Author: Richard Brown

A Couple More Thoughts on Shombies and Illuminati

Richard Brown 1 Comment

Towards the end of the discussion we began to discuss what was involved in conceiving of the shombie world. Dave seemed to think that doing so involved conceiving of a possible world where physicalism was true. This amounts to the claim that <>[](P–>Q), i.e. that it is possible that it is necessary that the physical facts of our world entail that there are qualitative facts. In S5 if something is possibly necessary it is necessary (this is intuitive since if there is one possible world where a necessary truth is true then, if it is to be necessary, it must be true at all worlds including the actual one). This way of reading the shombie/illuminati argument has it more like contemporary versions of the ontological argument.

This is a bad way of thinking about the shombie argument since once we start to think about the possibilities of necessities we have to think about the entire space of possible worlds and our intuitions about how this space will turn out become very untrustworthy. I agree with Dave that we are better off just trying to conceive of one specific possible world. So to conceive of the shombie world all we need to do is to conceive of one possible world where physical facts entail qualitative facts (or alternatively where qualitative properties just are physical properties). That world is negatively conceivable (call this the illuminati claim) and plausibly positively conceivable (call that the shombie claim). Now is this a world where physicalism is true? Well that depends on your views about the necessity of identity statements. As is well know there is a very simple proof of the necessity of identities. The following is taken from Kripke’s Identity and Necessity,

1. (x)(y)(x=y)–>(Fx–>Fy)

2. (x)[](x=x)

3. (x)(y) (x=y)–>([](x=x)–> [](x=y))

4. (x)(y) (x=y)–> [](x=y)

The first premise says that if x and y are the same thing then if x has a property y has that property also. The second premise just says that it is necessary that every object is self identical. The third premise is a substitution instance of 1. What it says is that if x and y are the same thing then if x has the property of being necessarily self identical then so does y. From that we get four since we know that [](x=x) is true (though I suppose one could block the argument if they denied this…maybe someone like Dave will do this?)

At any rate 3 makes clear the point I am trying to make. The postulated identity in the antecedent is one thing and the hypothesis that if there is an identity then it is necessary is another. Thus we do not have to conceive of the shombie world as one where it is possible that some necessary truth is true. We simply need to conceive of a world where P=Q for some subset of the total physical facts and some qualitative fact. It is true that this by itself will not rule out dualism. The dualist might counter that the shombie world is conceivable but dualism still might be true of our world (though as a side note this is a much less attractive view for the dualist since if they admit that it is possible to have shombies then we probably should be focused on empirically finding out which is true) but this can be ruled out independently with the above argument.

Towards the end of the discussion Barbara Montero brought up the possibly of what she called “blocker worlds”. A blocker world is one where there are illuminati, or shombies, but in addition there are nonphysical properties or entities that ‘block’ the qualitative properties. Thus in this world there are physical duplicates of us, the physical facts entail that there are qualitative facts, but the nonphysical blockers block these facts. Barbara and dave seemed to think this was a problem for physicalism. Jim Pryor suggested that it wasn’t since even though in this world there were no qualitative properties if one just took the physical facts by themselves we would get qualitative facts. Kati seemed to think that this world wouldn’t be a world where there was a minimal physical duplicate of me since a minimal physical duplicate is one that has only the properties that I do and nothing else. Others were pressing here suggesting that blocker world folk are minimal physical duplicates because they have all of the very same physical properties as I do. I seemed to think that these creatures wouldn’t really even be physical duplicates of me since the blockers were interfering with the physical processes in such a way as to make a physical difference (how else could they block the qualitative properties? Barbara suggested that the blockers were like aspirin, but this works by effecting physical processes…I mean clearly there is a physical difference between someone who took aspirin for a headache and someone who did not.)

Anyway, I have to get ready to head out to Benji Kozuch’s cog sci talk at the Graduate Center….

Shombies & Illuminati

Richard Brown 3 Comments

On Tuesday I attended a discussion with Kati Balog of her paper Zombies, Illuminati, and Metaphysical Gridlock which was very interesting.

She begins by thinking about zombies and the things they say. If there really is a world that is microphysically identical to ours but that lacks consciousness and one of those creatures is a microphysical duplicate of David Chalmers then it is reasonable to think that Zombie-Dave will be giving zombie versions of the zombie argument. Zombie-Dave will say things like that he doesn’t see how structure and function could suffice for phenomenal experience even though all the while he doesn’t have any phenomenal experience (he is a zombie after all). Zombie-Dave will even say things like from the fact that he can conceive of all of the physical facts without the phenomenal facts obtaining it follows that physicalism is false.

The response to this argument hinges on the distinction between negative and positive conceivability. To negatively conceive of something is simply to be unable to rule it out a priori. Positive conceivability involves something more. For Kati positive conceivability is just negative conceivability (i.e. not being able to rule something out a priori) plus the application of standard everyday concepts. This is a modest notion of positive conceivability that leaves no room for an act of the imagination or a special act of conceiving. At any rate, the response to the Zombie-Dave objection then becomes the claim that while Zombie-Dave might be able to negatively conceive of P & ~Q* (where Q* is Zombie-Dave’s conception of what a phenomenal property is) he cannot positively conceive of it.

One of the interesting things that came up at this point was whether or not zombie phenomenal concepts have any mode of presentation. Kati seemed to think that they did not. Since Zombie-Dave denies that phenomenal properties are physical/functional properties it seems right to say that the zombie phenomenal concepts do not present phenomenal properties as physical/functional properties. Given that zombies actually lack phenomenal properties then there doesn’t seem to be anything else that could be a mode of presentation. So Kati thinks that a zombie’s phenomenal concept refers to a physical state of the zombie’s brain but it doesn’t present that property to the zombie in any way. While this seems right, it also seems right to say that there is some way in which the zombie thinks about, discriminates, or is rationally responsive to things it would call red, blue, green, etc. If all one means by mode of presentation is something like ‘a way of thinking about the referent’ then there doesn’t seem to be any problem in saying that the zombie’s phenomenal concepts present the brain states in a certain way.

Kati then introduces Illuminati as an answer to the positive conceivability argument. Illuminati are creatures that are completely and only physical and which have phenomenal properties. I call illuminati ‘shombies’ Keith Frankish has called them ‘anti-zombies’. Are such creatures conceivable? Well, clearly they are at least negatively conceivable. As Kati argued, illuminati are negatively conceivable for the same reason that zombies are. There is nothing in our grasp of phenomenal properties that tells us one way or the other whether they are physical or not. This is an important point. Someone who denies this is what I call a type-A dualist. Type-A dualism seems to me to be clearly an unattractive view since it amount to no more than defining phenomenal properties into nonphysicality. The response to this negative conceivability argument is again to invoke positive conceivability and the a priority of the zombie intuition. if it is a priori that zombies are positively conceivable then it cannot be the case that illuminati are negatively conceivable.

At this point there are two ways one can go. Kati’s preferred way is to try and bolster the negative conceivability of illuminati by offering a coherent physicalist metaphysics and ontology. If one can give a coherent philosophical account in completely physicalist terms then it would seem to bolster the case for the negative conceivability of illuminati. Another way to go would be to invoke positive conceivability of shombies. The shombie argument starts from the premise that shombies are positively conceivable (or what I really think is that it starts from the premise that they seem positively conceivable). Kati is skeptical of this move since she thinks that someone like Dave might be working with a notion of positive conceivability that is strong enough to make shombies inconceivable. My feeling is that any such notion will also make zombies inconceivable. But we would have to see the actual account of positive conceivability in play. According to the one I have been working with, shombies certainly seem positively conceivable. I can actually imagine discovering that we are shombies via the use of a philosophical theory of consciousness (like a higher-order theory) and relevant brain science discoveries.

Dave’s objection was that zombies are more intuitively conceivable than shombies. Perhaps after the acquisition of a bunch of theory one could have the shombie intuition but regular people have the zombie intuition quite strongly and pre-theoretically. I wonder whether if some experimental philosopher actually asked a bunch of undergraduates whether Dave would be right. My experience is that at least some of my students have the shombie intuition quite strongly and seemingly pre-theoretically. But I really don’t see why this matters. That one theory is taken for granted doesn’t show that it is true. I can’t imagine that someone who objected to the heliocentric view of the universe by complaining that one needed a bunch of advanced physics under your belt to see how it could be true would be taken seriously. It might be true that our natural inclinations run towards dualism; I admit that mine did when I was an undergraduate…but I also thought that the structure of the atom mirrored the structure of the solar system.

So I think that the shombie argument is the natural next move in the debate, but Kati does not. She then wonders what we are to say if at the end of the day we end up with two theories that are mutually incompatible but equally simple, elegant, predictive, and supported by the empirical evidence? If the physicalist and the dualist are each able to offer complete metaphysical and epistemological frameworks that are both compatible with the totality of the findings from science then what are we to say? Do we say that there is a fact of the matter about which theory is correct? Or do we say that there is no determinate fact of the matter about which is true? Or finally do we say that the issue is merely terminological? Kati seems to think that the answer is indeterminate. I think it is probably just terminological at that point, though as I have been arguing recently, I wonder if the dualist can even give a coherent account of what nonphysical properties are and so I doubt that we will end up with two equal theories at the end of inquiry. It seems to me that if the debate between the dualist and the physicalist comes out to be indeterminate then there are no determinate facts at all. That is, once we go Quinian we can’t stop the backslide all way to the inscrutability of reference. I for one just can’t bring myself to believe this. I mean if it were true then the fact that there are no determinate facts is itself indeterminate; but what could that mean?

…or so it seems to me at first glance as I run out of the house to class…

The Singularity and Simulation

Richard Brown 1 Comment

(cross-posted at Brains)

There is a nice video of a recent talk by David Chalmers on the singularity available here. Dave also summarizes the argument in a recent post at his blog Fragments of Consciousness (here). He also gave this talk at the Graduate Center, which is where I saw it last Wednesday. It is an excellent talk and I hope it starts people talking about these interesting issues. Assuming you believe that AI is a possibility I find the general line he is pushing very persuasive and would be interested to hear what others thought about it.

One thought that I had was that if the second premise of the argument is right then we might have some kind of evidence that we are not living in a simulated world. If we were we would be the AI and the second premise says that once you have AI it will be a matter of years before you have AI+, but we haven’t had AI+ yet (i.e. strong A.I.) so we are not AI. When I asked about this Dave responded that ‘a matter of years’ should be interpreted as in the time scale of the next world up. If we are indeed in a simulated world then the simulators of our world could presumably manipulate the time scale in the simulated world. So what may seem like a long time to us could be a few seconds for them. Ah well, I guess we still can’t be sure that we aren’t in the Matrix.

This allows me to clarify the point of my previous post. In discussion with Dave about it he pointed out that what I describe is just one kind of dualism and that it is not the kind that the zombie argument deals with. This is a fair point. Looking back at the post I see that I was sloppy in presenting the argument. I should not have been saying that the zombie argument by itself is an argument that we are in a simulated world. What I should have said is that this account of what a nonphysical property is is the only one that is one the table. But when we adopt this as a theoretical account of what non-physical properties are even zombies can have them and so they do not seem to threaten physicalism. If there is some other account of what a nonphysical property is then we can examine it and one cannot say that an obvious example of a nonphysical property is seeing green or feeling pain. What is needed is an account of what it would mean to say that feeling pain is nonphysical. I, for one, can’t even conceive what that would mean except in the way Dave does in his matrix paper.

The Matrix & Nonphysical Properties

Richard Brown 3 Comments

(cross-posted at Brains)

I have long wondered what dualists mean when they speak of nonphysical properties. Today I was reading Chalmers’ paper The Matrix as Metaphysics and he says something that may shed some light on the way in which he thinks of nonphysical properties. He argues that the matrix scenario can be construed as a metaphysical hypothesis about the ultimate nature of the physical world. If this is right then there is a sense in which dualism is true. The mind is a distinct entity that exists outside of physical space-time and causal interacts with the physical body. This is because the physical theory that is true of reality in the matrix is a computational theory on which the ultimate things which exist are bits (zeros and ones). Thus brains in the matrix are ultimately composed of bits and when people in the matrix talk about brains they ultimately are talking about bits. The brain which is outside of the matrix is not composed of bits (let us assume). It is ultimately composed of something else (let’s say strings). Thus the brain outside the matrix, when viewed from the perspective of someone who is in the matrix, is nonphysical. It is not something that could be deduced from a completed matrix microphysics (which would be phrased in terms of ones and zeros).

One might wonder whether a completed matrix physics would have to be supplemented with (from the perspective of the matrix) nonphysical laws in order to capture outside the matrix facts or whether we might view the truly completed matrix physics as being expanded to include the outside the matrix physics. On this latter view the laws of matrix-physics would be a special subset of the laws of outside-physics. If this were true then the matrix-physics would not be complete until it was expanded to include outside-physics and physicalism could still be true. One might also wonder whether people in the matrix had largely true outside-physics beliefs since the matrix world is a deliberate simulation of outside-physics.

But even setting aside these issues there are strange results. Suppose that physicalism is true and that consciousness is a purely physical property of the brain. Let us also assume that this is true of a brain that is not in a matrix scenario. Call this scenario 1. Now imagine that a physical duplicate of this physicalist brain that has been in a matrix scenario since birth Call this scenario 2). Then physicalism is true in scenario 1 and dualism is true of scenario 2. But these brains are physically identical! Furthermore this shows that we could not resolve the dispute between the physicalist and the dualist until one was in a position to determine whether or not one is in a matrix scenario. Since Chalmers himself admits that he cannot a priori rule out that he is not in a matrix scenario he must also admit that he is not in a position to a priori tell if physicalism or dualism true. So, suppose that we are actually in a matrix scenario then conceiving of zombies is just conceiving of a computer simulation composed completely of NPCs (non-player characters). But this doesn’t show that physicalism is false, since physicalism is best construed as the claim that lines up with the first brain; since with this understanding of nonphysical physicalism turns out to be nothing but the hypothesis that we are not in the matrix.

But even if we were in the matrix there is a sense in which we can say that physicalism is still ultimately true since in the above envisioned world qualitative properties turn out to be identical to properties which are physical in terms of outside-physics (since these properties are the very same as the ones in the world where physicalism is true).

Zoombies are creatures that are nonphysically identical to me in every respect and which lack nonphysical qualitative properties. I have in the past suggested that one way to conceive of zoombies is as Cartesian minds that only have thoughts but no qualia but now we can put it in terms of matrix scenarios. A zoombie has all of the same nonphysical properties that I in fact do. Suppose that I am in fact in scenario 2 above. Then a creature that has all of the nonphysical properties that I in fact do will have a brain that is identical to my outside-brain. This is to imagine scenario 1.

The traditional zombie is a creature that is physically identical to me and lacks consciousness. Now suppose that I have a zombie twin who is in a matrix scenario since birth. My matrix zombie twin has nonphysical properties (which are the very same properties that I physically have) but no qualitative properties. So, whether one has nonphysical properties or not is simply a matter of whether one is in the matrix or not. Chalmers’ defense of nonphysicalism can thus be seen as a defense of the claim that we are in the matrix.

Busy Bees Busily Buzzing ‘Bout

Richard Brown 1 Comment

This last week was a very busy one. As you may have noticed from the side bar the call for papers for Consciousness Online is now out…spread the word!

Tuesday I attended a talk/discussion of a paper by Ned Block on Attention and direct realism. Direct realism, roughly, is the view that when one has a veridical experience, say of the subway train coming into the station, the phenomenology of one’s experience is is determined by, or is constituted by, the properties that the object actually has. So on this view when one sees the subway one is somehow directly in contact with the physical object. This is contrasted with the view that one’s phenomenology is instead determined by, or constituted by some kind of mental representation that is perhaps caused by a physical object but which represents the physical object with a set of mental properties.

Block was arguing that direct realists can’t explain a certain fact about attention. His argument revolved around an interesting phenomenon discovered in attention research. If one is staring at a fixed spot and while doing that focuses one’s attention on one of three circles that are interlocked (something that is hard but can be done) one sees the circle one is attending to as brighter than the others. With practice one is able to move ones attention around the three circles and light them up as one goes. Given this some researchers took two gratings one of which was slightly dimmer than the other and what they found was that when the subjects attended to the fixation spot they could tell which of the two gratings was actually brighter than the other. But when they attended to the fixation point and shifted their attention to the dimmer patch they judged it to be the same brightness as the other patch; that is the two patches looked equally luminous. Block’s argument was then that the direct realist did not have any objective thing in the figure that they could point to to explain the difference in phenomenology. the figures stayed the same. Nor did they have any principled reason to say that one of the two perceptions was illusion and the other veridical.

On Friday I went to Jeremy Grey’s talk at the cuny cog sci speaker series. he was presenting data on the relationship between intelligence, as measured by standard psychological measures correlates with self-control. He was arguing for what he called the Individual Differences view. He started with a famous and intriguing study that found a correlation between self-control in 6 years olds and their subsequent performance on the SAT’s. Kid were given the following two options. They could either take a marshmallow that sitting on the table in front on them right now or they could wait until the experimenter returned and have two marshmallows. The experimenter then left and the children were videoed. Some of the kids were able to wait for the two marshmallow reward while others gave in immediately and ate the one that they had in front of them. What was surprising was that 12 years later when they took their SATs the ones who did best were those that waited longest for the two marshmallow reward. That is, the longer they were able to resist the marshmallow in front of them and wait for the return of the experimenter (some made it others didn’t) the higher their SAT scores were. Grey did a series of studies on adults to test the relationship between intelligence and self control and he found that there was indeed this relationship. There were, however, some people who scored high on the standardized tests but also scored high on impulsivity tests (that is they would be classified as high intelligence and low self control). The even more surprising thing was that if you factored in a certain kind of genetic variation which results in a variation in the dopamine receptors one saw that the outliers had this variation while those who conformed to the model did not. He also pointed to a study which suggested that pre-school children who participated in daily self-control exercises improved their performance on standardized IQ tests and so there is room for optimism that one is not stuck at one’s current IQ/self-control level.

Attributing Mental States

Richard Brown 1 Comment

Friday I attended James Dow‘s talk at the CUNY Cogsci Speaker Series. He was concerned with answering the question of how people are able to ascribe various mental states to themselves. In particular he was interested in critiquing the account offered by Bermudez and developing an alternative account inspired by P.F. Strawson.

The standard account has it that we first come to see that we have certain mental states like belief, pain, etc and that these result in various behaviors (e.g. utterances as well as other behavior). We then notice that other people engage in these kinds of behaviors and then reason by analogy that since they exhibit behaviors like the ones that I do when I have a, say, pain these people must also be in pain. Thus the standard account has that we start with ascribing mental states to ourselves (I am in pain) and then use that ability to ascribe mental states to others (Doug is in pain).

Bermudez criticizes the standard account using a posteriori evidence from developmental psychology. In particular Bermudez uses data from the phenomenon of joint attention. In joint attention you have two observers each attending to some object, say a piece of fruit, and each aware that the other one is attending the same object. Bermudez argues that in order to be able to do this (and infants do it as early as 9 months old) the child must be representing the mother as seeing the object and attending to it. This, in turn, must mean that the child represents the mother under a psychological sortal; that is as seeing x and attending to x. This together with other evidence that suggests that the child does not at this point attribute mental states to itself  shows that the standard account can’t be right. Bermudez then argues that the best explanation of what is going on here is that the ability to attribute mental states to others constitutes the ability to attribute mental states to oneself.

Dow wanted to criticize Bermudez for using an a posteriori argument to establish something like logical dependence. Whatever that turns out to mean Dow’s basic concern was to develop the Strawsonian alternative and to argue that none of Bermudez’s arguments decide between his account and the new alternative. In short the Strawsonian alternative postulates that the child simply has the ability to pick out other persons. ‘Person’ here is used in the Strawsonian sense as of something which has both mental and physical attributes. Dow claims that in representing the mother as a person the child is neither representing their mother under any kind of sortal. They simply attend to the eye and where it is focused. This Strawsonian view is what Dow called a ‘no-priority’ view in that it holds that there is a logical dependance (whatever that is) between self and other ascriptions (at the very least it seems to mean that in order to have the one ability one must also have the other ability (and vice versa?) but that neither one develops before the other.

We were promised a transcendental argument that was supposed to establish this but we ran out of time.

Over drinks I had an interesting discussion with Josh Dulberger who was proposing a novel take on the simulation theory/theory theory debate. Traditionally these are thought to be opposed but Josh suggested that they need not be. He thought that the simulation might be used to generate data for the theory one employs of other people and their mental states. This is an interesting idea. He then suggested that if one thought this then one might be able to argue that the function of consciousness lay in enriching the data that one gets. Intuitively the idea is that consciousness gives one better access to ones own mental states and so boosts the amount of data that one has for one’s theory thus making the theory richer. He thought this was nice since one could adopt David Rosenthal’s higher-order thought theory of consciousness and then argue that Rosenthal is wrong that there is no function of consciousness using his own theory against him, so to speak. But there is a problem here. In Rosenthal’s account there must have been a time when there were people who were able to infer what mental states they were in from observing their own behavior (which includes verbal utterances). But since these people are not able to have these higher-order thoughts in a way that seems unmediated by inference they do not have conscious mental states yet. As they get better and better and attributing these kinds of states to themselves they get to the point at which these attributions no longer seem to be mediated by inference at which point they come to have conscious mental states. At the point just before they have conscious mental states their access to their unconscious mental states is just as good as it will be when they do have conscious mental states. They will have what seems to them to be a different kind of access to their mental states but they really just have the same access as before (only now it seems to them to be immediate and non-inferential). If this is right then Rosenthal’s account ia not committed to Dulberger’s claim that consciousness produces more data. Our Rosenthalian ancestors have all of the same data that we do even though all of their mental states are unconscious.

On an interesting side note Daniel Shargel pointed out an interesting difficulty for Rosenthal in this story. The Rosenthalian ancestors do not have any conscious mental states. At some point they acquire the appropriate concepts which enables them to have higher-order thoughts attributing (theoretical) mental states to themselves. On Rosenthal’s account the fact that these higher-order thoughts are mediated by inference means that they do not result in the target states becoming conscious. It is only once the higher-order thoughts are seemingly unmediated by inference that we get consciousness. But Shargel argues that when these Rosenthalians have their very first higher-order thought, whether mediated by inference or not, it will not seem to them to be so mediated since nothing seems any way to them (all of their mental states are unconscious). Shargel suggested that Rosenthal response to this was that the inference needs to be the product of some internal mental state. Since the Rosenthalians always make inferences based on external perceptions the higher-order thoughts they have are not of the right kind. I wonder if there is some other response he can give, but this is already too long!

Unconscious Trait Inferences

Richard Brown 1 Comment

Friday I attended social psychologist James Uleman‘s talk at the CUNY Cogsci Speaker Series. His topic was unconscious trait inferences. A trait inference is exemplified by the following. Subjects are presented with a sentence, for instance, “she solved the mystery half-way through” which reliably produces subjects to infer a trait. In this case the trait is something like clever or smart. Combining his interest in trait inference with a generally accepted paradigm in social psych that contrasts traits with actions as being more abstract Uleman designed an experiment to test whether or not people make unconscious trait inferences. His experimental setup is as follows. He presents subjects with a picture of someone engaged in an activity (say reading a book) and underneath has a sentence like the one previously mentioned, though in some of the trials he includes the trait explicitly (i.e. “she was so clever she solved the mystery halfway through”). He lets subjects look at the picture and the sentence for up to 8 seconds. Then they are presented with the picture and asked whether or not the trait was explicitly mentioned in the sentence which described that picture. So, to take our earlier example, if a subject is presented with a picture of a women reading a book described by the sentence “she solved the mystery halfway through” and then later presented only with the picture and asked if the trait term appeared explicitly or not. The interesting thing is that he is able to show that people falsely remember the trait being explicitly stated in sentences where it did not explicitly appear. This is taken as evidence that the inference to the trait is made unconsciously.

Even more interestingly, Uleman was able to show that if one simply varies the geographical distance one is able to show that people make less trait inferences about people who are near to them than they do about people who are distant from them. To show this he presented subjects with the same pictures of the same women reading the same book. The only difference is that before the picture was presented subjects either saw a picture of Washington Square Park (these are NYU students so this is the near condition) or a picture of something in Florence Italy (with ‘Florence’ explicitly stated in the picture. Subjects in the distant condition made a statistically significant higher number of false recognitions (i.e. they said that the trait term explicitly occurred in the description when it actually did not more times) than did those in the near condition. What this is taken to show is that people make more unconscious trait inferences about more distant people than they do about near people.

But why? The explanation offered went as follows. When a person is distant it is not a good strategy to focus on what they are doing at the moment as that can change. Rather what one wants to do is to focus on central, relatively stable attributes of the person. On the other hand when a person is near it is important to focus on what they are currently doing rather than on abstract stable traits. This is a curious finding (many people resisted it) but it makes sense to me. These inferences are ‘unconscious’ because subjects are not instructed to make them. They are not given a rating task (i.e. rate the following pictures on how clever they are) which requires them explicitly to make the inference.

Perhaps even more interestingly Uleman was able to show that the notion of distance in play is not simply geographical. You get the same results when the distance is social (i.e. class related) or even in terms of power. People who are asked to remember a time when they had power over someone else (defined as ability to give or deny something to someone) make more unconscious trait inferences than people who were asked to remember a time when someone else had power over them. The idea here is that when one feels powerful one feels “above it all” and when one is powerless one has the issues “hanging over them”.

In the second half of the talk Uleman presented data on the issue of whether or not people treat traits as causes or descriptions. That is, do people treat, say cleverness, as a cause of the women solving the mystery halfway through. I did not really follow this half of the talk, but I did pick up on one interesting tidbit. He seemed to be suggesting that people who are asked to make these kinds of trait inferences explicitly are more likely to stereotype and scapegoat an out group than those that make these kinds of inferences unconsciously…this is interesting, though I am not sure what to make of it…in fact I may be misremembering it….

Kripke on the Structure of Possible Worlds

Richard Brown 9 Comments

Yesterday I attended Saul Kripke’s talk at the Graduate Center (there are quite a few interesting talk coming up…also looking good are the cogsci talks). The title of the talk was “The Structure of Possible Worlds: a Preface to a statement” and was subtitled, “Prolegomena to a talk on possible worlds, some considerations” so maybe I didn’t really attend a talk. Much of the talk, um preface to a prolegomena to a talk, consisted of Kripke going over the history of his thinking on modality, punctuated with his usual wit and humor. My two favorite moments were (a) the one where, after talking for an hour (the scheduled length of the talk), he stops and says “I may have to go over because I haven’t come to the main point yet” and (b) the one where, while discussing modal realism, he says “what does God have to do to make a really existing merely possible world actual; give it a kiss?” Ah, that Kripke should have his own reality show…I bet it’d be really popular!

As to the content of the talk, or whatever, he seemed to be indicating that (he may have) changed his mind about how he conceives of possible worlds. Here is how he put it in the handout,

I used to think that a sufficient account of my view of what a possible world is might be given by something like a Russellian structured proposition describing it. Now i think that one cannot give an account of what a possible world is in and of itself, but only as part of the structure of all possible worlds, or at least that modal logic cannot rule this out. (Even if there is not a unique structure of all possible worlds, structures where the problem I have just described arise cannot be ruled out philosophically.)

A structured Russellian proposition is an abstract entity that has as its constituents the actual objects in the proposition. So, to modify the classic example, the proposition that Jennifer Aniston loves Brad Pitt has Jennifer Aniston (the actual person), Brad Pitt (the actual person), and the relation of loving as constituents and these constituents are ordered by the loving relation such that Jennifer loves Brad (which is a different ordering from Brad loving Jennifer). This is just a more refined way of putting the basic point of Naming and Necessity. When we ask if Al Gore could have won the election we are asking a question about the actual Al Gore. We are not talking about some distinct object which merely resembles Al Gore (or which merely has many of the same descriptions true of it). There is therefore no issue of how we re-identify Al Gore in various possible worlds; it is the actual Al Gore.

So why does he now think that we can’t do this? (Of course, he hasn’t really come to this conclusion officially. What he said was that it might be true and he sort of leaned towards thinking that it is true). The basic reasoning he employed relies on his argument that there can be objects which are indistinguishable in every way which are none the less distinct. There is no criterion of identity that distinguishes these objects; as Kripke put it “the only difference between them is that there is a difference.” His favorite example is the square root of -1, known as i. i is equal to whatever number equals -1 when multiplied by itself. This number is not on the real number line and so is known as an imaginary number. The problem comes when we realize that i and -i are distinct numbers and that everything that is true of the one is also true of the other. i and -i are distinct yet have no clear criterion of identity.

The problem that Kripke sees is that something like this might be true for possible worlds. To see this he talks about what he calls ‘grounded’ objects. So, let’s imagine a world where there is an person, let’s call him George, who does not exist in the actual world. Let’s say that George is the son of two people who actually exist but do not actually have kids. In the actual world the sperm and egg that come together to form George never meet, but they do in some possible world. So George is grounded in the actual gametes that he could be the product of. But there is also the possibility of ungrounded objects. So consider Georgette. Georgette is a women who does not exist in the actual world but does at some possible world, yet unlike George, Georgette is not the related to any past, present, or future actual person. The sperm and egg which come together to form Georgette at no time exist in the actual world; they are the products of an alternate history which does not overlap with the actual world (in this respect). Georgette is ungrounded.

So if there can be ungrounded entities then we can see that the following situation is possible. Imagine that there are two possible worlds each which has an extra hydrogen atom which is not related to anything that actually exists. These two extra hydrogen atoms are thus ungrounded. We can furthermore imagine that these two hydrogen atoms are indistinguishable from each other. Now if this is the case then the two possible worlds we are imagining are indistinguishable yet distinct. There is nothing that we could say about the possible world in and of itself that could distinguish between these two worlds. The only way that we could distinguish them is by noting their relative positions to each other in the overall space of possible worlds. Thus Kripke concludes that we cannot talk about individual possible worlds except in relation to the entire structure of modality, or to be more modest, he concludes that there is no a priori consideration that would rule this kind of automorphism out.

An interesting argument, and there is a lot more to say about it but I will come back to that later.