Category: Consciousness and Qualia

Semantic Engines, Consciousness, Intentionality, and Understanding

John Haugeland:

The basic idea of cognitive science is that intelligent beings are semantic engines – in other words, automatic formal systems with interpretations under which they consistently make sense. [This puts]  psychology and artificial intelligence on a more or less equal footing: people and intelligent computers [are] different manifestations of the same underlying phenomenon. (p. 15)

“The basic idea of cognitive science” is what John Searle calls Strong Artificial Intelligence.  “On the Strong AI view, the appropriately programmed digital computer does not just simulate having a mind, it literally has a mind.” (Mind: A Brief Introduction, Oxford UP 2004, p. 66) Searle does not, however, endorse Haugeland’s view. The latter — call it H — implies that there is no good reason to think that the  (natural) intelligence in us  must be any different  in kind from the (artificial) intelligence that is now present in some computers.   My view is that there is good reason to doubt it, and what’s more, deny it.

But if I have reason to deny (H), then it is incumbent upon me to explain why. No bare assertions allowed. (Quod gratis asseritur gratis negatur.) I will defend what Hauegeland calls the Hollow Shell Strategy.  The following quotations are from an  accurate summary of Haugeland’s above-cited article, a summary  written by someone else.

Hollow Shell: however well a semantic engine acts as if it understands, it can’t really do so because it’s missing some vital ingredient “X”. A robot based on a semantic engine would be a sham no matter how good it got. Candidates for “X” are consciousness, original intentionality and caring.

I will consider in this post just consciousness and original intentionality.

X = Consciousness: however smart computers might seem, they can’t really understand anything because they aren’t conscious. It’s natural to suppose something difficult and important has been left out. Haugeland responds by saying that no one understands consciousness – it’s mysterious regardless of your viewpoint. So, how do we know (a) that understanding is impossible without consciousness or (b) that semantic engines won’t ever have consciousness?

NON-INTENTIONAL CONSCIOUSNESS

The first step  is to distinguish between consciousness (non-intentional consciousness) and consciousness-of. The latter is what philosophers call intentionality or object-directedness.  A conscious state need not be an intentional or object-directed state.  Felt sensations of pain and pleasure, for example, needn’t present an object: they needn’t be of or about anything in the way my noticing that the door is ajar about the door’s being ajar. I have found that it is very difficult to get this distinction across to some people. I’ll try once more.

You are pedaling your bike and you feel  something you describe as ‘pain in the right knee.’ Is that felt pain literally in your knee? No.  What is literally inside the knee is at most the cause of the felt pain:  an X-ray reveals that you are ‘bone-on-bone’: the meniscus has worn away and the femur is grinding against the tibia.  The X-ray reveals the cause. The cause is physical; the pain you feel is not physical, but mental. Follow me so far?

The felt pain — the pain as sensory quale — does not show up on the X-ray slide, nor will the felt pain be empirically detected by an orthopedic surgeon who cuts into the knee and looks around in there.  The point is this: the felt pain is a conscious state that is not an intentional state.  The conscious state has no intentional object;  it is not object-directed; it does not present an object to the person in the state; it does not ‘take an accusative.’ These are four different ways of saying the same thing.

So far I have made two simple distinctions. The first is between the physical/material cause of the felt pain and the felt pain as mental /conscious sensory quale. The second is between conscious states that are not object-directed and those that are.

What is a conscious state? A conscious state is an occurrent episode of awareness. I have just alluded to a third distinction between the occurrent and the dispositional.  Another technical term is ‘act,’ not to be confused with ‘action.’ (A fourth distinction.) An act is an occurrent episode of awareness (consciousness) that takes an object. No act without an object and no object without an act.  ‘Object’ in this context means intentional object, where such an object may or may not exist without prejudice to its being an intentional object.  So we distinguish between merely intentional objects and intentional objects that also exist in reality, i.e., extra-mentally. (I’ve just made a fifth distinction).  If I ever get a tattoo, it will read Distinguo ergo sum.

There are other points to be made but I must move on.

If we use ‘qualia’ to refer to the non-intentional conscious states one of which I  have just  described, then Daniel Dennett is a qualia-denier. His opposite number is Thomas Nagel, a qualia-affirmer. To put it bluntly and polemically: Thomas Nagel is for me the real thing as philosophers go; Daniel Dennett is a brilliant sophist. But he is indeed brilliant and creative and his work cannot be ignored. His work functions as a very useful and richly elaborated foil against which to develop saner and sounder theories, theories that have a chance of being true.

Killer Quotation from Nagel from a  NYRB review.

I am reminded of the Marx Brothers line: “Who are you going to believe, me or your lying eyes?” Dennett asks us to turn our backs on what is glaringly obvious—that in consciousness we are immediately aware of real subjective experiences of color, flavor, sound, touch, etc. that cannot be fully described in neural terms even though they have a neural cause (or perhaps have neural as well as experiential aspects). And he asks us to do this because the reality of such phenomena is incompatible with the scientific materialism that in his view sets the outer bounds of reality. He is, in Aristotle’s words, “maintaining a thesis at all costs.”

That’s right. When a line of reasoning issues in an absurdity such as the absurdity that consciousness and its deliverances are illusions, then what you have is a reductio ad absurdum of one or more of the premises with which the reasoning began.  Dennett assumes physicalism and that everything can be explained in physical terms.  This leads to absurdity. But Dennett, blinded by his own brilliance — don’t forget, he counts himself one of the ‘brights’ — bites the bullet. He’d rather break his teeth than examine his assumptions.

Another thing struck me. Dennett makes much of Wilfrid Sellars’ distinction between the manifest and scientific images. ‘Image’ is not quite the right word. An image is someone’s image. But whose image is the scientific image? Who is its subject? It is arguably our image no less than the manifest image.  Nagel quotes Dennett as saying of the manifest image: “It’s the world according to us.”  But the same, or something very similar, is true of the scientific image: it’s the world in itself according to us.  Talk of molecules, atoms, electrons, quarks, and strings is our talk just as much as talk of colors and plants and animals and haircuts and home runs.

The world of physics is the world as it is in itself according to us.  Arguably, the ‘according to us’ gets the upper hand over the ‘in itself’ relativizing what comes within the former’s  scope much like Kant’s transcendental prefix, Ich denke.  Das ‘ich denke’ muss alle meine Vorstellungen begleiten koennen . . . .  “The ”I think’ must be able to accompany all my representations.” (KdrV, B 131-2)

Arguably, the world of physics is a mind-involving construct arrived at by excluding the mental and abstracting away from the first-person point of view and the life world (Husserl’s Lebenswelt) it reveals.  I am alluding to an idealist approach to the problem of integrating the first- and third-person points of view.  It has its own problems. But why is it inferior to a view like Dennett’s which eliminates as illusory obvious data that are plainly not illusory?

Time was when absolute idealism was the default position in philosophy. Think back to the days of Bradley and Bosanquet. But reaction set in, times have changed, and the Zeitgeist is now against the privileging of Mind and for the apotheosis of Matter.  (But again, matter as construed by us. Arguably, the scientific realist reifies theoretical constructs that we create and employ to make sense of experience.)  Because idealism is out of vogue, the best and brightest are not drawn to its defense, and the brilliant few it attracts are too few to make much headway against the prevailing winds.

Now I’ll tell you what I really think. The problem of integrating the first- and third-person points of view is genuine and perhaps the deepest of all philosophical problems. But it is insoluble by us.  If it does have a solution, however, it certainly won’t be anything like Dennett’s.

Although Dennett’s positive theory is worthless, his excesses are extremely useful in helping us see just how deep and many-sided and intractable the problems are.

IS CONSCIOUSNESS MYSTERIOUS?

Haugeland says it is, above.

My response is that consciousness is mysterious only if you insist on trying to understand it in materialist or  naturalist terms.  For if you try to understand it in these terms, then you are approaching it from a third-person point of view and are bound to miss it.  From the first-person point of view, it is self-evidently certain that there is consciousness.  Nagel 1; Dennett 0.

It is also self-evidently certain that there is consciousness of objects. ‘Consciousness of objects’ is a genitivus objectivus. So consciousness of objects = object-directed consciousness.  If something is appearing to you, you are conscious of it.

Is anything appearing to you right now?  How about the computer in front of you, the desk upon which it sits, and so on.  Plenty of objects are appearing. This is undeniable.  The question is not whether appearing objects exist apart from their appearing. The question is whether any objects are appearing. You will miss the simple point I am making if you think that consciousness, to be real, must itself appear as a particular object among objects.  It does not and cannot since it is what makes possible the appearing of objects in the first place.

Do you understand the sentences you are now reading? You do. So  the  sentences and the words  in them are  appearing.   If there were no object-directed consciousness, you would not understand anything you are reading or hearing.  Indeed, you would not be reading or hearing at all. Are the words appearing to a Cartesian res cogitans, an immaterial thinking substance? I didn’t say anything about that. I am simply pointing out the obvious phenomenological fact that objects are appearing. The words you are reading are appearing to you as objects of visual consciousness.

So the answer to (a) above in the second quotation is this:  we know that understanding without consciousness is impossible because  without consciousness nothing would appear that could be understood or not understood.  ‘Consciousness’ here is of course  consciousness-of.

As for (b) — How do we know that semantic engines (automatic formal systems) won’t ever  have consciousness?  — the question makes no sense since it contradicts the very idea of a semantic engine.  For Haugeland, a semantic engine is one in which the “semantics takes care of itself” once the proper syntactic operations are up and running. But there is no semantics in such a contraption and the semantics cannot “take care of itself” because you cannot get meaning from merely syntactic manipulations of uninterpreted symbols.   What makes a symbol a full-fledged symbol is its interpretation by an Original Intender (an Original Sinn-er, to make Fregean pun out of it).

Computers running programs are wholly meaning -free in and of themselves. All meaning originates from us.

X = Original Intentionality: a semantic engine’s tokens only have meaning because we give meaning to them; their intentionality is derivative, and they don’t mean anything by their tokens any more than books do. Genuine understanding is intentional in its own right and isn’t derived from something else. Haugeland’s response is that actions prove intentionality. Books don’t act, so their intentionality is the author’s, but this isn’t a real counter-example. Robots might be made which act in an intelligent, self-motivated and responsive manner – and seem to act on its opinions, defend them or modify them, which would suggest that they were the robot’s own opinions. So, the response is similar to the last one; original intentionality depends on whether the object has a suitable structure and / or dispositions relative to its environment, so how do we know semantic engines can’t have it?

Haugeland’s response to the objection from original intentionality gives me no good reason to change my view.  He allows that the derivative intentionality of a book comes from the original intentionality of its author but disallows books as counterexamples because they don’t act. That’s true of most books: they just ‘sit’ on the shelf and don’t do anything.  But audio books speak to us, thereby acting.  Will Haugeland say that audio books are self-motivated and act in an intelligent way?  I hope not. As for robots, all Haugeland is able to show is that robots SIMULATE human actions and SIMULATE human thought processes. Beyond that, he is simply begging the question.

I do grant, however, that it really would be wonderful if we could understand the mind-brain relation in us as just a special case of the relation of software to hardware /wetware in a machine running a program. It it easy to understand how cognitive scientists got unduly excited over the prospects of this research program.

The fundamental error of Dennett, Haugeland, et al. is the mistake of thinking that an exhaustive understanding of everything is achievable from a third-person point of view.  And so Dennett attempts the absurd task of explaining consciousness by explaining it away — which issues in the self-refuting claim that consciousness is an illusion.  I explain this in detail in my Substack article Can Consciousness be Explained? Dennett Debunked.

The Hard Problem in the Philosophy of Mind: Comments on Vlastimil Vohanka

May be an image of textMay be an image of text

COMMENTS

What is the so-called ‘hard problem’? Vlastimil thinks it is the problem of specifying which natural (physical) item a given mental item is identical to.  But this is a misuse of the term ‘hard problem.’ As  used in contemporary literature, its meaning is rather more specific.  Although the problem so-labelled has been around for a long time, the label ‘hard problem’ as used in contemporary discussions was first introduced by David Chalmers in a 1995 article and then fully explained in his 1996 book The Conscious Mind.  It is to be understood by contrast with (supposedly) ‘easy problems’ in the philosophy of mind.  So what is the hard problem?  And what is an example of a (supposedly)  easy problem?

The Hard Problem

Suppose I inadvertently touch a hot stove top, experience a pain sensation, and withdraw my hand.  My behavior (stove-touching, hand-withdrawal) is third-person accessible: I see it and others around  me see it. And if no one around me is present? They could see it if they were present.  But surely there is more to pain than pain-behavior. There is the sensation itself.  But what is the sensation itself?  There are two ways of considering it. In one way it too is third-person accessible or outwardly observable. In the other way it is not.

The first way is by taking the pain sensation to be a state or event or process literally internal to (spatially inside of) the organism. In the present example, the organism is the animal  that wears my clothes. So the sensation is a physical process occurring in BV’s brain-body composite. This physical process is then plausibly taken to be the salient cause of the aversive behavior whether physical or verbal  (the hand-withdrawal and/or various spoken obscenities.)

The second way I’ll call phenomenological. The subject of the painful experience simply attends by introspection to the felt pain precisely as it is felt while bracketing (in roughly Husserl’s sense) all such considerations as the causation of the experience and its location in the organism.  In so doing, the subject does not deny that the felt pain has a cause or that the pain has neural correlates located in the brain. He does not even doubt their existence.  He simply leaves those factors out of consideration, and makes no use of them, the better to focus on the undeniable (pace Dennett the Denier) phen0menological datum, the felt pain precisely as it is felt. The subject focuses his attention exclusively on the phenomenal or qualitative features of the experience he is enduring.

The so-called ‘hard problem’ can now be stated. It is the problem of giving an account, consistent with (metaphysical) naturalism, of these qualitative features or properties of sensations,  the so-called qualia (singular: quale), when the sensations are taken, not behavioristically, nor neuro-scientifically, but precisely as they themselves appear from the first-person point of view (POV) of  the subjects who consciously experience them. There is something it is like to experience the phenomenal pain consequent on touching a hot stove. This Nagelian ‘what it is like’ is undeniably real — is there anything more real than pain? — but it cannot be slotted into a wholly third-personal naturalistic ontology. The felt pain is ineluctably subjective: there is simply no place for it in a naturalist world-scheme.

Now suppose you are, for whatever reasons, committed to that naturalist scheme. Then you face a very hard problem indeed, so hard, I would say, as to count as insoluble.  The problem is well-formulated in the inconsistent triad  Vlastimil cites at the top of the page.  Scroll up and take another look at it.  The third proposition cannot be denied on pain of the naturalist’s ceasing to be a naturalist.  The first cannot be denied unless our naturalist is an eliminativist about consciousness, a position with absolute nothing to recommend it. The naturalist’s only hope is somehow to hold onto the second proposition  — consciousness is not physical.   But how? By going mysterian.  But first a word about easy problems, one of which is intentionality.

A (Supposedly) Easy Problem

Not every episode of consciousness is object-directed.  The felt pain I have been using as an example is not object-directed, or an instance of intentionality.  It has a cause outside the body, the hot stove, and a neural correlate inside the brain, Delta-A fiber excitation let us suppose, but the felt pain does not reveal or display or make manifest anything in the way my seeing of a glorious Arizona sunset reveals said sunset to me.  The pain is not of or about anything; the seeing is. We have to distinguish between consciousness and consciousness-of.  Of course, an episode of consciousness-of may have some associated qualia, if I am, say, watching a glorious sunset or moonrise, but there are qualia free object-directed mental states. But this is a complicated special topic we cannot now discuss.  I will just refer you to an earlier entry, Intentionality not a ‘Hard Problem’ for Physicalists?

Mysterianism

What our friend Vlastimil is calling mysterianism about consciousness I take to be the thesis that, while humans either do or can — unclear which he means — know that, in general, every mental item (whether intentional or non-intentional)  is identical to some natural item or other, what they cannot know, and what must remain a mystery to them, is whether any given mental item such as a phenomenal pain felt on a given occasion by a particular subject is both natural and mental.  They cannot know it as both mental and physical.  While experiencing headache pain, for example, I can and do know it self-evidently to be a mental datum, but  I cannot also know it to be wholly natural, even if in reality it is!  Thus our friend is advocating the mysterianism of Colin McGinn.

Vlastimil  distinguishes this mysterian form of naturalism about the mind from an old-school anti-naturalist mysterianism according to which we cannot recognize mental items to be natural items because there are no natural items to which they could be identical.

But here I must raise a question about the tenability of this distinction. If a particular felt pain cannot be identified with, and thus reduced to, a particular instance of Delta-A fiber stimulation because the former has properties the latter cannot have, and vice versa, why call this mysterianism?  Where is the mystery? What we would have  here is a straight-forward  argument which, if sound, shows that some or all mental items cannot be reduced to natural items. What we would have is an argument for dualism, whether property dualism or substance dualism.

To have a mystery in the strict sense you need propositions that appear and cannot fail to appear to intellects of our constitution as logically inconsistent, but in reality, and beyond our ken, are somehow consistent. Trinity and Incarnation are mysteries in this sense.

Similarly for McGinn: We cannot understand how qualia, which are undeniably real, are identical to natural items, and yet they are!  It is imply beyond our ken.  Our cognitive architecture is so structured as to disallow any insight on our part as to how this pain I am feeling is nothing more than a physical occurrence!

Is the World Inconceivable Apart from Consciousness? (Version 2.0)

That depends. It depends on what ‘world’ means.

Steven Nemes quotes Dermot Moran on the former’s Facebook page:

[1] In contrast to the outlook of naturalism, Husserl believed all knowledge, all science, all rationality depended on conscious acts, acts which cannot be properly understood from within the natural outlook at all. [2] Consciousness should not be viewed naturalistically as part of the world at all, since consciousness is precisely the reason why there was a world there for us in the first place. [3] For Husserl it is not that consciousness creates the world in any ontological sense—this would be a subjective idealism, itself a consequence of a certain naturalising tendency whereby consciousness is cause and the world its effect—but rather that the world is opened up, made meaningful, or disclosed through consciousness.  [4] The world is inconceivable apart from consciousness. [5] Treating consciousness as part of the world, reifying consciousness, is precisely to ignore consciousness’s foundational, disclosive role. (Dermot Moran, Introduction to Phenomenology, p. 144.)

This strikes me as confused. I will go through it line by line. I have added numbers in brackets for ease of reference.

Ad [1]. I basically agree.  But while conscious acts cannot be properly understood from within die natürliche  Einstellung, it doesn’t follow that they cannot be understood “at all” from within the natural attitude or what Moran is calling the natural outlook. So I would strike the “at all.” I will return to this issue at the end.

Ad [2].  Here the trouble begins. I grant that conscious acts cannot be properly understood in wholly materialistic or naturalistic terms. They cannot be understood merely as events in the natural world. For example, my thinking about Boston cannot be reduced to anything going in my brain or body when I am thinking about Boston. Conscious acts are object-directed. They have the property philosophers call ‘intentionality.’  Intentionality resists naturalistic reduction.

And I grant that there is a sense in which there would not be a world for us in the first place if there were no consciousness. A world for us is a world that appears to us as conscious beings.

But note the equivocation on ‘world.’ It is first used to refer to nature itself, and then used to refer to the openness or apparentness of nature, nature as it appears to us and has meaning for us.  Obviously, without consciousness, nature would not appear, but this is not to say that consciousness is the reason why there is a natural world in the first place.  To say that would be to embrace a form of idealism.

Ad [3] We are now told that this is not “a subjective idealism.” I agree.  But note that the world that is disclosed and made meaningful is not the world that is inconceivable without consciousness.  The equivocation on ‘world’ persists.  There is world in the transcendental-phenomenological sense as the ‘space’ within which things are disclosed and become manifest, and there is world as the things disclosed.  These are plainly different even if there is no epistemic access to the latter except via the former.

Ad [4] Therefore, to be precise, we should say that the world as the ‘space of disclosure’ is inconceivable without consciousness. But this ‘space of disclosure’ is not the same as the natural world, which is not inconceivable without consciousness.  If you say that it is, then you are adopting a form of metaphysical idealism, which is what Husserl in the end does.

In the end, he reduces Being to Meaning (Sein to Seinsinn) or Being (Sein) to ontic validity (Seinsgeltung).  Accordingly, beings in the world are constituted (a piece of Husserlian jargon) as beings by transcendental consciousness.  This is the upshot of the transcendental-phenomenological reduction. What we naively take in the natural attitude to exist in themselves, on their ontological own, so to speak, things like rocks and planets and galaxies, are in truth intentional objectivities constituted in transcendental consciousness.

Ad [5] In the final sentence, ‘world’ clearly refers to the physical realm, nature. I agree that it would be a mistake to reify consciousness, to identify it with any physical thing or process. Consciousness plays a disclosive role. It is pre-mundane, transcendental.  As OF the world — genitivus objectivus — consciousness is not IN the world. But the world in this sense, the world that consciousness is not IN, is conceivable apart from consciousness.  If it were not, then Roman Ingarden’s realism and Thomist realism, and other types, would not be conceivable, which they plainly are.

And so the confusion remains.  The world in the specifically phenomenological sense, the world as the ‘space’ within which things are disclosed — compare Heidegger’s Lichtung or clearing — is inconceivable without consciousness. But the world as that which is disclosed, opened up, gelichtet, made manifest and meaningful, is NOT inconceivable apart from consciousness. If you maintain otherwise, then you are embracing a form of metaphysical idealism.

So I’d say that Moran and plenty of others are doing the ‘Continental Shuffle’ as I call it: they are sliding back and forth between two senses  of ‘world.’  Equivalently, they are conflating the ontic and the broadly epistemic.  I appreciate their brave attempt at undercutting the subject-object dichotomy and the idealism-realism problematic.  But the brave attempt does not succeed.  A mental act of outer perception, say, is intrinsically intentional or object-directed: by its very sense it purports to be of or about something that exists apart from any and all mental acts to which it appears.  To speak like a Continental, the purport is ‘inscribed in the very essence of the act.’  But there remains the question whether the intentional object really does exist independently of the act. There remains the question whether the intentional object really exists or is merely intentional.  Does it enjoy esse reale, or only esse intentionale?

I recommend to my friend Nemes that he read Roman Ingarden’s critique of Husserl’s idealism.  I also recommend that he read Husserl himself (in German if possible) rather than the secondary sources he has been citing, sources some of which are not only secondary, but second-rate.

To return to what I said at the outset: Conscious acts cannot be properly understood naturalistically.  But surely a full understanding of them must explain how they relate to the goings-on in the physical organisms in nature that support them.  A satisfactory philosophy cannot ignore this. And so, to end on an autobiographical note, this was one of the motives that lead me beyond phenomenology.

Could an Advanced AI System be Conscious?

My Substack uploads continue. You can read them without subscribing simply by heading over there. Here is the latest. Perhaps David Brightly can poke a hole or two in it.

Speaking of Brightly, here is a post of his in which he engages me on the topic of the irreducibility of intentionality.  Mr. Brightly is a model of clarity, precision, and gentlemanliness.

Could Qualia Terms and Neuroscience Terms have the Same Reference?

Could Frege’s sense-reference distinction be put to work? I think not.

Top o’ the Stack.

I made the point a while back that the vocabularies of phenomenology and neuroscience are radically disparate, such that nonsense arises when one says things like, ‘This burnt garlic smell is identical to a brain state of mine.’ To which a Vietnam veteran, altering the example, replied by e-mail:

. . . when a neuroscientist says your smelling this odor as napalm is nothing but a complex neural event activating several regions of the brain…, he isn’t claiming you can replace your talk about smells with talk about neural signals from the olfactory bulb. Different ways of talking have evolved for different purposes. But he is saying that beneath these different ways of talking & thinking there is just one underlying reality, namely, neural events in our brain.

The idea, then, that is that are are different ways of referring to the same underlying reality. And so if we deploy a simple distinction between sense and reference we can uphold the materialist/physicalist reduction of qualia to brain states. Well, I doubt it. In fact, I deny it.

Mind without Consciousness?

David Brightly in a recent comment writes,

[Laird] Addis says,

The very notion of language as a representational system presupposes the notion of mind, but not vice versa.

I can agree with that, but why should it presuppose consciousness too?

In a comment under this piece you write,

Examples like this cause trouble for those divide-and-conquerers who want to prise  intentionality apart from consciousness with its qualia, subjectivity, and what-it-is-like-ness,  and work on the problems separately, the first problem being supposedly tractable while the second is called the (intractable) Hard Problem (David Chalmers). Both are hard as hell and they cannot be separated. See Colin McGinn, Galen Strawson, et al.

Could you say a bit more on this?

I’ll try.  You grant that representation presupposes mind, but wonder why it should also presuppose consciousness.  Why can’t there be a representational system that lacks consciousness?  Why can’t there be an insentient, and thus unconscious, machine that represents objects and states of affairs external to itself? Fair question! 

Here is an example to make the problem jump out at you. Suppose you have an advanced AI-driven robot, an artificial French maid, let us assume, which is never in any sentient state, that is, it never feels anything.  You could say, but only analogically, that the robot is in various ‘sensory’ states, states  caused by the causal impacts of physical objects against its ‘sensory’ transducers whether optical, auditory, tactile, kinaesthetic . . . but these ‘sensory’ states  would have no associated qualitative or phenomenological features.  Remember Herbert Feigl? In Feiglian terms, there would be no ‘raw feels’ in the bot should her owner ‘feel her up.’  Surely you have heard of Thomas Nagel. In Nagelian terms, there would be nothing it is like for the bot to have her breasts fondled.  If her owner fondles the breasts of his robotic French maid, she feels nothing even though she is programmed to respond appropriately to the causal impacts via her linguistic and other behavior.   “What are you doing, sir? I may be a bot but I am not a sex bot! Hands off!” If the owner had to operate upon her, he would not need to put her under an anaesthetic. And this for the simple reason that she is nothing but an insensate machine.

I hope Brightly agrees with me that verbal and nonverbal behavior, whether by robots or by us, are not constitutive of  genuine sentient states. I hope he rejects analytical (as opposed to methodological) behaviorism, according to which feeling pain, for example,  is nothing more than exhibiting verbal or nonverbal pain-behavior.  I hope he agrees with me that the bot I described is a zombie (as philosophers use this term) and that we are not zombies.  

But even if he agrees with all that, there remains the question: Is the robot, although wholly insentient, the subject of mental states, where mental states are intentional (object-directed) states?  If yes, then we can have mind without consciousness, intrinsic intentionality without subjectivity, content without consciousness.

Here are some materials for an argument contra.

P1 Representation is a species of intentionality. Representational states of a system (whether an organism, a machine, a spiritual substance, whatever) are intentional or object-directed states.

P2 Such states involve contents that mediate between the subject of the state and the thing toward which the state is directed.  Contents are the cogitata in the following schema: Ego-cogito-cogitatum qua cogitatum-res. Note that ‘directed toward’ and ‘object-directed’ are being used here in such a way as to allow the possibility that there is nothing in reality, no res, to which these states are directed.  Directedness is an intrinsic feature of intentional states, not a relational one.  This means that the directedness of an object-directed state is what it is whether or not there is anything in the external world to which the state is directed. See Object-Directedness and Object-Dependence for more on this.

As for the contents, they present the thing to the subject of the state. We can think of contents as modes of presentation, as Darstellungsweisen in something close to Frege’s sense.     Necessarily, no state without a content, and no content without a state.  (Compare the strict correlation of noesis and noema in Husserl.) Suppose I undergo an experience which is the seeing as of  a tree.  I am the subject of the representational state of seeing and the thing to which the state is directed, if it exists, is a tree in nature.  The ‘as of‘ locution signals that the thing intended in the state may or may not exist in reality.

P3 But the tree, even if it exists in the external world, is not given, i.e., does not appear to the subject, with all its aspects, properties, and relations, but only with some of them. John Searle speaks of the “aspectual shape” of intentional states. Whenever we perceive anything or think about anything, we always do so under some aspects and not others.  These aspectual features are essential to the intentional state; they are part of what make intentional  states the states that they are. (The Rediscovery of the Mind, MIT Press, 1992, pp. 156-157) The phrase I bolded implies that no intentional state that succeeds in targeting a thing (res) in external world is such that every aspect of  the thing is before the mind of the person in the state.

P4 Intentional states are therefore not only necessarily of something; they are necessarily of something as something.  And given the finitude of the human mind, I want to underscore the fact that  even if every F is a G, one  can be aware of x as F without being aware of  x as G.   Indeed, this is so even if necessarily (whether metaphysically or nomologically) every F is a G. Thus I can be aware of a moving object as a cat, without being aware of it as spatially extended, as an animal, as a mammal, as an animal that cools itself by panting as opposed to sweating, as my cat, as the same cat I saw an hour ago, etc.  

BRIGHTLY’S THEORY (as I understand it, in my own words.)

B1. There is a distinction between subpersonal and personal contents. Subpersonal contents exist without the benefit of consciousness and play their mediating role in representational states in wholly insentient machines such as the AI-driven robotic maid.  

B2. We attribute subpersonal contents to machines of sufficient complexity and these attributions are correct in that these machines really are intentional/representational systems.

B3. While it is true that the only intentional (object-directed) states of which we humans are aware are conscious intentional states, that they are  conscious is a merely contingent fact about them. Thus, “the conditions necessary and sufficient for content are neutral on the question whether the bearer of the content happens to be a conscious state. Indeed the very same range of contents that are possessed by conscious creatures could be possessed by creatures without a trace of consciousness.” (Colin McGinn, The Problem of Consciousness, Blackwell 1991, p. 32.

MY THEORY

V1. There is no distinction between subpersonal and personal contents. All contents are contents of (belonging to) conscious states. Brentano taught that all consciousness is intentional, that every consciousness is a consciousness of something.  I deny that, holding as I do that some conscious states are non-intentional. But I do subscribe to the Converse Brentano Thesis, namely, that all intentionality is conscious. In a slogan adapted from McGinn though not quite endorsed by him, There is no of-ness without what-it-is-like-ness. This implies that only conscious beings can be the subjects of original or intrinsic intentionality.  And so the  robotic maid is not the subject of intentional/representational states. The same goes for the cerebral processes transpiring  in us humans when said processes are viewed as purely material: they are not about anything because there is nothing it is like to be them.  Whether one is a meat head or a silicon head, no content without consciousness! Let that be our battle cry.

And so, when the robotic maid’s voice synthesizer ‘says’ ‘This shelf is so dusty!’ it is only AS IF ‘she’ is thereby referring to a state of affairs and its constituents, the shelf and the dust.  ‘She’ is not saying anything, sensu stricto, but merely making sounds to which we original-Sinn-ers, attribute meaning and reference. Thinking reference (intentionality) enjoys primacy over linguistic reference. Cogitation trumps word-slinging. The latter is parasitic upon the former.  Language without mind is just scribbles, pixels, chalk marks, indentations in stone, ones and zeros. As Mr. Natural might have said, “It don’t mean shit.” An sich, und sensu stricto.

V2. Our attribution of intentionality to insentient systems is merely AS IF.  The robot in my example behaves as if it is really cognizant of states of affairs such as the dustiness of the book shelves and as if it really wants to please its boss while really fearing his sexual advances.  But all the real intentionality is in us who makes the attributions.  And please note that our attributing of intentionality to systems whether silicon-based or meat-based that cannot host it is itself real intentionality. It follows, pace Daniel Dennett, that intentionality cannot be ascriptive all the way down (or up). But Dennett’s ascriptivist theory of intentionality calls for a separate post.

V3. It is not merely a contingent fact about the intentional state that we our introspectively aware of that they are conscious states; it is essential to them.

NOW, have I refuted Brightly ? No! I have arranged a standoff.  I have not refuted but merely neutralized his position by showing that it is not rationally coercive.  I have done this by sketching a rationally acceptable alternative. We have made progress in that we now both better understand the problems we are discussing and our different approaches to them.

Can we break standoff? I doubt it, but we shall see.

More on the Unity of Consciousness: From Self to Immortal Soul?

Suppose I see a black cat. The act of visual awareness in a case like this is typically, even if not always, accompanied by a simultaneous secondary awareness of the primary awareness.  I am aware of the cat, but I am also aware of being aware of the cat.  How does the Humean* account for one's awareness of being aware? He could say, plausibly, that the primary  object-directed awareness is a subject-less awareness. But he can't plausibly say that the secondary awareness is subject-less.   For if both the primary awareness (the awareness of the cat) and the secondary awareness (the awareness of the primary awareness) are subject-less, then what makes the secondary awareness an awareness of the primary awareness? What connects them? The two awarenesses cannot just occur; they must occur in the same subject, in the same unity of consciousness.

Suppose that in Socrates there is an awareness of a cat, and in God there is an awareness of Socrates' awareness of a cat.  Those two awarenesses would not amount to there being in Socrates an awareness of a cat together with a simultaneous secondary awareness of being aware of a cat.  But it is phenomenologically evident that the two awarenesses do co-occur. We ought to conclude that the two awarenesses must be together in one subject, where the subject is not the physical thing in the external world (the animal that wears Socrates' toga, for example), but the I, the self, the subject.

What I have just done is provide phenomenological evidence of the existence of the self that Hume claimed he could not find. Does it follow that this (transcendental) self is a simple substance that can exist on its own without a material body? That's a further question.  To put it another way: do considerations anent the unity of consciousness furnish materials for a proof of the simplicity, and thus the immortality, of a substantial soul?  Proof or paralogism? 

__________

*A Humean for present purposes  is one who denies that there is a self or subject that is aware; there is just awareness of this or that. Hume, Sartre, and Butchvarov are Humeans in this sense.

AI and the Unity of Consciousness

Top AI researchers such as Geoffrey Hinton, the "Godfather of AI,"  hold that advanced AI systems are conscious.  That is far from obvious, and may even be demonstrably false if we consider the phenomenon of the unity of consciousness.  I will first explain the phenomenon in question, and then conclude that AI systems cannot accommodate it.

Diachronic Unity of Consciousness, Example One

Suppose my mental state passes from one that is pleasurable to one that is painful.  Observing a beautiful Arizona sunset, my reverie is suddenly broken by the piercing noise of a smoke detector.  Not only is the painful state painful, the transition from the pleasurable state to the painful one is itself painful.  The fact that the transition is painful shows that it is directly perceived. It is not as if there is merely a succession of consciousnesses (conscious states), one pleasurable the other painful; there is in addition a consciousness of their succession.  For there is a consciousness of the transition from the pleasant state to the painful state, a consciousness that embraces both of the states, and so cannot be reductively analyzed into them.  But a consciousness of their succession is a consciousness of their succession in one subject, in one unity of consciousness.  It is a consciousness of the numerical identity of the self through the transition from the pleasurable state to the painful one.  Passing from a pleasurable state to a painful one, there is not only an awareness of a pleasant state followed by an awareness of a painful one, but also an awareness that the one who was in a pleasurable state is strictly and numerically the same as the one who is now in a painful state.  This sameness is phenomenologically given, although our access to this phenomenon is easily blocked by inappropriate models taken from the physical world.  Without the consciousness of sameness, there would be no consciousness of transition.

What this phenomenological argument shows is that the self cannot be a mere diachronic bundle or collection of states.  The self is a transtemporal unity distinct from its states whether these states are taken distributively (one by one) or collectively (all together).

May we conclude from the phenomenology of the situation that there is a simple, immaterial, meta-physical substance that each one of us is and that is the ontological support of the phenomenologically given unity of consciousness?  May we make the old-time school-metaphysical moves from the simplicity of this soul substance to it immortality? Maybe not! This is a further step that needs to be carefully considered. I don't rule it out, but I also don't rule it in. I don't need to take the further step for my present purpose, which is merely to show that a computing machine, no matter how complex or how fast its processing, cannot be conscious.  No material system can be conscious.  For the moment I content myself with the negative claim: no material system can be conscious. It follows straightaway that no AI system can be conscious.

Diachronic Unity of Consciousness, Example Two

Another example is provided by the hearing of a melody.  To hear the melody Do-Re-Mi, it does not suffice that there be a hearing of Do, followed by a hearing of Re, followed by a hearing of Mi.  For those three acts of hearing could occur in that sequence in three distinct subjects, in which case they would not add up to the hearing of a melody.  (Tom, Dick, and Harry can divide up the task of loading a truck, but not the ‘task’ of hearing a melody, or that of understanding a sentence.)  But now suppose the acts of hearing occur in the same subject, but that this subject is not a unitary and self-same individual but just the bundle of these three acts, call them A1, A2, and A3.  When A1 ceases, A2 begins, and when A2 ceases, A3 begins: they do not overlap.  In which act is the hearing of the melody?  A3 is the only likely candidate, but surely it cannot be a hearing of the melody.

This is because the awareness of a melody involves the awareness of the (musical not temporal)  intervals between the notes, and to apprehend these intervals there must be a retention (to use Husserl’s term) in the present act A3 of the past acts A2 and A1.  Without this phenomenological presence of the past acts in the present act, there would be no awareness in the present of the melody.  This implies that the self cannot be a mere bundle of perceptions externally related to each other, but must be a peculiarly intimate unity of perceptions in which the present perception A3 includes the immediately past ones A2 and A1 as temporally past but also as phenomenologically present in the mode of retention.  The fact that we hear melodies thus shows that there must be a self-same and unitary self through the period of time between the onset of the melody and its completion.  This unitary self is neither identical to the sum or collection of A1, A2, and A3, nor is it identical to something wholly distinct from them.  Nor of course is it identical to any one of them or any two of them.  This unitary self is co-given whenever one hears a melody.  (This seems to imply that all consciousness is at least implicitly self-consciousness. This is a topic for a later post.)

Diachronic -Synchronic Unity of Consciousness

Now consider a more complicated example in which I hear two chords, one after the other, the first major, the second minor.   I hear the major chord C-E-G, and then I hear the minor chord C-E flat-G.  But I also hear the difference between them.   How is the awareness of the major-minor difference possible? One condition of this possibility is the diachronic unity of consciousness. But there is also a second condition. The hearing of the major chord as major cannot be analyzed without remainder into an act of hearing C, an act of hearing E, and an act of hearing G, even when all occur simultaneously.  For to hear the three notes as a major chord, I must apprehend the 1-3-5 musical interval that they instantiate.  But this is possible only because the whole of my present consciousness is more than the sum of its parts.  This whole is no doubt made up of the part-consciousnesses, but it is not exhausted by them.  For it is also a consciousness of the relatedness of the notes.  But this consciousness of relatedness is not something in addition to the other acts of consciousness: it includes them and embraces them without being reducible to them.  So here we have an example of the diachronic-synchronic unity of consciousness.

These considerations appear to put paid to the conceit that AI systems can be conscious.

Or have I gone too far? You've heard me say that in philosophy there are few if any rationally compelling,  ineluctably decisive, arguments for substantive theses.  Are the above arguments among the few? Further questions obtrude themselves, for example, "What do you mean by 'material system'?"  "Could a panpsychist uphold the consciousness of advanced AI systems?"

Vita brevis, philosophia longa.

Searle, Dennett and Zombies

Another in a series of Substack uploads debunking the brilliant scientistic sophistry of the late Daniel Dennett. 

I have over a thousand dollars in pledges. Should I monetize or not? It seems rude and arrogant not to graciously accept gifts. On the other hand, philosophy for me is a labor of love, a vocation, a high calling . . . . 

Galen Strawson on Nicholas Humphrey on Consciousness

Substack latest.

Strawson is right against Humphrey, but his own theory is worthless.

See also: The Problem of Consciousness and Galen Strawson's Non-Solution

UPDATE (10/3).  A friend referred me to this article which I judge to be very bad indeed. See if you can make out what is wrong with it.

Age Quod Agis: Agent and Awareness

Too much attention is wasted on what we did do and what we will do, and not enough on what we are doing. Age quod agis. "Do what you are doing."  A excellent maxim. A non-philosopher will take it as such and then move on. The philosopher lingers and goes deeper.

Verbally a tautology, the admonition expresses a non-tautological truth: attend to what you are doing.  I cannot fail to do what I am doing, but I can fail to attend to what I am doing. The admonition is in the same logical boat with "Be here now!" and "Live in the present!"

How could I fail to be here now? Where else would I be? And when else would I be? But that would be to miss the point. The tautological form of words expresses a non-tautological thought: Attend to the moment and be aware of your situation.

For a human being, to be is not merely to exist as a thing among things, but to be aware. The Being of a human being involves an element of material facticity — you are this indigent material thing right here — but also an element of transcendence in that, as aware, you are way beyond the miserable chunk of matter your awareness inhabits. You are way beyond it by being aware of the not-self. The not-self includes not only everything other than your body, but also your body inasmuch as your body and its parts are objects of awareness and thus not identical to you as subject of awareness. You are not merely a thing in the world, but also, as the subject of awareness,  a being  for whom there is a world.

As for living in the present, this is not a mere biological living. As a bit of nature's fauna, how could you biologically live other in the temporal present? To live in the present, as per the admonition, is to attend to the present, to impede the outward scatter of your thoughts, to bend back the outward intentionality (object-directedness) of mind to the present moment and its contents. You draw in your thoughts from the diaspora of the past and the future and the elsewhere in space and the elsewhere in general and bring them home. You could call it 'bringing it all back home.'  You could call it spiritual intro-version, or swimming upstream to the Source of thought's river. ("Man is a stream whose source is hidden" (Ralph Waldo Emerson).

The Being of the human being is a living, but not a merely biological living, not a mere living as understood by the objectifying natural science of biology. The ineluctable subjectivity ingredient in the Being of human beings cannot be understood from the point of view of biology.

Consider now the sentence 'I am hungry' asserted by BV.  It is true now at 12:45 PM.  What is it about? It is about BV, a publicly identifiable person. What does it predicate of BV? It predicates the property of being hungry. The predicational tie is signified by the copula 'am.' Does this copula express merely the object BV's instantiation of the property? No, it also expresses the speaker's awareness that he himself is hungry. Property-possession in a human being is more than a merely objective relation. This fact complements the earlier one about the ineluctable subjectivity of the Being of human beings. Both the Being and the Being-propertied of human beings is unlike anything else in the world.

The Grain Problem

Ed Buckner writes,

Here is another problem that needs to be carefully phrased.

I want to say that the pitch of a musical note is continuous through time. I mean, at any point in continuous time, i.e. time as specified by the real numbers, the pitch of the note (e.g. middle C) is the same.

However, the “physical” property that grounds the pitch is not continuous, but rather a cycle of different events.

That strikes me as a problem for the kind of physicalism according to which qualities “as we perceive them” are identical with the properties that ground them. For pitch is temporally continuous, the oscillation that grounds it is not temporally continuous, ergo etc.

It is a problem indeed, Ed, although I have questions about your formulation of it.

The problem is known in the trade as the Grain Problem. Whether it surfaces before Sir Arthur Eddington, I don't know, but he raises it, or at least anticipates it with his question about the 'two tables.'  A lot of work was done on the Grain Problem by the great American philosopher Wilfrid Sellars, son of the rather less distinguished Roy Wood Sellars, but nonetheless a quantity to be reckoned with in his day.

Sellars Wilfrid IntentionalityHere is Sellars fils  in his seminal essay, "Philosophy and the Scientific Image of Man," reprinted in his Science, Perception, and Reality (Routledge, 1963). The portion I am about to quote is from pp. 35-37. I take the text from Chrucky's online version.

It is worth noting that we have here a recurrence of the essential features of Eddington's 'two tables' problem — the two tables being, in our terminology, the table of the manifest image and the table of the scientific image. There the problem was to 'fit together' the manifest table with the scientific table. Here the problem is to fit together the manifest sensation with its neurophysiological counterpart. And, interestingly enough, the problem in both cases is essentially the same: how to reconcile the ultimate homogeneity of the manifest image with the ultimate non-homogeneity of the system of scientific objects.

BV: Whether we are discussing colors with Sellars or sounds with Buckner, it is the same problem, that of reconciling the homogeneity of the manifest or phenomenal sensory quality with the non-homogeneity of the underlying  scientific explanatory posits. For Sellars, of course, these posits are not mere posits but ultimately real, as you will see if you read below the fold.

Buckner's formulation above leaves something to be desired, however. He cites the continuous perception over time of the same note, middle C, let us say. But then in the very next sentence he reverts to a rarefied mathematical concept of continuity, thereby mixing phenomenological description with a mathematico-scientific construct.   He thereby conflates phenomenal continuity with mathematical continuity.  When I hear middle C sounding from an organ, say, over a non-zero interval of time, five seconds say, do I hear a series of points of time — a series of temporally extension-less moments — the cardinality of which is 2-to-the-aleph-nought? No. (The cardinality of the set of real numbers (cardinality of the continuum) is) 

And then Ed goes on to say that "the 'physical' property that grounds the pitch is not continuous, but rather a cycle of different events." But that is not right either. Middle C depicted on an oscilloscope shows up as a sine wave:

Middle_C _or_262_hertz _on_a_virtual_oscilloscope

Obviously the sine wave is continuous. What Ed wants to say, of course, is that the heard sound, the phenomenal sound, does not fluctuate as does the physical reality does, the physical reality that "grounds the pitch." Ed is equivocating on 'continuous.'

But I know what he is getting at, and it is a genuine problem. I am merely complaining about his  formulation of it. Now back to Sellars, whose solution to the problem is not clear to me.

 

Continue reading “The Grain Problem”