(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.

Check out these amazing pictures from the Hubble telescope….woa…

(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.

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.

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!

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….

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.

Wow, it’s already September! I have had a very busy summer. I have just recovered from a massive hard drive failure which has cost me all of my photos and three new papers that I have been working on. Damn you Dell!! I know, I know; I should have backed it up. I don’t know what happened. I moved the damn thing to my new house (closer to LaGuardia) and it was working fine (so it didn’t break in the move) then it just stopped working. They tell me that no data can be recovered from it unless I am prepared to pay in excess of $2,000!! So that stuff is as good as gone. I should have been working on that stuff here instead of in a word file!!!! I did print out a hard copy of some of it, but even so I have substantially revised it since then…Oh well, live and learn. I am now the proud owner of a macbook pro and loving it…though I have Pages instead of Word and I am wondering if I should get MS Office. At any rate,  I am planning on posting the rework so I should be posting more often.

On a personal note, I got married! I just returned from my honeymoon (a seven day cruise to Mexico) and am trying to get my life back together again. Between the moving, the computer dying, and the two weeks out of town for the wedding/honeymoon, and the semester starting, I feel like everything is ass over teakettle.

I also recently became the editor of the Higher-Order Theories of Consciousness category over at PhilPapers, which I am loving. This is a fantastic resource that makes research in philosophy the way it should be: fun and exciting. And finally some exciting things are in the works for this year’s Consciousness Online and the official call for papers should be going out in the next week or two.

There has been a bit of discussion of zoombies in the blogosphere of late and I want to keep track of them all so that I can respond to them so I am posting links to them.

1.) Intentional Objects‘ David Gawthorne

Richard Brown’s Zoombies and Shombies

2.) Brain Scam’s Tony Alterman:

Zombie, Scmombie –Richard Brown’s Efforts to Ressurect Materialism

(and his reply to my reply) Return of the Zombie

3.) And then there’s Richard Chappell’s responses.

In our latest exchange he has acknowledged that the primary and secondary intensions of statements in Q may diverge but seems to think that translating those statements into “semantically neutral” language will still let the argument go through. So, just was “the watery stuff isn’t H2O” comes out true at Twin Earth, “the painful stuff isn’t C-Fiber firing” comes out true at the zombie world. But this move won’t work. Here is what Chalmers has to say about this:

Given the discussion above, one might try generating an anti-materialist argument by simply substituting primary possibility for metaphysical possibility in the original argument.

(1) P&~Q is conceivable

(2) If P&~Q is conceivable, P&~Q is 1-possible

(3) If P&~Q is 1-possible, materialism is false.

_______________

(4) Materialism is false.

On this reading, (1) and (2) are both plausible theses, but (3) is not obviously plausible. The reason is that materialism requires not the 1-impossibility of P&~Q but the 2-impossibility of P&~Q. That is, materialism requires that it could not have been the case that P were true without Q being true. This is a subjunctive claim about ordinary metaphysical possibility, and so invokes 2-impossibility rather than 1-impossibility.

A materialist might reasonably question (3) by holding that even if there is a world W verifying P&~Q, W might be a world with quite different ingredients from our own. For example, it might be that W does not instantiate true microphysical properties (those instantiated in our world), such as mass and charge, but instead instantiates quite different properties: say, pseudo-mass and pseudo-charge, which stand to mass and charge roughly as XYZ stands to H₂O. Likewise, it might be that W does not lack true phenomenal properties, but instead lacks quite different properties: say, pseudophenomenal properties. If so, then the possibility of W has no bearing on whether true microphysical properties necessitate true phenomenal properties. And it is the latter that is relevant for materialism.

Still, it may be that the gap between 1-possibility and 2-possibility could be closed. In particular, when a statement S has the same primary intension and secondary intension, then a world will verify S iff it satisfies S, so S will be 1-possible iff it is 2-possible. If P and Q both have primary intensions that coincide with their secondary intensions, then so will P&~Q, and we could run the following argument:

 

(1) P&~Q is conceivable

(2) If P&~Q is conceivable, P&~Q is 1-possible

(3) If P&~Q is 1-possible, P&~Q is 2-possible.

(4) If P&~Q is 2-possible, materialism is false.

_______________

(5) Materialism is false.

Here, the truth of (3) requires that both P and Q have primary and secondary intensions that coincide. (from The 2-D Argument Against Materialism)

I am now officially a reviewer for the newly established Philosopher’s Digest. This looks like it could potentially be a very valuable resource…I hope to have my first review up in a month or so…I’ll keep you posted.