"andy-k" wrote in
news:YI%%7k.84661$P83.17850@newsfe20.ams2:
> Perhaps I'm using the words wrongly here, but I was under the
> impression that it is 'realism' that presumes an external world, and
> that 'empiricism' pertains to what can be ascertained through
> experience (not necessarily experience of anything external, but
> rather what arises within experience, and particularly in that aspect
> of experience categorized as "sense data", whatever it's source).
> Either way, this seems something of a digression at this point.
Realism posits an external world. Empiricism then then proposes that
knowledge of that world can only be gained by observing it. Realism is a
metaphysical theory; empiricism an epistemological theory which presumes
the realist metaphysic. As you note, empiricists do not regard all
experience as informative (dreams, hallucinations, fantasies, etc. are
disregarded). Only sensory experience is regarded as informative, since,
per realism, it conveys information about an external world. If an external
world is not presumed then sensory experience has no special role or
status. [Note that I'm using "realism" there in its broadest sense.]
But you're right, it is a digression.
>> Well, if solipsism is assumed, then there is no "communication,"
>> since communication is understood to be transmission of information
>> from one mind (or other sending source) to another mind (or other
>> receiver). If there is only one mind, mine, then what I'm calling
>> "communication" is just noise of inexplicable origin.
> If that's the way you stipulate your use of the word 'communication'
> then your objection stands, but I would stipulate a definition that
> also pertains to the transfer of information between machines (cf.
> Shannon and Weaver's classic paper entitled "The Mathematical Theory
> of Communication").
Sure. That's why I said "or other sender/receiver." But if we are
solipsists there is no "other," either human or machine. We are carrying on
both sides of the conversation, so to speak. Nor is any information
exchanged, since "exchange," like "communication," presumes a second party.
>> So how do we move beyond solipsism? Or do you assume there is no
>> rational ground for doing so?
> I don't think we have to 'move' anywhere -- our instinctive empathy
> commits us to a default belief in other minds. My rejection of
> solipsism rests on nothing more than an in-built prejudice, and I
> accept it on those terms.
There are no rational grounds for rejecting solipsism, then? Can in-built
prejudices be overcome? Can they be misleading or outright wrong?
>> Empathy is not a behavior pattern. It is one of those internal
>> subjective states we are discussing, of the same class as love,
>> admiration, jealousy, emnity, etc. --- i.e., affective responses we
>> may experience in the presence of other people (or some animals). It
>> is not a source of *knowledge* regarding those other entities, i.e.,
>> it will not inform us whether the entity before us is alive, is
>> human, is in pain, or is conscious. We have to make those judgments
>> *before* the affective state of empathy can make an appearance.
>>
>> You are trying to promote empathy, an affective subjective state,
>> into some kind of perceptual or cognitive tool. That was the point of
>> the gorilla- thing example --- whether "empathy" would allow you to
>> decide among the alternative explanations (and you seemed to agree it
>> would not). The question is, How do we *decide* whether another
>> entity is conscious, in pain, etc., not how we feel about that entity
>> once we've decided.
> I'm saying that I don't think we 'decide' whether another
> entity is conscious, in pain, etc. -- we respond *instinctively*:
Well, again, that was the point of the gorilla-thing story. We can't
*respond* (at least not appropriately) until we have ascertained the nature
of the stimulus. The question is, How do we make that determination? You
seem determined to skip over that step.
To be sure, in most everyday encounters between people appearance alone
suffices to identify the creature before us as a person. We have no reason
to think about it consciously, and usually don't. But before we begin
speaking to the entity before us, or otherwise responding to it, we will
have determined the entity to be a person, by some means or other. We don't
launch into conversation with mannequins in department stores. Only when
the visual cues are unclear, or anomalous, are we obliged to ponder that
question consciously: "Is this a person?"
I'm not denying that we may have some built-in response sequences to
particular stimuli. E.g., mothers (of most mammal and bird species) respond
instinctively to a crying infant (or kitten or puppy). Most animals
instinctively avoid fire and abrupt edges of high surfaces. But before any
of those responses can be invoked the organism must identify the stimulus.
I have two cats. Several years ago I bought a plaster cat statue to
decorate the floor under a glass table. The statue was about the size of a
Maine Coon. Soon after I placed it one the cats wandered down from
upstairs. She passed the table, did a double-take, stopped in her tracks,
arched her back, and hissed at the statue. But then she relaxed, walked up
to the statue, sniffed it all around, and thereafter ignored it. Clearly
the cat responded instinctively, after deciding there was a strange and
very large new cat in her house, based on its immediate appearance. But
then she investigated further and decided the statue was not a cat after
all. (She then gave me a long look: "I suppose you think you're funny. . .
.")
Saying a response is instinctive does not answer the question about how we
identify the stimulus.
>> What would it take for the biologists to pronounce this organism
>> "conscious?" Would empathy or "instinct" suffice? Or would you simply
>> conclude it is conscious a forteriori (per panpsychism)?
> Ordinarily we would have no grounds for pronouncing this organism
> 'conscious' since it is too far removed from other members of our own
> species. Our instinct gives us no room to doubt that other members of
> our species are conscious (no decision taking place here at all),
Of course it does. We often cannot tell by appearance alone --- and
certainly not by "instinct" --- whether another human is alive, dead, brain
dead, in a coma, or just "playing possum." That's why we check for pulse,
check for pupil response, do EKGs, etc. Whenever a person (or something
that appears to be a person) is not exhibiting behaviors we take as
criterial for consciousness, we have room for doubt about whether he is
conscious, and are obliged to investigate further (just like the cat).
One implication of your comment there is that for any organism "too far
removed" from our species, we could never have any grounds for imputing
consciousness to it. That implies, in turn, that consciousness must be
impossible among extraterrestrials (unless they miraculously happen to be
like us).
Suppose one of the organisms in my story began sprouting shoots up from its
roots, each topped by a small red blossom. A few months after this blooming
begins one of the scientists is perusing an aerial photo of the forest in
which this specimen is rooted, and notices that the pattern of the red
blossoms, which are arrayed across several hectares, spell out very
clearly: "Invaders go home --- or else."
Would we then consider it conscious?
>> I'm not clear on this. Are you saying we have an "instinctive belief"
>> that a responsive system (I presume of the right kind) is conscious
>> and an "instinctive belief" that a nonresponsive system is not
>> conscious,
>
> Yes (see above).
>
>
>> i.e., that this instinct tells us whether the system is conscious or
>> not, independently of the behavioral evidence? Or are you saying that
>> we have no grounds for making that distinction at all?
>
> No, the instinct doesn't "tell us whether the system is conscious or
> not" -- it predisposes us to respond in a particular manner to certain
> systems (particularly other people).
Again, responding and identifying are two different things.
>> Ah, so that is where the denying the antecedent comes from.
>> Your mistake is in the premise: it is not a conditional but a
>> biconditional, e.g.,
>>
>> B <--> C, not B --> C. That is because C (consciousness) is a
>> *hypothetical construct* imputed *only* to systems evincing certain
>> characteristic behaviors in certain circumstances.
> Stipulating that as your definition of consciousness your argument
> is consistent, but other people are working to a variety of different
> definitions to the one you stipulate, and given their own definitions
> their own arguments may also be consistent. I think your aim here is
> to show me that my definition and argument are inconsistent, but your
> reasoning as far as I have been able to discern it rests on the claim
> that your definition is the right one and mine is the wrong one. I'd
> need a better understanding of why my definition is the wrong one,
> other than that it is different to yours, if I am to recognize the
> inadequacy of my position.
For clarity I'll give my definition of "consciousness" again (I gave this
in a response to impromptu, which you may not have seen):
------------
"Consciousness" has two senses in ordinary discourse:
1. A system is conscious if it has a *capacity* for having various
subjective states, e.g., unique states correlated with the smell of coffee,
the color of a rose, different pains, etc., and also such affective or
motivational states as anger, desire, sadness, fear, contentment, etc.
2. A system is conscious if it is currently in one or more of those
states.
The first is the capacity sense, the second the occurrent sense. A
person who is sleeping is conscious in the first sense, but not the
second. He is *unconscious*. A dead person or a stone is conscious in
neither sense. It is *nonconscious*.
The *evidence* that a system is conscious in either sense, however, is
behavior, and secondarily, structure. Conscious systems (in either
sense) evince various characteristic behaviors. But those behaviors are
not the *definiens* of consciouness, any more than tracks in the snow
are part of the *definiens* of a moose. They are merely evidence of a
moose. The subjective states implied by the term "conscious" are
unobservable, persisting or recurring states *imputed* to a system in
order to explain that system's observable behaviors. They are
intervening variables, just as a "talent" is an intervening variable
imputed to a pianist to explain her ability to produce pleasing music.
We impute those states to a system if it evinces the requisite behaviors
in the requisite circumstances; we declare that it is conscious in the
second sense. If it has displayed those behaviors in the past, and still
appears to be functioning normally, we take it to be conscious in the
first sense, even though it is now quiescent.
There is nothing at all mysterious about this concept. Consciousness,
and the subjective states comprising it, functions analytically much
like such concepts as electrons, quarks, quantum foam, electromagnetic
waves, etc. --- non-observable hypothetical constructs postulated to
explain some observable phenomena. The constructs are considered "real"
if the theories which postulate them prove to have explanatory power,
i.e., are able to predict future observations.
-------------
As I said earlier, I don't think our definitions differ, at least not
materially. Where we seem to differ is in how we acquire that concept and
how it works in everyday language, and perhaps as to the epistemological
category to which it belongs.
> I agree that we learn the term via social conditioning, and your claim
> that the instinct "would not happen were the physical and behavioral
> cues not present" has never been in dispute. What I don't understand
> is how, in your scheme, you make the transition from learning the term
> to associating it with our own subjective states.
The same way we learn the terms for the subjective states themselves: as
children we are told (in effect) "this is blue," "that is yellow," etc. And
whenever something before us is described as "blue," we find ourselves
experiencing a unique sensation specific to those things (and different
from sensations we get from other things). So we learn to call that
specific sensation "blue," as well as the things that give rise to it.
Similarly, we learn that "conscious" applies to things that also, judging
by their behavior, can see colors, smell odors, hear sounds, etc. So we
suppose that "consciousness" refers to the ability to do those things,
which we have along with other persons and many animals. And just as with
the terms for the various states, we associate the term "consciousness"
with the various states we experience when we make those discriminations,
considered collectively. We impute similar states to others because others
seem to be like us in other ways, and because those states serve to explain
much of their behavior. E.g., "Why is that person sitting quietly and
listening to that music?" Because it provokes in him a sensation he finds
pleasant.
It is probably worth noting that most of us learn the adjectives
"conscious," "unconscious," and perhaps even "nonconscious" long before we
learn the abstract noun "consciousness."
>> That does not mean, of course, that we could not impute subjective
>> states to other kinds of entities if we wished, i.e., extend the
>> class of eligible objects. The question there would be, What would be
>> the advantage of doing so? How would doing so help us predict the
>> behavior of those entities?
> You're still assuming that the existence of subjective states
> associated with a system *necessarily* involves behavior, and whilst I
> understand that this is the way you are defining the word
> 'consciousness', surely you must accept that your definition is not
> the only way the word is understood. Your last question above then
> becomes irrelevant -- the advantage lies not with the prediction of
> behavior but with hypotheses concerning the origin of subjective
> states.
What is necessary is that the terms "subjective states" and
"consciousness" have some public truth conditions. It is a matter of
historical and linguistic accident that those terms were chosen to explain
certain kinds of behavior, i.e., that they were coined to denote a
hypothetical construct presumed to underlie those behaviors. We can easily
imagine that the terms could have had other meanings. What we can't imagine
is that they could have acquired any currency in public languages without
public truth conditions.
Well, I guess we could imagine that --- the language is riddled with
noncognitive terms. :-) But "conscious" and its relatants are common coin
in everyday speech, and do have public (and not very controversial) truth
conditions.
Either "consciousness" and "subjective states" have public truth conditions
or they are noncognitive, i.e., there is no way in principle to distinguish
things which have them from things which don't. The world looks the same
either way. They become pseudo-properties whose attribution to any
particular thing conveys no information about that thing.
>> The gist of our disagreement seems to turn on the origins of our
>> understanding of that concept (consciousness). You seem to think that
>> we have some "instinctive" (innate, prelinguistic) grasp that it
>> refers to our experienced subjective states. My claim is that there
>> are no innate, prelinguistic grasps of any terms; the meanings of all
>> of them are arbitrary and can be learned only in speech communities,
>> by observing the contexts of their use. We can't apply terms in the
>> public language to any inner states until we have learned their
>> public truth conditions.
> This is why I speak about the *idea* of consciousness and not the
> *concept* of consciousness. If the concept necessarily invokes the
> term, then this is a case in which the idea (that other people have
> subjective states) precedes the concept.
I wouldn't disagree with that. Its quite possible that prelingual people
and perhaps even animals impute (unreflectively) subjective states to
others, and not only others of their own species. Can a cat tell when
another cat (or its human) is angry? Probably, but they will do so on the
basis of the other animal's appearance and behavior. It will be up to us
linguistic types to spell out the relationships between the states and the
behavior.
Why would we impute this ability (to impute anger to another animal) to the
cat? Because it helps explain the cat's behavior (if my cat thinks I'm
angry at her, she'll lay low for a while until the storm blows over).
> The belief that there is an engine under the hood is testable, whereas
> the belief in another's subjective states is not. This is not a
> dispute about whether or not a belief is justified by induction, but
> about whether or not a belief is testable.
Sure it's testable. It's just not testable by direct observation. What is
testable are the behavioral implications of imputing those states.
> The two competing hypotheses are that subjective states are emergent
> and that subjective states are fundamental. Shouldn't both of these
> hypotheses count as possible explanations unless and until we have
> some means of eliminating one of them?
Yes. And we have a means of eliminating the "fundamental" option. It is
that imputing that property to most things has no testable consequences,
i.e., it has no explanatory power (it is noncognitive). So we eliminate
that option via Occam's Razor.
