The Philosophical Zombie Problem – A Proposal Of An Approach To Test For Sentience
June 18, 2010 by Oliver
The Zombie argument, as developed by David Chalmers was directed primarily at proponents of physicalism who assert that in real or hypothetical intelligence systems, the physical facts determine all other facts about consciousness and sentience (take note of the significance of the expression “all other” in this definition of physicalism). The main thrust of Chalmer’s argument was that one can coherently conceive of a world composed of nothing but zombies; a zombie being a system indistinguishable from a human being in all respects(including, presumably, any identifiable “neural correlates of consciousness”), but lacking in conscious or sentient experience or “qualia.” Chalmer’s hypothetical world would be a world indistinguishable from our world, down its micro-details, but lacking in conscious beings which are capable of conscious experience. The physicalist argument denies this proposition by arguing that since the physical facts(of the neural correlates of consciousness, for instance) determine “all other” facts concerning consciousness, it is not possible to have beings indistinguishable from human beings, but lacking in conscious experience or sentience. While Chalmers insisted that his proposition is logically possible for our world, physicalists like Dennet rejected the idea as “metaphysically” impossible and therefore, also, “logically” impossible.
It would appear that Chalmers weakened his proposition of the possibility of philosophical zombies by trying to use it against physicalism. The argument is best used against that particular specie of physicalism called behaviorism, which asserts that mental states exist as nothing but behavior (i.e. observable psychological traits) and neural mechanisms. One can counter behaviorism by considering the possibility of an android machine indistinguishable from a human being in observable behavioral traits; that is, psychological traits, both emotional and intellectual, but which is lacking in sentient experience or qualia, and possibly consequently, also, in the so-called “neural correlates of consciousness.” Needless to say is the fact that such a machine, composed of bolts, nuts, silicon chips and electronic circuits is feasible at the current state of scientific theoretical knowledge even if not at the current state of technological know-how; but given the fact that scientific theory and technology is yet to make any progress towards identifying the “neural correlates of consciousness”(which physicalists are dogmatically convinced exist), or even say exactly what sentience is, talk-less of implementing it in Artificial Intelligence systems, we can ignore the problem of the “neural correlates of consciousness” in our definition of indistinguishability with regard to our hypothetical zombie. Of course, a physicalist might prefer to persist in the argument (possibly correctly) that the physical facts of neural circuitry, composed either of organic matter in a biological system or inorganic components as in an Artificial Intelligence or machine system, may be sufficient to implement and thus determine the “emergent” phenomenon of sentience. The fact is that we do not have any conclusive evidence for or against such argument. One can either accept it or reject it depending on one’s current philosophical predilection!
The philosophical zombie argument can readily be used to challenge not the physicalist explication which insists that there are “neural correlates,” which determine “all other” facts of consciousness, but behaviorism which asserts that there is nothing to a physical human system besides the observable facts of behavior and hypothetical “correlated” neural functions. In the behaviorist position we do not speak of the “neural correlates” of consciousness, for the behaviorist insists that neural activity is consciousness itself. The question raised by this apparently counter-intuitive position is: Are our intuitive notions of sentience mistaken? Are we really different in any way from a hypothetical android machine? Have we been misconceiving behavior in the intuitional language of sentience?
Some physicalists actually take the position that we have been misconceiving behavior in the intuitional language of sentience by arguing that the concept of sentience is an incoherent concept and therefore, in effect, only a fancy word for behavior and neural function. That, at least, is the position some prominent behaviorists appear to have taken.
One might, on the other hand, point out the fact that there is a tendency among “scientific thinkers,” anxious to identify with trendy materialistic thought, to equate observable form and function with consciousness and sentience, and thus, represent the notion of consciousness only in terms of functions implementable in Artificial Intelligence systems. The Turing Test, for instance, is widely misunderstood, especially among laymen, to be a test for consciousness(whatever consciousness is). Alan Turing’s original ill-posed question, “Can machines think,” leads, in the context of Descartes’ famous maxim, “I think therefore I am,” to the notion that if machines can think then they must be conscious entities. Here we have a confusion of computationally implementable logic with what humans intuitively understand as “thinking” in the sentient sense of “self-talk.”
It has for long been asserted that, regardless of what sentience might be, if it is really anything at all, there is no scientifically objective way of testing for it, and as such scientists would better ignore the problem of sentience, given especially, what appears to be the fact that ignoring human self-reports of sentience in scientific research make no difference to scientific theory and thought about human behavior. It may, however, be pointed out that if the physicalist notion that there are “neural correlates” of consciousness waiting to be discovered is true, then scientists might have, in the long run, to abandon the current position and begin investigating the objective implications of the existence of such “correlates” of sentience in human beings unambiguously reporting sentient experience.
This article suggests a conceptual model for testing for sentience and deriving results which may prove useful for narrowing down on the possible “neural correlates of consciousness,” by comparing aspects of measurable neural function in persons scoring high on the sentience test with those scoring low. One is, of course, making the presumption that individuals differ in their levels of sentience. Why do we make this presumption? Is it a valid or reasonable presumption to make about human populations?
To begin with, the test set up is not based on any formal definition of what sentience is or is not. It is not even assumed, at the outset, that sentience is “real,” or is a cognitive illusion(as some behaviorists insist). These considerations are irrelevant to the success of the test.
The main consideration of relevance, in design and implementation of this test, is that, based on human self-reports, sentience(whatever it is) is a wholly subjective private experience and therefore, the only approach available to us, at the moment(given our inability to identify its neural correlates), is to test for sentience by the respondents’ self-reportage. The test relies on the self-reportage of respondents under constraints imposed, in the test situation, in such a manner that the respondents’ self-report can be reasonably taken as reliable for the purpose of the test. We need to appreciate, from the fact of wholly private nature of sentience experience, that I do not subscribe to the overbearing professional view that some “expert” mind theorist or cognitive scientist can assert or deny a respondent’s sentience on his behalf. I am of the view that, given the wholly private nature of sentient experience, it is fruitless exercise for armchair behaviorist-theorist and his non-behaviourist opponent to argue over the question of whether sentience is an illusion or cognitive misconception or not. In spite of the trendiness of “objective,” measures in scientific research, we are subjected to the constraint of reliance, even if grudgingly, on the respondent as the final authority on the subject of his own subjective status with regard to the condition of sentience. The function of the person administering the test is only to guide the respondent in a manner which imposes constraints on what the respondent may misunderstand as, or wrongly identify as sentience in himself.
The test itself is deceptively simple but potentially powerfully revealing of the system’s subjective self-experience, or otherwise, of sentience:
TEST QUESTION: Imagine an interaction at “arms length” with a system you know is an android but is indistinguishable from a human being in all observable behavioral traits. Besides the fact of material constitution, how would you begin conceptualizing your difference(s), as a human being, from such a machine? Give you response at any desired length. Unlimited time is allowed.
The conditions for the test are stated below:
(1)The test is best administered on “naive” subjects. That is, individuals who have had no formal training in the mind sciences, or in philosophy of mind, nor have independently been exposed to relevant literature in that field. This means, effectively, individuals who hold only “folk notions” of whatever is understood or meant by sentience in the given population or culture.
(2)The “arms length” condition implies an interaction with the system under circumstances which disguise the fact that the android is not made of organic material. That is, the machine is made of “synthetic” material which effectively mimics human physical material constitution from the circumstance of a casual perspective in which the observer does not have the opportunity of close or probing physical examination which might reveal or make obvious the fact that the android is made of “synthetic material.”
(3)Unlimited time is allowed to give the respondent time to think originally and make the best responses possible in the circumstance of his previous experience.
We can anticipate two broad categories of responses:
(1) Respondents who, being confused by the question, make irrelevant responses.(e.g. “If you can eat it is sentient. If you can’t eat it then it is not sentient”).
(2) Respondents who, being unable to conceive of any additional element or function in themselves apart from outward behavioral traits, effectively assert that outward behavior is equivalent to sentience.
(3) Respondents,who insist, effectively(regardless of non-specialist or “naive” language in which the response is expressed), that there is an additional subjective aspect to human constitution that is distinct from mere outward appearances of behavior and function.
IQ test means Intelligence Quotient,with the help of IQ tests intelligence test you can measure your ability of your work in any field, if you get high IQ level in IQ intelligence test that means there is chances of highly complex jobs.
Related posts:
