The Philosophical Zombie Problem - A Proposal of an Approach to Test For Sentience

The Zombie argument, as developed by Davidmay prove useful for narrowing down on the
Chalmers was directed primarily at proponents ofpossible "neural correlates of consciousness," by
physicalism who assert that in real or hypotheticalcomparing aspects of measurable neural function
intelligence systems, the physical facts determinein persons scoring high on the sentience test with
all other facts about consciousness and sentiencethose scoring low. One is, of course, making the
(take note of the significance of the expressionpresumption that individuals differ in their levels of
"all other" in this definition of physicalism). Thesentience. Why do we make this presumption? Is
main thrust of Chalmer's argument was that oneit a valid or reasonable presumption to make
can coherently conceive of a world composed ofabout human populations?
nothing but zombies; a zombie being a systemTo begin with, the test set up is not based on
indistinguishable from a human being in allany formal definition of what sentience is or is
respects(including, presumably, any identifiablenot. It is not even assumed, at the outset, that
"neural correlates of consciousness"), but lacking insentience is "real," or is a cognitive illusion(as some
conscious or sentient experience or "qualia."behaviorists insist). These considerations are
Chalmer's hypothetical world would be a worldirrelevant to the success of the test.
indistinguishable from our world, down itsThe main consideration of relevance, in design and
micro-details, but lacking in conscious beings whichimplementation of this test, is that, based on
are capable of conscious experience. Thehuman self-reports, sentience(whatever it is) is a
physicalist argument denies this proposition bywholly subjective private experience and
arguing that since the physical facts(of the neuraltherefore, the only approach available to us, at
correlates of consciousness, for instance)the moment(given our inability to identify its neural
determine "all other" facts concerningcorrelates), is to test for sentience by the
consciousness, it is not possible to have beingsrespondents' self-reportage. The test relies on the
indistinguishable from human beings, but lacking inself-reportage of respondents under constraints
conscious experience or sentience. While Chalmersimposed, in the test situation, in such a manner
insisted that his proposition is logically possible forthat the respondents' self-report can be
our world, physicalists like Dennet rejected thereasonably taken as reliable for the purpose of
idea as "metaphysically" impossible and therefore,the test. We need to appreciate, from the fact of
also, "logically" impossible.wholly private nature of sentience experience,
It would appear that Chalmers weakened histhat I do not subscribe to the overbearing
proposition of the possibility of philosophicalprofessional view that some "expert" mind
zombies by trying to use it against physicalism.theorist or cognitive scientist can assert or deny a
The argument is best used against that particularrespondent's sentience on his behalf. I am of the
specie of physicalism called behaviorism, whichview that, given the wholly private nature of
asserts that mental states exist as nothing butsentient experience, it is fruitless exercise for
behavior (i.e. observable psychological traits) andarmchair behaviorist-theorist and his
neural mechanisms. One can counter behaviorismnon-behaviourist opponent to argue over the
by considering the possibility of an androidquestion of whether sentience is an illusion or
machine indistinguishable from a human being incognitive misconception or not. In spite of the
observable behavioral traits; that is, psychologicaltrendiness of "objective," measures in scientific
traits, both emotional and intellectual, but which isresearch, we are subjected to the constraint of
lacking in sentient experience or qualia, andreliance, even if grudgingly, on the respondent as
possibly consequently, also, in the so-called "neuralthe final authority on the subject of his own
correlates of consciousness." Needless to say issubjective status with regard to the condition of
the fact that such a machine, composed of bolts,sentience. The function of the person
nuts, silicon chips and electronic circuits is feasibleadministering the test is only to guide the
at the current state of scientific theoreticalrespondent in a manner which imposes constraints
knowledge even if not at the current state ofon what the respondent may misunderstand as,
technological know-how; but given the fact thator wrongly identify as sentience in himself.
scientific theory and technology is yet to makeThe test itself is deceptively simple but potentially
any progress towards identifying the "neuralpowerfully revealing of the system's subjective
correlates of consciousness"(which physicalists areself-experience, or otherwise, of sentience:
dogmatically convinced exist), or even say exactlyTEST QUESTION: Imagine an interaction at "arms
what sentience is, talk-less of implementing it inlength" with a system you know is an android but
Artificial Intelligence systems, we can ignore theis indistinguishable from a human being in all
problem of the "neural correlates ofobservable behavioral traits. Besides the fact of
consciousness" in our definition of indistinguishabilitymaterial constitution, how would you begin
with regard to our hypothetical zombie. Of course,conceptualizing your difference(s), as a human
a physicalist might prefer to persist in thebeing, from such a machine? Give you response
argument (possibly correctly) that the physicalat any desired length. Unlimited time is allowed.
facts of neural circuitry, composed either ofThe conditions for the test are stated below:
organic matter in a biological system or inorganic(1)The test is best administered on "naive"
components as in an Artificial Intelligence orsubjects. That is, individuals who have had no
machine system, may be sufficient to implementformal training in the mind sciences, or in
and thus determine the "emergent" phenomenonphilosophy of mind, nor have independently been
of sentience. The fact is that we do not haveexposed to relevant literature in that field. This
any conclusive evidence for or against suchmeans, effectively, individuals who hold only "folk
argument. One can either accept it or reject itnotions" of whatever is understood or meant by
depending on one's current philosophicalsentience in the given population or culture.
predilection!(2)The "arms length" condition implies an
The philosophical zombie argument can readily beinteraction with the system under circumstances
used to challenge not the physicalist explicationwhich disguise the fact that the android is not
which insists that there are "neural correlates,"made of organic material. That is, the machine is
which determine "all other" facts of consciousness,made of "synthetic" material which effectively
but behaviorism which asserts that there ismimics human physical material constitution from
nothing to a physical human system besides thethe circumstance of a casual perspective in which
observable facts of behavior and hypotheticalthe observer does not have the opportunity of
"correlated" neural functions. In the behavioristclose or probing physical examination which might
position we do not speak of the "neuralreveal or make obvious the fact that the android
correlates" of consciousness, for the behavioristis made of "synthetic material."
insists that neural activity is consciousness itself.(3)Unlimited time is allowed to give the respondent
The question raised by this apparentlytime to think originally and make the best
counter-intuitive position is: Are our intuitiveresponses possible in the circumstance of his
notions of sentience mistaken? Are we reallyprevious experience.
different in any way from a hypothetical androidWe can anticipate two broad categories of
machine? Have we been misconceiving behavior inresponses:
the intuitional language of sentience?(1) Respondents who, being confused by the
Some physicalists actually take the position thatquestion, make irrelevant responses.(e.g. "If you
we have been misconceiving behavior in thecan eat it is sentient. If you can't eat it then it is
intuitional language of sentience by arguing thatnot sentient").
the concept of sentience is an incoherent concept(2) Respondents who, being unable to conceive of
and therefore, in effect, only a fancy word forany additional element or function in themselves
behavior and neural function. That, at least, is theapart from outward behavioral traits, effectively
position some prominent behaviorists appear toassert that outward behavior is equivalent to
have taken.sentience.
One might, on the other hand, point out the fact(3) Respondents,who insist, effectively(regardless
that there is a tendency among "scientificof non-specialist or "naive" language in which the
thinkers," anxious to identify with trendyresponse is expressed), that there is an additional
materialistic thought, to equate observable formsubjective aspect to human constitution that is
and function with consciousness and sentience,distinct from mere outward appearances of
and thus, represent the notion of consciousnessbehavior and function.
only in terms of functions implementable inOf course, we must tolerate a degree of
Artificial Intelligence systems. The Turing Test, forfuzziness in a test that must, by virtue of its
instance, is widely misunderstood, especiallynature, be imprecise in its results. For instance,
among laymen, to be a test forsome individuals will make responses based on
consciousness(whatever consciousness is). Alanpreconceptions, or culturally acquired notions of
Turing's original ill-posed question, "Can machineswhat the right answer should be, rather than
think," leads, in the context of Descartes' famouswhat they really think it is. Respondents can also
maxim, "I think therefore I am," to the notionbe expected to show different levels of certainty
that if machines can think then they must beas to the accuracy of their self-report, which
conscious entities. Here we have a confusion ofleads to the need for the design of a necessarily
computationally implementable logic with whatimprecise "Sentience Quotient Scale." A large
humans intuitively understand as "thinking" in thepopulation sampling would be necessary to cancel
sentient sense of "self-talk."out for the effect of imprecision on the part of
It has for long been asserted that, regardless ofrespondents. The selection for subsequent
what sentience might be, if it is really anything atanalysis, of differences in neural function which
all, there is no scientifically objective way ofmay be correlated with consciousness would be,
testing for it, and as such scientists would betterof course, between individuals on the extremes
ignore the problem of sentience, given especially,of the scale.
what appears to be the fact that ignoring humanThe potential objective significance of this test is
self-reports of sentience in scientific researchthat it might be found to be useful data for
make no difference to scientific theory andlocating those aspects of neural function which
thought about human behavior. It may, however,correlate with self-report of sentience in human
be pointed out that if the physicalist notion thatbeings. A more far-reaching potential significance
there are "neural correlates" of consciousnessof this test is that, we might be able to answer a
waiting to be discovered is true, then scientistsquestion that has vexed me for a long time. Why
might have, in the long run, to abandon thedo even specialists in the cognitive sciences and
current position and begin investigating themind theory disagree so significantly with each
objective implications of the existence of suchother over the "hard problem" of consciousness?"
"correlates" of sentience in human beingsIs it possible that human populations are so
unambiguously reporting sentient experience.significantly variable in levels of sentience as to
This article suggests a conceptual model forwarrant the categorization of human populations
testing for sentience and deriving results whichinto zombies and sentient humans?