CSC 190A/290A/08, A S S I G N M E N T 1 ============================================ DUE: WED., FEB. 6, 5pm (hand in to TA) NOTE: Questions should be answered independently by each student. (That's not to say you shouldn't talk with others about the ideas encountered in the readings and your thinking about them -- by all means do!) Any text taken from elsewhere should be quoted and attributed (to avoid those dreaded charges of plagiarism). Keep your answers as crisp and tightly reasoned as you can. It's better to say a few things clearly and persuasively than to keep writing just to fill the page. But do have fun -- this is intriguing, and quite accessible, material, so doing some digging and thinking and writing should be more of a pleasure than a pain! ALL WORK SHOULD BE IN ELECTRONIC FORM, BUT PRINTED OUT FOR HANDING IN. No handwritten papers, unless written in the finest and most legible medieval calligraphy, and suitably illuminated. -=-=-=-=-=-=-=-=-=-=- 1. (General/preliminary questions) ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ a. What is Pinker's point, in relation to consciousness, in his discussion of Star-Trek-like "beaming" of a person from one place to another? Is it the same point as Blackmore's in ch.5? b. What is meant by "mind reading", as the term is used in the Economist article, "Who do you think you are?" c. Is Teuscher, in CACM, Nov.'06, arguing against biologically motivated AI, or is he arguing *for* something? Or both? Explain in a few sentences (up to a paragraph). d. What is your current reasoned view on whether a robot could feel pain? (A "reasoned view" here means a view backed by the most careful argumentation you can manage. Opinions are cheap, but reasoned views are more difficult to come by :-) You might consider both what is possible with current technology and AI techniques, and what might be possible in principle. Answer in about a page (no more than two). You need not, at this point, back up your discussion with arguments and facts to be found in the literature; and you need not reach a firm conclusion; just give a thoughtful discussion. You might think about the "partial zombies" question (2) before answering. 2. Zombie jamboree -- essay question ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ The idea of "zombies" may be absurd (as Blackmore claims) if we understand it in the strong sense of someone that seems in all respects like an intelligent alert, normal person, yet has NO ACTUAL PHENOMENAL EXPERIENCE WHATSOEVER. However, "blindsight" is one bit of evidence that some visual perceptions that are conscious in fully sighted people may be unconscious, yet present, in people with damage to their visual cortex. More generally, it is really only the final, high-level stages of perception that register consciously -- e.g., we do not register the process of edge detection in vision, or the complex "signal processing" of the auditory stream in speech recognition. This should make us wonder whether the idea of "zombies" is really so far-fetched. However, such unconscious processing is "inferior" to conscious processing, notably in that it is inaccessible to *language*: people with blindsight cannot *talk* about what they are perceiving subliminally, and even people with normal sight cannot *talk* about the edge-detection, or auditory signal processing that they are performing. Another line of attack on the anti-zombie stance is to consider ways of "faking" perception. First of all, a seemingly perceptually aware, conscious creature might just be remote- controlled. But even setting aside remote control, we can imagine at least "partial zombies" that are in some way faking perceptions that they don't actually experience. For example, a partial zombie might be a blind man with a vision gadget, disguised as sunglasses, that whispers detailed descriptions of what's in front of him in his ears - allowing him to act and talk just as if he could see. Someone might ask, what do you see in front of you?", and he might reply, "Well, I see you, with your baseball cap and striped shirt, and I see the blue sky behind you", based on what the vision gadget is whispering to him. You might say he's lying, and maybe that's his own view as well. Or maybe he has become so accustomed to the device that he has begun to think of the whispering voice as what seeing is all about. (BTW, visual prostheses have been tested with some success that consist of a flat grid with embedded electrical nodes placed on the tongue, displaying images captured by a digital camera as patterns of electrical stimulation.) Perhaps even more plausibly suppose a certain deaf woman is equipped with glasses that contain a microminiaturized speech and sound recognition system that displays the results (words spoken, or sounds heard --- "woof, woof", "squeal", and so on) to her visually, in a way that is undetectable by onlookers. Let's assume she also speaks perfectly well (this is unlikely for a person deaf from birth, but perhaps she became deaf as an adult). Similar partial zombies can be imagined lacking "real" touch sensitivity (perhaps just on some part of the body, like an arm), or a sense of taste, etc. This notion raises some interesting questions, in particular: a. Can it be said of the vision-zombie that he "can see", or of the auditory zombie that she "can hear"? (In considering this, you might also imagine even more advanced prostheses implanted in the body, made of organic materials, that feed their information directly to the appropriate brain areas (e.g., the auditory regions in the case of the blind man -- cf. synesthesia. In such cases the prostheses can really be treated as part of the person.) b. Suppose the visual zombie loses more and more senses (perhaps because of a degenerative disease), except for hearing. For each additional loss -- touch, taste, olfactory, etc., he receives an additional prosthetic device that conveys information in auditory form. Let's let him keep his sense of balance, and proprioceptive capabilities, which are in any case less prominent in human consciousness than the primary senses; we don't want him to fall over and be unable to act. This seems to show that "near-zombies" are possible -- someone who claims to hear, see, feel, taste, and smell yet "really" only hears. Does this show that a weakened form of Chalmer's claim holds, i.e., that subjective experience *is* something "extra"? c. Does this tell us anything about robot consciousness? Write an essay -- a reasoned commentary -- of about 3 pages (1500 words) on issues (a), (b) and (c). If you detect any faulty assumptions in the formulation of the issues, point them out, and say whether they can be "patched up" or not. Apart from that, draw the strongest conclusions that you think can be adequately supported (and of course, make the case for them!) The essay should clearly state its overall thrust (or thrusts) at the outset, and then proceed systematically, clearly stating each point you are establishing, and the reasoning leading to it. (Or, for issues that you believe remain unresolved, give good reasons why it is not possible -- perhaps at this time -- to resolve them.) This is not an easy question -- you should expect to spend a good deal of time trying to gain some clarity of understanding, and to do a significant amount of rewriting, before you have a satisfactory essay. 3. Identity theory, stream of consciousness ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ a. The recently posted notes for lecture 2 state, "Blackmore objects to the identity theory -- how can the pain *be* the above activation pattern?". Let's translate this question into the instructor's "artificial gravity" analogy, i.e., how can the outward gravity experienced by the rim-rider in the rotating enclosure *be* the manner of motion and acceleration of the rim-rider? Is this analogous identity claim true or false? If true, explain why. If false, is one causing the other? Are they both effects of a common cause? Is none of these alternatives true? Answer in a few sentences (up to a paragraph). (The answer will be graded on the degree of understanding and clarity of thinking shown, not on agreement with the instructor. For example, perhaps you want to find fault with the analogy.) b. Let's set aside Blackmore's pessimism about making sense of William James "stream of consciousness", and instead ask whether a robot might also be able to sense, and report on, such a stream of consciousness, particularly a *visual* stream of consciousness. What sorts of hardware, movement and control capabilities, and internal representational and processing capabilities would a robot need for this? This is a tough question, but we want a brief, perhaps bullet-by-bullet answer, trying to be as clear as you can on this topic, in a page or less.