The Turing Test: Affirming the Consequent?

[polytropos]

You need to explain why, come the day when we know as much about ourselves as we do about our machines, your argument won’t evaporate. So far you’ve only said there’s something mystic in the air which will deny us a full explanation.

Well, the main issue with your point #3 is that it’s blatantly question begging, “If we knew the architecture of the brain, we would know how thoughts progress.” That’s what’s being debated, so of course it can’t be used as an argument for the conclusion.

The other thing is that my argumentative strategy has been not at all as you’ve characterized it. I’ve “only said there’s something mystic in the air which will deny us a full explanation”? I have given arguments about why it seems to be in principle impossible to submit the mind to an explanation under methodological naturalism. The issue isn’t that we don’t know some stuff about the mind; it’s that explaining qualia or intentionality by way of a system that stipulates that subjective qualities are not empirically observable is nothing short of a category error.

But my point is that even without whatever it is you think will still be missing, once we know the language of thought then, as I’ve explicated constantly… the fact is that it has no semantic meaning.

But again, this just begs the question as to whether there is a “language of thought” (which seems dubious even if we grant materialism).

Unfortunately, you don’t answer my point. If you cannot come up with a test, you either have to accept Turing’s or accept that you are saying God cannot ever, under any circumstances, make a machine think. Either way, Turing has got you.

Not at all.
1). I’m not committed to saying that we have an ability to identify whether another species is intelligent.
2). I have given reasons why Turing’s fails to provide evidence that another species is intelligent.
3). I have given other reasons why it is doubtful that a machine could be intelligent.

I haven’t claimed that “God cannot ever, under any circumstances, make a machine think.” I’m pointing out that the machine’s ability to function in human conversation is not evidence of its thinking, and that the material facts don’t give us reason to believe that it thinks (and in fact give us reason not to believe that it does).

None of that has anything to do with God, unless you can provide some reason as to why God would make a machine think once some computer scientist manages to get it to cover all its bases and pass the Turing test. If God can make a machine think, then he can do it without recourse to absurd computational power (and without the help of software engineers). That is the main reason why the theological objection carried no force in the first place: even if a theist is committed to believing that a machine could be made intelligent by God, he is not committed to the involvement of a software engineer in the process, nor in the resulting intelligence being physically based (while the Turing test proponent, on the other hand, has a useless argument without those things).

What is confusing? Some materialists reject qualia because they are apparently unverifiable, subjective, and confused notions. Other materialists (Paul and Patricia Churchland, Alex Rosenberg) reject thoughts because they are apparently unverifiable, subjective, and confused notions. Such is the materialist project of denying the most basic aspects of human existence.

They should get together with ID fans, who also want to castrate science to let in their beliefs along with astrology and whatever else. Interesting that so many people idolize science to the extent that they want their superstitions to be let in.

See above. Also, the penultimate sentence would look just as good if you replace “qualia” with “evolution”.

This is frankly uncharitable and immature. Your position is completely insulated from criticism. First, in #151, you assume that every philosopher who believes in qualia is acting out of some religious prejudice and scientific ignorance. Given the names of a couple credible and well-renowned philosophers who are atheists (I also forgot Tallis, who is a former neuroscientist, from the list), you decide to liken them to superstitious “ID fans” and astrologists.

These are naturalists and atheists, some of whom are fairly well-versed in neuroscience. Unfortunately, I don’t think the ad hominem attacks and likening to the intelligent design bogeyman suffices to sweep them under the rug.

It means if there is no evidence for a hypothesis, and no logical requirement either, and no reason under the sun to support it, then bin it, why carry around useless baggage which is neither use nor ornament?

Well, fortunately we do have evidence for qualia, ie. the fact that we are not (or, at least, I am not) philosophical zombies. Hard to believe that knowing that we aren’t philosophical zombies is “useless baggage.”

 

[polytropos]

Your example of a meaningless statement makes my point. It is meaningful to us.

Agh, I mean it is not meaningful to us. Hopefully this doesn’t cause a massive miscommunication.

 

[polytropos]

Another thought to add to the intentionality/semantics related arguments I’ve been making.

A computer has “derived intentionality,” to use Searle’s term. It does not have intrinsic intentionality; the activity of a computer is not about things except as an extension of what human thoughts are about, ie. if I perform a calculation with the aid of a computer, then it is aiding my mathematical thoughts, but its performance is purely algorithmic, and if it’s a calculator or basic program, then there is not a sense in which it cares what operations it’s performing or which Arabic numeral the pixels returned on the screen constitute.

But now suppose that there is a machine that passes the Turing test. It therefore has the capability to simulate other entities which pass the Turing test. For instance, it simulates a video game in which two characters have a conversation. Since it can simulate human conversation, the conversation between the two virtual entities will be quite realistic and convincing.

Are the virtual entities passing the Turing test, though? That seems doubtful. Now, I am a human who (I hope) can pass the Turing test. So I can imagine a conversation between two people and play it over in my head. But the “entities” in my head are not passing the Turing test; they don’t exist, and the meaning of their conversation is only an extension of my own linguistic abilities.

On the contrary, the virtual entities generated by the computer seem to be on the same metaphysical par as the computer generating them (they have the same capabilities and are equally convincing in conversation). But my intuition is that it is absurd to say that they are thinking. But if the speech capabilities of the computer are equivalent to theirs, then we have to say that the computer is not thinking either.

(Someone could certainly challenge this intuition. I need to sleep, so I haven’t thought this experiment all the way through yet. But it seems a bit absurd to say that a thinking entity could contain a principally limitless number of other thinking virtual entities. And it also seems implausible to say that the machine is thinking while they are not.)

 

[inocente]

Well, are we still dealing with the same Turing test after all those precautions…?

Yes of course, it’s standard practice.

Another point is that if it is possible to “cheat” in the Turing test, then that test does not appear to be as useful, as it was supposed to be. That is, a machine that is not able to think can also pass it.

You keep coming back to this but it’s possible to cheat in any test or exam yet that doesn’t stop anyone. We’ve already been round this anyway, you isolate the machine and run scrupulous protocols to prevent cheating, it’s not rocket science, it’s standard experimental design.

The point is that if the computer can find the answers to other problems reasonably fast, it will have to be slower while answering something like “5643646346384634 + 6387617368786387 = ?”. And that is the type of question that can be foreseen and will be prepared for.

Yes of course. But probably the best environment would be a messaging system like CAF, where both the human and the machine can spend as long as they want to post their answers, then it’s not an issue. All this stuff is just a matter of protocol, scientists design experiments everyday, there’s nothing new here even though you may never have thought about it before.

*Unfortunately, it doesn’t get any better with “X=‘someone who can pass the Turing test’”… It is at least not certain that one needs to be able to think in order to pass the Turing test and it is almost certain that one can fail the Turing test while being able to think.

Also, I don’t think anyone is arguing in favour of that proposition with “X=human”. For example, we can probably agree that angels exist and that they are able to think, but angels are not humans. Thus they are a counterexample. By the way, I suppose that some books of the Bible do indicate that angels are perfectly capable of passing the “informal” Turing test…

Also we could happily agree that aliens capable of thinking are possible (maybe they exist, maybe not). They would not be “conventional” humans, thus they can also be a counterexample. By the way, it is not certain that they would pass the Turing test.

What is doubted here is something different: that the computer that passes the Turing test is to be considered a thinking being for that reason alone. By the way, if we got some special revelation from God, saying that a specific machine has been given the ability to think, that would be a different piece of evidence…*

I’m not sure what you’re saying, except I think you’ve ruled yourself out.

Turing invented a game by which it can be demonstrated that a machine does or does not think. No one on this thread has proposed an alternative, you guys have just picked holes. But by only picking holes and not proposing any alternative you’re saying you won’t accept that a machine can think under any circumstances, because you’ve given yourselves no way to ever accept it.

That means there’s no reason why anyone should listen to any of you, because you have already made up your minds.

 

[inocente]

An aside: Is it? Certainly we could use apples in a barrel as a rudimentary calculator (like children count on their fingers). But how do you know you are not just adding billions of atoms to billions of atoms? How do you know that you are not multiplying the quantity of apples in the first bucket by 2.4? There is a natural interpretation, of course, but the “translation” into mathematical language is indeterminate. But the same difficulties in principle face a computer. Its outputs and (name removed by moderator)uts have meaning because I attribute meaning to them and develop a computer to accomplish certain ends of my own. But the computer isn’t objectively telling me anything - its operations are always indeterminate.

An example: for the computer, is 0 + 1 = 1 addition (“0 plus 1 equals 1”), or formal disjunction (“p v q is true where p is false and p is true”)? The answer is neither: it depends on what a human is using it for.

I’ve never seen the worth of this logical positivist navel gazing. It’s like trying to analyze how you walk and getting confused about how you do it and falling over. Ordinary people have had no problem doing sums for thousands of years. It works, end of story.

*I have no glaring objections to the definition. But the issue is simply that the Turing test can’t be used to verify most of those, because the coincidence of outputs does not need to be accompanied by reasoning, abstract thought, comprehension, “catching on,” “making sense,” or “figuring out.” All of those terms, in any ordinary use of the English language, denote internal processes. As long as we believe that the subject in question has a mind, then it makes sense to suppose that his performance is an indicator of what he’s thinking. But the performance alone does not allow us to conclude that he has a mind (as I said to Jewel, we need additional premises: 1). he is a human, 2). I am a human, 3).I know that I think).

(Note, one could soften the definitions for reasoning, abstract thought, comprehension, mind, etc., but then the conclusion would be trivial.)*

But you’re doing what Turing is attempting to avoid, getting all hung up about definitions of words. I’ll say the same to you as to MPat: until you propose a workable alternative to Turing’s test, no one will listen to you because you give yourself no way to ever accept the a machine can think, you’ve already made up your mind.

So what is your alternative to Turing’s game, what would make you accept that God is omnipotent and can let a machine think?

Hm? No one said everyone experiences qualia in the same way. I could have an inverted spectrum compared to you (and no one would know it) but it still wouldn’t make qualia less coherent.

Not to mention: some people are colorblind. No one is denying that colorblindness has a physical basis depending on what information reaches the brain. The question of how what does reach the brain is experienced as a quale remains intractable because, of course, the data was irrelevant to it.

You have no means to keep alive this belief in qualia though, since you can never make progress when you’ve already decided it’s intractable. Science will gradually chip away at the solution, and as with other mystical beliefs through history, the belief in qualia will melt away, it will be an ex-parrot.

As more and more is explained, the next generation have fewer and fewer reasons for beliefs that are inexplicable.

I agree they’re not “mystical.” I’ve said that I think qualia and consciousness have a “material” basis, just that the conceptual tools required to explain them are unavailable to philosophical naturalism, based on the way it has defined itself historically.

You are definitely a cup-half-empty kind of guy. But no, it’s do-able, just needs a bit of time. It may need a few generations, kids are much better at shaking off shackles and innovating.

 

[Peter_Plato]

You have no means to keep alive this belief in qualia though, since you can never make progress when you’ve already decided it’s intractable. Science will gradually chip away at the solution, and as with other mystical beliefs through history, the belief in qualia will melt away, it will be an ex-parrot.

He said as he hacked away at the branch upon which he sat.

That is the problem with “going out on a limb,” all the while assuming that the limb will be sufficient to hold a great deal of weight when it cannot even withstand the cutting edge of scrutiny.

The point being that eliminative materialism cannot explain much that is self-evidently true and by merely dismissing that, it ventures blindly out “on a limb,” all the while disregarding the clear signs that its “going forth” is leading to ever smaller and less solid ground.

I saw a cat once being enticed by a crow to venture onto ever smaller and higher branches up in a birch tree. I can’t figure out if you are playing the role of the crow or the cat, but I do know what happened to the cat when it went too far out on that small, high branch.

 

[inocente]

Well, the main issue with your point #3 is that it’s blatantly question begging, “If we knew the architecture of the brain, we would know how thoughts progress.” That’s what’s being debated, so of course it can’t be used as an argument for the conclusion.

What I mean is currently we have no language for describing thoughts. We need to find patterns in brain activity and layout, then we can start to analyze thoughts.

The other thing is that my argumentative strategy has been not at all as you’ve characterized it. I’ve “only said there’s something mystic in the air which will deny us a full explanation”? I have given arguments about why it seems to be in principle impossible to submit the mind to an explanation under methodological naturalism. The issue isn’t that we don’t know some stuff about the mind; it’s that explaining qualia or intentionality by way of a system that stipulates that subjective qualities are not empirically observable is nothing short of a category error.

I’m going to stick with what I said. The issue is that the hypotheses of qualia and intentionality disallow explanation on the basis of evidence and thus are not scientific, yet there is an insistence that they are kept. They are defined such that they can’t be analyzed empirically, yet can’t be analyzed by logic alone either. Who needs them, it’s as if they’ve been specifically designed to forever remain mystic.

But again, this just begs the question as to whether there is a “language of thought” (which seems dubious even if we grant materialism).

That’s a sweeping statement. You appear to be saying that thought cannot be analyzed, even in principle. But there’s linguistics, moral grammar and so on. Why shouldn’t thoughts be capable of analysis?

*Not at all.
1). I’m not committed to saying that we have an ability to identify whether another species is intelligent.
2). I have given reasons why Turing’s fails to provide evidence that another species is intelligent.
3). I have given other reasons why it is doubtful that a machine could be intelligent.

I haven’t claimed that “God cannot ever, under any circumstances, make a machine think.” I’m pointing out that the machine’s ability to function in human conversation is not evidence of its thinking, and that the material facts don’t give us reason to believe that it thinks (and in fact give us reason not to believe that it does).

*

But to me this is just saying, in various different ways, that you have no alternative to Turing’s test, and so give yourself no means to ever accept that a machine can think.

What is confusing? Some materialists reject qualia because they are apparently unverifiable, subjective, and confused notions. Other materialists (Paul and Patricia Churchland, Alex Rosenberg) reject thoughts because they are apparently unverifiable, subjective, and confused notions. Such is the materialist project of denying the most basic aspects of human existence.

I’ve not read these people and can’t comment on what they may be saying. But for instance, Churchland’s home page implicitly acknowledges thinking: “I explore the impact of scientific developments on our understanding of consciousness, the self, free will, decision making, ethics, learning, and religion and issues concerning the neurobiological basis of consciousness, the self, and free will, as well as on more technical questions concerning to what degree the nervous system is hierarchically organized, how the difficult issue of co-ordination and timing is managed by nervous systems, and what are the mechanisms for the perceptual phenomenon of filling-in.”

*This is frankly uncharitable and immature. Your position is completely insulated from criticism. First, in #151, you assume that every philosopher who believes in qualia is acting out of some religious prejudice and scientific ignorance. Given the names of a couple credible and well-renowned philosophers who are atheists (I also forgot Tallis, who is a former neuroscientist, from the list), you decide to liken them to superstitious “ID fans” and astrologists.

These are naturalists and atheists, some of whom are fairly well-versed in neuroscience. Unfortunately, I don’t think the ad hominem attacks and likening to the intelligent design bogeyman suffices to sweep them under the rug.*

Hang on, hang on. I was responding to you saying that science “is intrinsically limited to resolving questions of mind, and some overhaul is needed if it is to succeed”. That is precisely the argument made by ID, that science must allow theories not based on evidence - newscientist.com/article/dn8178-astrology-is-scientific-theory-courtroom-told.html

To me it’s the same, I can’t see the difference. I mean I agree with your characterization of ID, but when anyone says science must change to let in their theory, it doesn’t matter what their theory is, they are still attacking the very foundation on which science is built.

Well, fortunately we do have evidence for qualia, ie. the fact that we are not (or, at least, I am not) philosophical zombies. Hard to believe that knowing that we aren’t philosophical zombies is “useless baggage.”

Hmmm. Define things called qualia. Call someone without these things a p-zombie. Say oh look there are no p-zombies. Therefore there must be qualia! Can’t believe you would make such a circular argument.

 

[polytropos]

I’ve never seen the worth of this logical positivist navel gazing. It’s like trying to analyze how you walk and getting confused about how you do it and falling over. Ordinary people have had no problem doing sums for thousands of years. It works, end of story.

Noting that 0 + 1 could represent addition or formal disjunction isn’t logical positivism. It’s just a fact about computing.

I didn’t say there’s a problem with ordinary people doing sums either. I’m saying when we program a computer to do sums for us, the meaning of its symbols and manipulation is derived from our intentions, not from the syntax.

But you’re doing what Turing is attempting to avoid, getting all hung up about definitions of words.

It’s rather tough for me to respond to both you and Jewel, since you’re saying I’m getting hung up on definitions, and Jewel is saying that I need to provide more definitions (even though Turing won’t). I’ll take this as evidence that I’m striking a nice middle ground.

That said, noting that reasoning, abstract thought, comprehension, “catching on,” “making sense,” or “figuring out” are all inextricably internal and subjective is hardly “getting all hung up about definitions of words.” If we ignore all of their definitions, then we’ve excluded any interpretation of the Turing test’s conclusions. As I’ve mentioned before, I’m fine with that; I admit that the Turing test could give us weak AI (ie. a machine that passes the Turing test behaves “like a human” and can be defined as “artificially intelligent.”) But if you are ignoring the definitions of reasoning, abstract thought, comprehension, etc. it’s hard to say what exactly you’ve proven at the end of it all. (And one has to watch out for equivocation, if one defines terms after the demonstration.)

As I’ve said to Jewel, I think philosophy (particularly critical philosophy) can get off of the ground without defining all of its terms.

Not to mention… you provided the definition of intelligence from Wikipedia. I observed: okay, I have no major problems with that, but the Turing test does not meet the definition that you linked to.

I’ll say the same to you as to MPat: until you propose a workable alternative to Turing’s test, no one will listen to you because you give yourself no way to ever accept the a machine can think, you’ve already made up your mind.

So what is your alternative to Turing’s game, what would make you accept that God is omnipotent and can let a machine think?

Sorry, but this does not make sense. The weakness of the Turing test does not oblige its critics to find an alternative, because the critics’ point is that Turing test simply doesn’t accomplish what it sets out to do (unless it is taken to show weak AI, which it does). If one method fails, it’s not “right” until someone comes up with an alternative; if I wrote out a completely fallacious proof of Euler’s Riemann zeta conjecture, there would be no requirement for anyone to accept it just because they can’t produce a valid proof. (More than a few good mathematicians have burnt out trying to solve it.)

The bolded part seems the crux of the issue. You are assuming that there is a way to find out whether a machine thinks. I don’t see why that is an epistemic truth; unless someone has twisted their definition substantially, thinking is a subjective process. It seems to be a category error to declare that there is an objective test to detect it.

Another analogy: I claim that the existence of God can be demonstrated objectively. You claim that something like Aquinas’s First Way fails. I obviously can’t say that until you produce an alternative, better demonstration, you better accept Aquinas’s First Way, even if you think it’s faulty. The only reason one should accept any demonstration is if it shows what it’s meant to show. Turing’s test does not, so even in the absence of an alternative, we have no reason to accept it.

You have no means to keep alive this belief in qualia though, since you can never make progress when you’ve already decided it’s intractable.

I think the subjective phenomena are intractable to systems that are defined by the stipulation that subjective phenomena are not empirically observable.

 

[polytropos]

The issue is that the hypotheses of qualia and intentionality disallow explanation on the basis of evidence and thus are not scientific, yet there is an insistence that they are kept. They are defined such that they can’t be analyzed empirically, yet can’t be analyzed by logic alone either. Who needs them, it’s as if they’ve been specifically designed to forever remain mystic.

If you think that the problem with their definitions is that they are “defined such that they can’t be analyzed empirically,” then feel free to define the phenomena they describe in such a way that they are receptive to scientific explanation. Neuroscientists haven’t done so. Materialists generally opt for elimination rather than explanation. So that leaves eliminativism… or one could admit that science might just not explain everything. Or one could leave a vague blank check, “Believe me, science will explain it. Someday…”

That’s a sweeping statement. You appear to be saying that thought cannot be analyzed, even in principle. But there’s linguistics, moral grammar and so on. Why shouldn’t thoughts be capable of analysis?

Well, some philosophers have proposed a subconscious “mental language” (sometimes called “mentalese”), but it is widely regarded as implausible. I was not sure if that was what you were referring to. There is certainly data in our brains that corresponds to certain occurrences. I can also internally “vocalize” things to myself. But it seems very plausible to say that propositional knowledge is not necessarily linguistic. I’m not sure if any of that is what you were referring to.

I’ve not read these people and can’t comment on what they may be saying. But for instance, Churchland’s home page implicitly acknowledges thinking: …

From Wikipedia: “Along with his wife, Churchland is a major proponent of eliminative materialism, the belief which claims that everyday mental concepts such as beliefs, feelings, and desires are part of a “folk psychology” of theoretical constructs without coherent definition, destined to simply be obviated by a thoroughly scientific understanding of human nature.” It is hard to see why your reasons for eliminating qualia and intentionality would not carry over to beliefs and desires (which are intentional phenomena, anyway). (But more on that below.)

Hang on, hang on. I was responding to you saying that science “is intrinsically limited to resolving questions of mind, and some overhaul is needed if it is to succeed”. That is precisely the argument made by ID, that science must allow theories not based on evidence - newscientist.com/article/dn8178-astrology-is-scientific-theory-courtroom-told.html

I said, “generally they think that materialism is simply incomplete and that neuroscience would have to be overhauled to explain qualia.” (That seems to be what you’re referring to, unless you responded to something that you did not quote.) Materialism and science aren’t the same thing.

To me it’s the same, I can’t see the difference. I mean I agree with your characterization of ID, but when anyone says science must change to let in their theory, it doesn’t matter what their theory is, they are still attacking the very foundation on which science is built.

They are hardly “attacking.” They are noting that science, if it is to explain phenomena of a sort which it has previously defined as unexplainable, would have to change. They aren’t suggesting that any scientific theories are wrong. They are claiming that materialistic interpretations of scientific theories are wrong.

I also don’t think I offered a “characterization of ID.” I just stated what the implications of your references seemed to be. “Intelligent design” is a pretty broad term.

Hmmm. Define things called qualia. Call someone without these things a p-zombie. Say oh look there are no p-zombies. Therefore there must be qualia! Can’t believe you would make such a circular argument.

I’m not giving a formal argument when I say, “Well, fortunately we do have evidence for qualia, ie. the fact that we are not (or, at least, I am not) philosophical zombies.” That is literally just a restatement of the definition of qualia. That which is not a philosophical zombie has qualia. I am just articulating it in that way to drive home how fundamental qualia are to our existence, and what a denial of them is tantamount to.

To revisit a previous point, an eliminativist could say, “Define things called thoughts. Call someone without these things a p-zombie. Say oh look there are no p-zombies. Therefore there must be thoughts!” Since the reference to philosophical zombies is intended as a restatement of the definition, the circularity is not a concern (for circularity is not an indicator of inconsistency, and is to be expected when someone is emphasizing an idea). But the statement gives us no reason to doubt the existence of thoughts.

 

[MPat]

Yes of course, it’s standard practice.

You keep coming back to this but it’s possible to cheat in any test or exam yet that doesn’t stop anyone. We’ve already been round this anyway, you isolate the machine and run scrupulous protocols to prevent cheating, it’s not rocket science, it’s standard experimental design.

I do not mean “cheating” as in, let’s say, bribing the judges. The “cheating” in question is, essentially, making the computer pass the test. The computer is being made to “look” like a human. In a sense, that is not “cheating” - this is just the intended way to pass the test.

That is, in a sense, you get the computer that is actually mistaken for a human and now you are claiming that it is not enough…? Well, the Turing test is simply to get the computer mistaken for a human, nothing more. There is no rule that passing by guessing the questions and preparing for them doesn’t count.

And your claim is that passing the Turing test is sufficient. If there are any other conditions, your claim will have to be abandoned or modified.

Yes of course. But probably the best environment would be a messaging system like CAF, where both the human and the machine can spend as long as they want to post their answers, then it’s not an issue. All this stuff is just a matter of protocol, scientists design experiments everyday, there’s nothing new here even though you may never have thought about it before.

However, it was the “experiment” designed by Turing that is being discussed. Design it differently and you will get “Turing-inocente test” instead.

Turing invented a game by which it can be demonstrated that a machine does or does not think.

No, it is not certain that it can actually demonstrate that. It has been claimed that it can, it has been claimed that it cannot. We are discussing that.

No one on this thread has proposed an alternative, you guys have just picked holes. But by only picking holes and not proposing any alternative you’re saying you won’t accept that a machine can think under any circumstances, because you’ve given yourselves no way to ever accept it.

That means there’s no reason why anyone should listen to any of you, because you have already made up your minds.

Well, why are you talking to the ones who should not be listened to…?

Anyway, I did propose an alternative: get us the private revelation declared worthy of belief by the Church, claiming that the specific computer is able to think. Doesn’t that count as “test”…?

Also, you have yet to give us a proper explanation of qualia. You only indicated that you have faith (with little to support it) that “science” will explain everything. Eventually. Well, in such case, does it mean we shouldn’t listen to you either…? Or does it mean that it is OK to have faith that we will also get a way to find out if a given object can think - eventually…? For example, after the death or the end of the world…?

 

[Peter_Plato]

Yes of course, it’s standard practice.

You keep coming back to this but it’s possible to cheat in any test or exam yet that doesn’t stop anyone. We’ve already been round this anyway, you isolate the machine and run scrupulous protocols to prevent cheating, it’s not rocket science, it’s standard experimental design.

The reason this “test” for thinking doesn’t work is precisely because the “thinker” in question has not only been supplied with all the responses, but a listing of all the challenges and a complete set instructions for addressing the different kinds of challenges by the programmer prior to the test. It is a faulty test for “thinking” in the same way that providing college students with a comprehensive set of instructions and answers for taking a test circumvents the need for intelligent thinking and therefore cannot be used to measure the quality of that thinking - i.e., if prompt Q1, reply A1; if Q2, reply A2; etc.

Since the Turing Test is constructed precisely to circumvent the need for intentional thinking, by preloading questions, instructions and responses, it cannot be legitimately used to determine whether intentional thinking is actually present. It is clearly a test of the programmer’s capability to prepare or setup the experimental “subject” rather than as a real test of the autonomous ability of that subject. Since the burden of the test is completely on the programmer, none remains for the theoretical subject.

In short, the reason it is a faulty experiment, is because we would have huge issues with using the same method for testing the quality of intentional thinking for subjects we know are capable of intentional thinking, college students, for example. Why would we use such a faulty means on subjects where the presence of subjective thinking, itself, is in question?

 

[inocente]

Noting that 0 + 1 could represent addition or formal disjunction isn’t logical positivism. It’s just a fact about computing.

I didn’t say there’s a problem with ordinary people doing sums either. I’m saying when we program a computer to do sums for us, the meaning of its symbols and manipulation is derived from our intentions, not from the syntax.

I meant your “billions of atoms to billions of atoms” question and so on.

I can’t see the supposed problem about meaning. There is a wide gulf between a computer and something with intelligence. I don’t see how anything with intelligence could only see the world as sets of symbols.

*It’s rather tough for me to respond to both you and Jewel, since you’re saying I’m getting hung up on definitions, and Jewel is saying that I need to provide more definitions (even though Turing won’t). I’ll take this as evidence that I’m striking a nice middle ground.

That said, noting that reasoning, abstract thought, comprehension, “catching on,” “making sense,” or “figuring out” are all inextricably internal and subjective is hardly “getting all hung up about definitions of words.” If we ignore all of their definitions, then we’ve excluded any interpretation of the Turing test’s conclusions. As I’ve mentioned before, I’m* fine with that; I admit that the Turing test could give us weak AI (ie. a machine that passes the Turing test behaves “like a human” and can be defined as “artificially intelligent.”) But if you are ignoring the definitions of reasoning, abstract thought, comprehension, etc. it’s hard to say what exactly you’ve proven at the end of it all. (And one has to watch out for equivocation, if one defines terms after the demonstration.)

As I’ve said to Jewel, I think philosophy (particularly critical philosophy) can get off of the ground without defining all of its terms.

Not to mention… you provided the definition of intelligence from Wikipedia. I observed: okay, I have no major problems with that, but the Turing test does not meet the definition that you linked to.

I think we’re talking at cross purposes. Turing is interested in able to decide whether a machine can think. But he wants to avoid discussion of what is a machine, since some would say that a human is a biological machine, and the discussion will disappear into metaphysical clouds and never return. He wants to avoid trying to define what is meant by thinking for the same reason. He is aware that our concepts of machine, human, thinking, intelligence are squidgy. So in the absence of universally agreed hard definitions, what will convince most people that something made by engineers can think (and perhaps can suffer)?

Sorry, but this does not make sense. The weakness of the Turing test does not oblige its critics to find an alternative, because the critics’ point is that Turing test simply doesn’t accomplish what it sets out to do (unless it is taken to show weak AI, which it does). If one method fails, it’s not “right” until someone comes up with an alternative; if I wrote out a completely fallacious proof of Euler’s Riemann zeta conjecture, there would be no requirement for anyone to accept it just because they can’t produce a valid proof. (More than a few good mathematicians have burnt out trying to solve it.)

I asked “So what is your alternative to Turing’s game, what would make you accept that God is omnipotent and can let a machine think?”

So OK, remove the bit about Turing, what would make you accept that God is omnipotent and can let a machine think?"

*The bolded part seems the crux of the issue. You are assuming that there is a way to find out whether a machine thinks. I don’t see why that is an epistemic truth; unless someone has twisted their definition substantially, thinking is a subjective process. It seems to be a category error to declare that there is an objective test to detect it.

Another analogy: I claim that the existence of God can be demonstrated objectively. You claim that something like Aquinas’s First Way fails. I obviously can’t say that until you produce an alternative, better demonstration, you better accept Aquinas’s First Way, even if you think it’s faulty. The only reason one should accept any demonstration is if it shows what it’s meant to show. Turing’s test does not, so even in the absence of an alternative, we have no reason to accept it.*

This is just taking us further into the Slough of Definition and the Mire of Meaning. But there’s a job to be done, and it won’t get done if we say everything is subjective and so nothing can be done.

So let’s bring in the Marines. Are there are things we can do which mud cannot? “Yes, Sir.” Do these include curiosity, creativity, self awareness, empathy, and so on? “Yes, Sir.” Are these objective qualities, would you know them when you see them in others? “Yes, Sir.” Do you think their absence demonstrates absence of ability to think? “Yes, Sir.”

Job done.

I think the subjective phenomena are intractable to systems that are defined by the stipulation that subjective phenomena are not empirically observable.

As I said, you have no means to keep alive this belief in qualia since you can never make progress when you’ve already decided it’s intractable. Meanwhile us naughty scientist engineer Marine types will make progress until no one is left to wonder what these subjective phenomena were supposed to be. It’s cooties all over again.

 

[inocente]

If you think that the problem with their definitions is that they are “defined such that they can’t be analyzed empirically,” then feel free to define the phenomena they describe in such a way that they are receptive to scientific explanation. Neuroscientists haven’t done so. Materialists generally opt for elimination rather than explanation. So that leaves eliminativism… or one could admit that science might just not explain everything. Or one could leave a vague blank check, “Believe me, science will explain it. Someday…”

The Wikipedia article on qualia lists examples: a headache, the taste of wine, and the perceived redness of an evening sky. Currently we have no idea whether those three things have anything in common. Some philosophers would like us to believe they do, but they have no reason to bunch them together other than to bunch them together. Oh look, they say, a bunch of disparate subjective feelings we can’t explain, let’s bunch them together.

We can safely ignore those philosophers and their notions of cooties.

Well, some philosophers have proposed a subconscious “mental language” (sometimes called “mentalese”), but it is widely regarded as implausible. I was not sure if that was what you were referring to. There is certainly data in our brains that corresponds to certain occurrences. I can also internally “vocalize” things to myself. But it seems very plausible to say that propositional knowledge is not necessarily linguistic. I’m not sure if any of that is what you were referring to.

I’m just saying that you cannot have composed those sentences without in some way first forming them in your mind, and that process is open to analysis (in principle if not yet in practice).

From Wikipedia: “Along with his wife, Churchland is a major proponent of eliminative materialism, the belief which claims that everyday mental concepts such as beliefs, feelings, and desires are part of a “folk psychology” of theoretical constructs without coherent definition, destined to simply be obviated by a thoroughly scientific understanding of human nature.” It is hard to see why your reasons for eliminating qualia and intentionality would not carry over to beliefs and desires (which are intentional phenomena, anyway). (But more on that below.)

You were referring to his wife, not to her husband, and you said they “reject thoughts because they are apparently unverifiable”. The article you quote doesn’t back you up - it says they believe there may be no need for concepts like “beliefs” or “feelings” as they don’t correspond to “objective phenomena, such as activation patterns across neural networks”, whereas of course thinking does correspond to activation patterns.

I agree with them if they’re saying that most of our terminology about the mind is too loose to hang a hat on, but that’s a long way from your claimed “materialist project of denying the most basic aspects of human existence”.

*I said, “generally they think that materialism is simply incomplete and that neuroscience would have to be overhauled to explain qualia.” (That seems to be what you’re referring to, unless you responded to something that you did not quote.) Materialism and science aren’t the same thing.

They are hardly “attacking.” They are noting that science, if it is to explain phenomena of a sort which it has previously defined as unexplainable, would have to change. They aren’t suggesting that any scientific* theories are wrong. They are claiming that materialistic interpretations of scientific theories are wrong.

I also don’t think I offered a “characterization of ID.” I just stated what the implications of your references seemed to be. “Intelligent design” is a pretty broad term.

They are placing the horse before the cart. In both cases they are saying science must change to accommodate their beliefs irrespective of evidence, which would destroy the entire foundation of science.

I’m not giving a formal argument when I say, “Well, fortunately we do have evidence for qualia, ie. the fact that we are not (or, at least, I am not) philosophical zombies.” That is literally just a restatement of the definition of qualia. That which is not a philosophical zombie has qualia. I am just articulating it in that way to drive home how fundamental qualia are to our existence, and what a denial of them is tantamount to.

This still seems to be a prime candidate for Circular Reasoning Of The Year Award 2013. P-zombies usually come up in attempts to disprove physical theories of mind. In reality we internally experience things which may or may not having anything in common which we vaguely call feelings.

To revisit a previous point, an eliminativist could say, “Define things called thoughts. Call someone without these things a p-zombie. Say oh look there are no p-zombies. Therefore there must be thoughts!” Since the reference to philosophical zombies is intended as a restatement of the definition, the circularity is not a concern (for circularity is not an indicator of inconsistency, and is to be expected when someone is emphasizing an idea). But the statement gives us no reason to doubt the existence of thoughts.

I think you’ll find that the p-zombie was invented by dualists in an attempt to attack physicalism, so dualists only have themselves to blame if it backfired on them. I don’t know of anyone who says thoughts don’t exist, they would seem pretty dippy, not to mention flaky and whacky, given we can watch MRI Christmas trees of neurons light up when people think.

 

[inocente]

I do not mean “cheating” as in, let’s say, bribing the judges. The “cheating” in question is, essentially, making the computer pass the test. The computer is being made to “look” like a human. In a sense, that is not “cheating” - this is just the intended way to pass the test.

That is, in a sense, you get the computer that is actually mistaken for a human and now you are claiming that it is not enough…? Well, the Turing test is simply to get the computer mistaken for a human, nothing more. There is no rule that passing by guessing the questions and preparing for them doesn’t count.

And your claim is that passing the Turing test is sufficient. If there are any other conditions, your claim will have to be abandoned or modified.

I got lost. It is straight-forward to design protocols to prevent the makers of a machine or the machine itself deceiving the interrogator.

However, it was the “experiment” designed by Turing that is being discussed. Design it differently and you will get “Turing-inocente test” instead.

Turing doesn’t go into detailed specifics on the design, he leaves it open. Have you actually read the paper?

csee.umbc.edu/courses/471/papers/turing.pdf

No, it is not certain that it can actually demonstrate that. It has been claimed that it can, it has been claimed that it cannot. We are discussing that.

Yes, I understand that you won’t accept anything short of 100% proof. You don’t ask for 100% proof that someone loves you or that the café didn’t poison your coffee, but that’s not as vital as an internet discussion of course.

Well, why are you talking to the ones who should not be listened to…?

Yes, that confirms my thoughts.

Anyway, I did propose an alternative: get us the private revelation declared worthy of belief by the Church, claiming that the specific computer is able to think. Doesn’t that count as “test”…?

Sorry, I can’t even begin to make sense of this or what came after.

 

[inocente]

The reason this “test” for thinking doesn’t work is precisely because the “thinker” in question has not only been supplied with all the responses, but a listing of all the challenges and a complete set instructions for addressing the different kinds of challenges by the programmer prior to the test. It is a faulty test for “thinking” in the same way that providing college students with a comprehensive set of instructions and answers for taking a test circumvents the need for intelligent thinking and therefore cannot be used to measure the quality of that thinking - i.e., if prompt Q1, reply A1; if Q2, reply A2; etc.

Since the Turing Test is constructed precisely to circumvent the need for intentional thinking, by preloading questions, instructions and responses, it cannot be legitimately used to determine whether intentional thinking is actually present. It is clearly a test of the programmer’s capability to prepare or setup the experimental “subject” rather than as a real test of the autonomous ability of that subject. Since the burden of the test is completely on the programmer, none remains for the theoretical subject.

In short, the reason it is a faulty experiment, is because we would have huge issues with using the same method for testing the quality of intentional thinking for subjects we know are capable of intentional thinking, college students, for example. Why would we use such a faulty means on subjects where the presence of subjective thinking, itself, is in question?

What?

Back on planet earth, there is Turing’s actual paper where no scripts are mentioned.

Why don’t you read it, then we could discuss reality.

csee.umbc.edu/courses/471/papers/turing.pdf

 

[MPat]

I got lost. It is straight-forward to design protocols to prevent the makers of a machine or the machine itself deceiving the interrogator.

You know, the whole point of the test is to try to deceive the interrogator…

If it is too hard to understand, let’s look at those cases of passing the Turing test:

1. The creator of machine bribes the interrogator.
2. The creator of machine steals the questions and builds the machine that can answer them to pass the test.
3. The creator of machine guesses the questions and builds the machine (just like the one in case 2) that can answer them to pass the test.
4. The creator of machine simply makes the machine that just happens to be like the ones in case 2 and case 3.
5. The creator of machine makes a machine that simulates the human brain (or something) and it passes the test.

Case 1 includes clear cheating, case 2 also includes clear cheating, cases 4 and 5 are clearly “honest”. Yet the machines in cases 2 to 4 are identical. And the machines in cases 2 to 5 all pass the same Turing test. Now, what do you think: which of the cases describe the machines that can be concluded to think?

At this point, given your previous statements, I expect you to answer “In cases 4 and 5.” or “In case 5.”…

Turing doesn’t go into detailed specifics on the design, he leaves it open. Have you actually read the paper?

csee.umbc.edu/courses/471/papers/turing.pdf

That’s because if you need non-trivial “specifics of design”, it is a sign that this test doesn’t work that well.

And yes, I did read much of the paper (and I remember reading it at some other time).

Yes, I understand that you won’t accept anything short of 100% proof. You don’t ask for 100% proof that someone loves you or that the café didn’t poison your coffee, but that’s not as vital as an internet discussion of course.

Well, the whole question is “Does it give us 100% proof?”. Otherwise we get just one more heuristic that can give us a right or wrong answer, without any idea how likely each of them is. And we already have many heuristics to choose from, for example, “Does it have human DNA?”, “Does it look like a human from afar?” or even a coin flipping. What makes you think that the Turing test is significantly better than those…?

Sorry, I can’t even begin to make sense of this or what came after.

OK, what exactly is not clear about that? You are challenging us to show that, under some circumstances, we can accept God showing His omnipotence in giving a machine ability to think. I am offering you such circumstances. That is, if God is willing to give a machine such an ability, He can also give a vision to someone and say so - that would be a good time to consider the possibility we’re talking about. That is not unimaginable: He did send a vision to St. Peter that has informed him that Gentiles can be baptised right away.

Now, do you have any objections to such a “test”? Are you going to say that it is not possible? I doubt that. Are you going to say that it is not scientific? Well, you didn’t ask for a “scientific” test (and asking for that would not work that well with your challenge). Are you going to say that it is unlikely that “my” test will be passed any time soon? Sure. But I also doubt if the Turing test will be passed (consistently) very soon either…

Is that clearer? Or is some part unclear? If so, which one?

 

[MPat]

I can’t see the supposed problem about meaning. There is a wide gulf between a computer and something with intelligence. I don’t see how anything with intelligence could only see the world as sets of symbols.

Er, aren’t you supposed to claim that there is no “wide gulf between a computer and something with intelligence”…?

I think we’re talking at cross purposes. Turing is interested in able to decide whether a machine can think. But he wants to avoid discussion of what is a machine, since some would say that a human is a biological machine, and the discussion will disappear into metaphysical clouds and never return.

Where do you take that from? In the paper itself, Turing dedicates the whole 3rd section to this question - and answers that he is talking about “digital computers”! He also specifically tells us “we wish to exclude from the machines men born in the usual manner.”. Later, in sections 4 and 5, he tells us more about the machines he has in mind. It is certainly wrong to say that “he wants to avoid discussion of what is a machine”.

He wants to avoid trying to define what is meant by thinking for the same reason.

Well, I suppose we could say that he evaded the straight answer with his “joke” about “Gallup poll” and that this answer is “close enough”… Still, I’d say that he simply didn’t know any definition that would have been useful to him…

He is aware that our concepts of machine, human, thinking, intelligence are squidgy. So in the absence of universally agreed hard definitions, what will convince most people that something made by engineers can think (and perhaps can suffer)?

A “Gallup poll”…? Sorry, but Turing seems to be rejecting this standard in the very first paragraph…

I asked “So what is your alternative to Turing’s game, what would make you accept that God is omnipotent and can let a machine think?”

So OK, remove the bit about Turing, what would make you accept that God is omnipotent and can let a machine think?"

I think I have answered that already…

This is just taking us further into the Slough of Definition and the Mire of Meaning. But there’s a job to be done, and it won’t get done if we say everything is subjective and so nothing can be done.

So let’s bring in the Marines. Are there are things we can do which mud cannot? “Yes, Sir.” Do these include curiosity, creativity, self awareness, empathy, and so on? “Yes, Sir.” Are these objective qualities, would you know them when you see them in others? “Yes, Sir.” Do you think their absence demonstrates absence of ability to think? “Yes, Sir.”

Er, what makes you think that, let’s say, absence of empathy “demonstrates absence of ability to think”…?

Also, you seem to be arguing like this: (not A => not B) => (A => B). That is: no empathy etc. implies no thinking, therefore, empathy implies thinking. But logic doesn’t work that way…

Job done.

As I said, you have no means to keep alive this belief in qualia since you can never make progress when you’ve already decided it’s intractable. Meanwhile us naughty scientist engineer Marine types will make progress until no one is left to wonder what these subjective phenomena were supposed to be. It’s cooties all over again.

In other words, you have a strong faith in science. Good for you. But maybe you should leave “cooties” out of this…? Do they actually add anything to your conclusion or argument…?

 

[Peter_Plato]

What?

Back on planet earth, there is Turing’s actual paper where no scripts are mentioned.

Why don’t you read it, then we could discuss reality.

csee.umbc.edu/courses/471/papers/turing.pdf

Where will the functional capacities of these artificially intelligent machines come from if not from programmed scripts? That would seem the presumption behind the entire paper, whether Turing mentions it or not. You aren’t seriously proposing that the machines, provided a certain level of complexity of circuitry will no longer require programming. Sophisticated hardware will no longer require software? Seriously?

If that is what you intend, that neither solves the problem because the software coding would then be hard programmed into the design of the components, still requiring anticipatory scripting, though pushed back a level. The machine would still be “responding” to cues, not “thinking” in any meaningful sense.

 

[inocente]

You know, the whole point of the test is to try to deceive the interrogator…

If it is too hard to understand, let’s look at those cases of passing the Turing test:

1. The creator of machine bribes the interrogator.
2. The creator of machine steals the questions and builds the machine that can answer them to pass the test.
3. The creator of machine guesses the questions and builds the machine (just like the one in case 2) that can answer them to pass the test.
4. The creator of machine simply makes the machine that just happens to be like the ones in case 2 and case 3.
5. The creator of machine makes a machine that simulates the human brain (or something) and it passes the test.

Case 1 includes clear cheating, case 2 also includes clear cheating, cases 4 and 5 are clearly “honest”. Yet the machines in cases 2 to 4 are identical. And the machines in cases 2 to 5 all pass the same Turing test. Now, what do you think: which of the cases describe the machines that can be concluded to think?

At this point, given your previous statements, I expect you to answer “In cases 4 and 5.” or “In case 5.”…

You lost me. If and when a machine plays Turing’s game and wins, then the results will be peer reviewed, published, and every critic in the world will be on the case. The experiment will be repeated many times by different teams trying to disprove the results. Religions which don’t believe God would let a machine think will buy some of the machines and test them for themselves. Some of the machines will be put on the internet and every man and his dog will try to catch them out.

That’s because if you need non-trivial “specifics of design”, it is a sign that this test doesn’t work that well.

Well, the whole question is “Does it give us 100% proof?”. Otherwise we get just one more heuristic that can give us a right or wrong answer, without any idea how likely each of them is. And we already have many heuristics to choose from, for example, “Does it have human DNA?”, “Does it look like a human from afar?” or even a coin flipping. What makes you think that the Turing test is significantly better than those…?

You lost me again. You don’t ask for 100% proof that someone loves you or that the café didn’t poison your coffee, or in virtually anything in life, so please explain why you need 100% proof in this one case.

*OK, what exactly is not clear about that? You are challenging us to show that, under some circumstances, we can accept God showing His omnipotence in giving a machine ability to think. I am offering you such circumstances. That is, if God is willing to give a machine such an ability, He can also give a vision to someone and say so - that would be a good time to consider the possibility we’re talking about. That is not unimaginable: He did send a vision to St. Peter that has informed him that Gentiles can be baptised right away.

Now, do you have any objections to such a “test”? Are you going to say that it is not possible? I doubt that. Are you going to say that it is not scientific? Well, you didn’t ask for a “scientific” test (and asking for that would not work that well with your challenge). Are you going to say that it is unlikely that “my” test will be passed any time soon? Sure. But I also doubt if the Turing test will be passed (consistently) very soon either…

Is that clearer? Or is some part unclear? If so, which one?*

You lost me again. Why do you think it’s called private revelation? If someone won’t accept Turing’s test, why would they accept what someone else says God told them? Suppose God reveals it to a Hindu, why would a Catholic believe her? Why would God follow your orders and tell anyone anyway? Are you proposing we close all universities and just fill up books with all the things people say God told them?

Could we try to get back to that little blue-green planet called Earth?

 

[inocente]

Where will the functional capacities of these artificially intelligent machines come from if not from programmed scripts? That would seem the presumption behind the entire paper, whether Turing mentions it or not. You aren’t seriously proposing that the machines, provided a certain level of complexity of circuitry will no longer require programming. Sophisticated hardware will no longer require software? Seriously?

If that is what you intend, that neither solves the problem because the software coding would then be hard programmed into the design of the components, still requiring anticipatory scripting, though pushed back a level. The machine would still be “responding” to cues, not “thinking” in any meaningful sense.

We don’t know enough about intelligence to program it. In any case it would take too long and have too many bugs.

But we can write programs to allow a machine to learn and evolve capabilities. The technology is now cheap enough for anyone to play, and there are basic research programs along with the military and commercial applications, so this field is developing.

There are probably lots of videos and websites but here’s a few to give a flavor.

Intro to neural networks - youtube.com/watch?v=DG5-UyRBQD4
Evolving a robot soccer team - youtube.com/watch?v=cP035M_w82s
Building “self-aware” robots - ted.com/talks/hod_lipson_builds_self_aware_robots.html