is not (a less contentious example might be one involving a mathematical sentence). Yet both have the same surface explicit surface syntax.

In the tensed representations of this sort that lack a NOW or similar concept or locution, no explicit piece of the representation represents the time (the time of the representation itself does this). The temporal 'at t' parameter which is in the truth conditions is hence not explicitly specified, which is why tenseless and tensed representations of these sorts will look the same. In these present-tensed representations, it is the time of the representation itself (roughly) that enters as the value of the implicit parameter. In other, perhaps less primitive ones, we may have a sentence with an implicit temporal parameter that is not pointed at the present but where the value of the parameter is some time which is particularly salient or otherwise demonstrated. So 'Go to my house' may have the current time as its value - I want you to go now - or it may have some other time contextually specified - such as some time soon, or after you've picked up my laundry, etc.

As mentioned earlier, NOW perhaps, then, acts as something which makes the parameter itself explicit, often indicating the time of the representation but without specifying explicitly which exact time that is - it is a stand-in for that time, whichever it may be. So NOW can be used to explicitly disambiguate representations that can take one or more tensed readings and/or a tenseless one. This might explain part of the reason why NOW always takes wide scope in sentences (particularly modal or temporal ones), since it's really the implicit current value of the temporal parameter that enters into the semantics – the NOW simply indicates its presence or place in those semantics.

All of this perhaps explains part of why tensed views of time are so attractive or natural to many people (put aside whether such views are true or not) - since there is no surface difference between present-tensed and tenseless representations, it is easy to confuse being F at t (where t is the current time) with being F simpliciter. If there is no explicit parameter – just NOW as a placeholder – then it will be easy to confuse ‘being F’ with ‘being F now’. And since NOW is a relatively simple concept, irreducible to tenseless ones, etc., then if one mistakes properties of representations for properties of what is represented (or for properties of what is expressed) one will take it that NOW expresses a relatively simple property which is not reducible to any tenseless ones – that is, there must be irreducibly tensed properties. And not only that, since the reference of NOW shifts over time, given this same confusion we will get confused notions of tensed “temporal passage”. A kind of primitive use/mention confusion.

Additional note: NOW is, perhaps, thus different from PRESENT. PRESENT seems to mean something akin to LOCATED AT or IN THE PRESENCE OF. So 'A is present' "means" 'A is located at' or 'A is in the presence of'. Unlike NOW, PRESENT does not take wide scope. Since this sort of representation is tensed, what is really going on is that the temporal parameter which would complete the representation and tell us at what time A is located is left implicit. So PRESENT takes a temporal parameter like many other predicates and doesn't specify a time at all on its own but will pick up that parameter, like any other similar predicate, from that governing whatever clause in which it happens to be embedded. (Compare the difference in parameter between 'A is present now', 'A is present at noon', etc.)

Friday, January 11, 2008

A Quick Note on Chalmers on the Phenomenal Concept Strategy

Chalmers (in numerous places) objects to the phenomenal concept strategy as a response to his explanatory gap argument against mainstream forms of physicalism. On most versions, he says, they make phenomenal concepts out to be indexicals or demonstratives. Yet indexicals and demonstratives have different 1- and 2-intensions whereas phenomenal concepts do not (hence why, from a third person point of view - according to Chalmers - when considering the truth conditions of someone else's indexical or demonstrative thoughts or other representations, the indexicality disappears, whereas this is not the case for pure phenomenal representations).

However, I think Chalmers is way too quick. If phenomenal concepts or representations are at least partially self-reflexive or represent phenomenal properties with those very same properties then we will have phenomenal concepts which do not have differing 1- and 2-intensions after all. And the phenomenal concept strategy will still work fine for physicalists of most stripes. It is simply a mistake for Chalmers to think that any sort of reflexivity or being recognitional makes a concept somehow automatically indexical or demonstrative. If it is essential to the concept that it has the content it does and hence does not have differing intensions then it will not be an indexical or demonstrative concept, contra Chalmers. So more work would need to be done to defend Chalmers' arguments agains this popular strategy.

Tuesday, January 8, 2008

The Function of Clocks and the Purpose of Tensed Thoughts

So it's been a long while since my last post. The end of the quarter, vacations, illnesses, a desire to fast from the blog world for a short time to recover, and a bad bout of food poisoning all conspired to keep away for a while. This is the first time I've looked at any blog in a while, so hopefully I'll start to get back into the swing of things and get back into posting on my own blog and having discussions on others. It'll be interesting to see what happened in some discussions I was involved with when I suddenly went MIA.

In any case, this post is a follow-up to the previous one. I'm not really happy with the argument of the last post - there are a number of points that are really weak or where the argument I now think isn't so good. So I'm going to try another take on why we need tensed thoughts here in this post. (Where the "we" in question is just finite, spatiotemporal agents who act at only particular times and places adjacent to their own particular times and places and do so as a result of (sometimes) deliberation and (all the times) intention)

Since we are talking about creatures which act at a time, for action, we need to coordinate the proposed action with time or else any success we meet in our endeavors will be a wild accident at best. So given information about a time at which I am located and need to act at, I need to be able to act reliably at that very time. This requires keeping track of the time in some way. To be reliable in coordinating temporal information with action, then, we need to make certain temporal information relevant in the proper way or action - or intention-in-action-producing at the relevant times. So, in other words, we need a clock-like system which somewhat reliably coordinates action with which times are appropriate to act at and so which representations are appropriate to act on.

Now consider an odometer. An odometer represent a lot of different speeds. But that by itself is not its function - its function is to indicate (via the pointer) what the speed of the vehicle currently is. So lots of speeds are represented but only one in particular is represented as the speed of the vehicle - and this is done without explicitly representing anything but the speed itself. Now consider a carbon monoxide detector where a light labeled "Carbon Monoxide" lights up whenever the compound is present. Clearly, this detector represents carbon monoxide but its function is to detect the presence of carbon monoxide. So when the light is lit up, it represents carbon monoxide as present - even though its presence is not itself explicitly represented. This works even if the detector tokens a full representation which doesn't mention the presence of carbon monoxide, so long as the function of the system is still the same. So if the light is labeled "Carbon Monoxide is Very, Very Bad", the lighting up of the light and thus the system's coordination of the presence of the compound with the representation is still representing carbon monoxide as present despite this fact being extrinsic to the representation itself.

What we can learn from looking at these few examples is that if it is the function of a system to detect, indicate, or otherwise track that some F is G, and it does this by tokening a representation of F, it thereby represents F as G. Now apply this to clock systems. Clock-like systems can represent a lot of different times or be involved in coordinating times with representations which explicitly refer to a lot of different times. It is the function of the clock-like system to track the current time. And by doing so, it represents that time as being present or being now. So a clock-like system is essentially a tensed system - simply put, just a system for keeping track of the time.

What a tensed system like a clock in effect does is to attach temporal representations to the appropriate time in such a way as to be in a certain way infallible. This is especially apparent in the case of representations where time is not explicitly represented at all and yet which are still only about temporal matters (and are hence in this sense tensed) - the time enters into the represent implicitly or is represented by itself. An explicit NOW concept is perhaps at least partly a placeholder making explicit the implicit presence of the current time. Tensed systems and representations, then, constitute a kind of direct access to time that we need as agents to act.