I would be tempted to use a Kantian analogy here: "percepts without concepts are blind". Thus the recognition process applies some kind of filter (concepts) to the incoming huddle of sensory input (blind percepts), the end result of which is meaningful (non-blind) qualia. So I would not say that the real-time recognition process is like qualia, because the process needs to be distinguished from the product.

@ Karl Pfeifer

The Kantian analogy is great! Thanks.

Deep learning is using concepts (thousands of images in the training sequence). So the process would not be so blind.

On the other side is no relation of the training material and the neuronal net which is trained. The result of the perception has no meaning for the program, the meaning is only in the mind of the user.

Anyway, I feel a similarity in the result.

The construction or re-construction of sensory experience is what creates perception, whether limited or connected to others makes little difference-and qualia is employed to identify an acknowledged process. That qualia can differ in time and impact, and whether in several forms at once-I would say, yes.

Relative to the visual sense in general, I have arrived at a useful working hypothesis. It is that cognition in all its aspects is probably achieved by learned experience and high speed pattern comparison. Laying down an appropriate comparative basis starts at birth. Think if you will of the acquisition of language or watch the touchy feely approach of an infant who is in effect busy storing the comparative instances it needs to explain its environment. At a much deeper level, you might consider how a skilled tennis player gains the ultra-high-speed ability to perceive and predict the precise trajectory of a 100 mph service stroke without doing any calculations.

David, arn't these specific phenomena in a varied field or canopy of phenomenons.

Yes indeed Stanley, that is a key point of my analytical approach. As an aid to personal comprehension, I treat all possible quale as the product of some sort of instance-specific and super fast pattern recognition process. Then, to achieve a wide enough range of sensory cognition, I regard the brain as a truely phenominal 'comparator'. The principle is that for any given instance of cognition to occur you must have an appropriate comparative basis from learned-experience and this may be from life in general or from a particular set of relevant instances.

Stanley,

thanks for your answer. Your answer above reinforces my assumption, which I have asked as a question.

David,

" It is that cognition in all its aspects is probably achieved by learned experience and high speed pattern comparison. "

I agree with your hypothesis. Please go one step further: we learn to interpret our sensory input with our body movements. Our fingers scan the surface of things and we look at these things. So we learn the geometrical shape. It is an integration process we have also with distances. We have to move to "understand" what distance means and how far something is.

Could you clarify the original question, please? I do not see what the term qualia has to do with your question - you're asking about meaningful object recognition, right? How do qualia, as subjective experiential qualities, figure in your question?

Joachim,

thank you for your demand to clarify my initial question.

Today we have fast algorithms like tensor flow, yolo and others in deep leaning to detect objects in a picture. These algorithms add a new item to the picture: the object identification. IMHO this is a little bit like qualia in human perception. We do not see colored pixels but colored impressive objects. This analogy was the intention to my initial question.

I was not asking about meaningful object recognition because meaning is a hard nut for philosophers and computer scientists. A machine is not able to detect meaning at all because meaning is only relevant for living organisms which has needs and feelings or emotions as well as remembers and associations which are linked to an object. Meaning is always subjective. Today computers are not able to compute meaning but I think it is the next step in AI. The meaning of an object can be understood as a whole network of relations which is active or highlighted when an object is identified. May be in qualia we have this cloud of meaning included ...

I would be happy to get your approval here. Of course, I am also pleased with critical remarks that broaden my view of things.

Wilfried,

thank you very much for your reply! Personally, I am finding it a bit hard to accept that qualia should come easier to a machine than meaning. But perhaps there is only a difference in terminology here. My primary understanding of the term "qualia" is associated with philosophers like Thomas Nagel or David Chalmers, who think of qualia as subjective experiences. For each "quale" there is something it is like to experience this quale. E.g. what it is like to lie in ths grass on a sunny spring day, what it is like to taste chocolate ice-cream, what it is like to perceive one's surroundings with sonar etc. Given this understanding I find it difficult to imagine that a machine - even if it can perceive objects - will automatically have an experience of what it is like to have this perception.

Does your use of the term "qualia" differ from the one I delineated?

Joachim,

I agree with your statement for qualia. I also feel a discrepancy of my subjective experience and the simple naming of an object by a computer program.

Your example "lieing in the gras ..." is very complex. There are lots of impressions ant the quale of this situation is like the sound of an orchestra of qualia. To analyse a quale it is better to reduce the impressions to one category.

If we reduce the visual impression to one color (red) we have really more than only the name of the color ("red") but also a lot of very subjective associations. If I have understood your answer correctly, you answered the initial question with "no" because of the meaning wich is linked with the name of the color or the name of the object.

Well I agree. Within the network of linked associations (within the meaning) we can find the subjective experience.

Wilfried,

thank you again for your further clarification. Given your explanation, I would indeed answer "no". I think there are many forms of object recognition which are insufficient for having qualia. For one, if a machine is able to distinguish all cases of "red" from all cases of "non-red", I would be inclined to say it can recognize red. Clearly, there are many ways of doing so which do not require qualitative experience on the part of the machine. In fact, even in humans, we might say that our eyes alone are able to distinguish cases of red from cases of non-red - but whatever creates the associated qualia is not part of our eyes, but a higher cognitive function.

Perhaps I also gravitate toward a sparser meaning of "meaning" than you do. I think if a machine can token a symbol in cases of red, and not token it in cases of non-red, it can (symbolically/linguistically) "mean" (i.e. refer to) red. Again, there are many ways in which a machine can mean red (in this sense) without having qualitative experience.

Therefore, I do not agree that there already needs to be " a lot of very subjective associations" involved, as you say. Both recognition and meaning are possible without experiential subjectivity. (Even though I think there is subjectivity in the sense of agency: Entities which mean things are agents. My point is, they need not be experiencing agents.)

"Our sensory input is quite different to the quale we are aware in a scene."

This is the dogma. Let's take a page from the non-Euclidean geometers and assume that it's incorrect.

What sort of geometry arises?

______________________

A theory that yields "maybe" as an answer should be recognized as an inaccurate theory.

~ 't Hooft

It seems clear that the present quantum mechanics is not in its final form [...] I think it very likely, or at any rate quite possible, that in the long run Einstein will turn out to be correct.

~Dirac

There is now in my opinion no entirely satisfactory interpretation of quantum mechanics.

~Weinberg

Can it really be true that Einstein, in any significant sense, was as profoundly "wrong" as the followers of Bohr maintain? I do not believe so. I would, myself, side strongly with Einstein in his belief in a submiscroscopic reality, and with his conviction that present-day quantum mechanics is fundamentally incomplete.

~Penrose

Whatever the meaning assigned to the term 'complete,' the following requirement for a complete theory seems to be a necessary one: every element of the physical reality must have a counterpart in the physical theory.

~EPR

If you ask a physicist what is his idea of yellow light, he will tell you that it is transversal electromagnetic waves of wavelength in the neighborhood of 590 millimicrons. If you ask him: But where does yellow come in? he will say: In my picture not at all, but these kinds of vibrations, when they hit the retina of a healthy eye, give the person whose eye it is the sensation of yellow.

~Schrödinger

Thus the colors with their various qualities and intensities fulfill the axioms of vector geometry if addition is interpreted as mixing; consequently, projective geometry applies to the color qualities.

~Weyl

The principle of duality in projective geometry states that we can interchange point and line in a theorem about figures lying in one plane and obtain a meaningful statement. Moreover, the new or dual statement will itself be a theorem – that is, it can be proven.

~Kline

While a proper understanding of M-theory still eludes us, much is now known about it. In particular the various geometric results that have emerged from string theory become related in interesting but mysterious ‘dualities’ whose real meaning has yet to be discovered.

~Atiyah

______________________

See also Smolin's paper:

"Matrix models as non-local hidden variables theories"

https://arxiv.org/abs/hep-th/0201031

Quantum mechanics does not help to understand consciousness.

Roman,

thank you for your highly qualified answer.

I would be very happy if you could redound some sentences to my initial question.

Joachim,

your answer is complex so I try to answer part per part.

"... but whatever creates the associated qualia is not part of our eyes, but a higher cognitive function. " I agree. Naturally a quale is not part of the sense organ but is part of a lower brain function. This necessary brain function was the reason of my question above. First we have to detect objects and then we add value to the recognized objects. These objects are the categories of our recognition process. In your example colors are such categories and you ask for the quale of "red" as Nagel did. We have a very impressive quale by seeing red and I think therein is nothing mysterious. It is first of all the representation of a special visual stimulus and second the additum value. May be there are also some physiological reactions ...

The value is a subjective qualification which is hardly to identify.

" Both recognition and meaning are possible without experiential subjectivity. "

Nowadays we have robots (autonomous cars or simple vacuum cleaner). They do not have any quale but they do react to stimuli. Naturally this is possible but they are not aware of it. It is a simple stimulus-response-interaction. Qualia and consciousness must have some advantages.

We are able to manipulate objects in our mind and we know that. Well, robots are also able to manipulate data but it is a search process (algorithm).

"Quantum mechanics does not help to understand consciousness."

O, well, if you say so. Any other proclamations you'd care to make?

When writing about transcendental issues, be transcendentally clear.

~Descartes

Joachim,

" a machine can mean red (in this sense) without having qualitative experience "

That's true. My agreement does not mean that I am satisfied with this answer. A better formulation would be that a machine can react to red color in different ways. Yes, it can but the machine is not able to reflect about the stimulus. There is nothing behind.

"a machine can mean red (in this sense) without having qualitative experience"

Why not build a machine that has qualitative experience?

Then we could dispense with such qualifiers as "in this sense."

Brian,

thank you for your constructive contributions.  I hope for your cooperation.

Wilfried,

Thanks for your gracious reply.

Wilfried,

thank you for your replies! I am afraid I still do not understand why object recognition should be equivalent to subjective qualitative experience. You say that a machine cannot reflect about a stimulus. But if by reflection you mean a phenomenon which involves qualia, the argument seems circular. In short, why should a machine which cannot have subjective qualitative experience not recognize objects? Surely you would agree that a sufficiently complex machine could (1) show all behavior associated with human object recognition and (2) be outfitted with all non-qualitative cognitive abilities (such as processing information and drawing logical inferences). Why should this not suffice for the machine to be able to recognize objects?

Joachim,

let me try to answer step by step.

My initial question was a question. I do not have an answer but I like to discuss the problem for clarification.

We have algorithms which are able to identify objects in images like tensor flow. This is an excellent ability. My first association was: it is like a quale if we identify things in our visual perception because we do not see colored shapes but the whole object like the machine.

" In short, why should a machine which cannot have subjective qualitative experience not recognize objects? ".

I confirm a machine is able to identify objects but I doubt that recognition is the same as identification. Recognition may have some aspects which are not present in the identification process. Naturally the machine is now able to associate some attributes to the identified object. It is also able to compute a three dimensional representation of all identified objects to model a scene of the surrounding. Recognition is more than that ...

" I think if a machine can token a symbol in cases of red, and not token it in cases of non-red, it can (symbolically/linguistically) "mean" (i.e. refer to) red. "

For recognition we need meaning and this is a really hard nut for a machine. Meaning is not only the identification of a category like "red" which is the result of a color sensor output. Meaning is not identification but meaning is the relation of the object to the perceiving subject. This is a much deeper sense of the expression.

In a human brain and surely in the brain of mammalian, birds and some other animals we can find such a recognition process combined with a lot of associations and bodily reactions like feelings, moods or even emotions. There is the object represented in a dynamical network of stored events, associations, subjective experiences and so on.

I think we can find qualia in the simultaneous conjuncture of the object identification, spacial representation and the qualification by meaning. The results of the three computational activities may also be accelerated by feedback in each step like a resonance process.

In short: I would answer my initial question with "no".

Wilfried,

perhaps I was a bit unclear. My problem is that I have a hard time finding convincing reasons for answering your question in the positive.

I agree that in many organisms, predominantly humans, recognition is tied to qualitative experience. But I see no reason to assume that we cannot have object recognition without this experience (especially in artificial systems).

I also agree that "meaning is a hard nut for a machine". Certainly, a machine tool which serves to identify objects need not be able to mean anything by it in a stronger sense. Still, I do not see why we should assume that any being (organism or machine, natural or artificial) should necessarily have capabilities for qualitative meaning in order to recognize objects.

Surely, to some degree this is a matter not just of investigating what usually is associated with processes of object recognition, but of plausibly defining "object recognition" and "meaning" in the first place. In this sense, my point is this: Equating object recognition and qualitative experience would require revisionist senses of either "object recognition" or "meaning" or both.

Joachim,

"... I see no reason to assume that we cannot have object recognition without this experience ..."

There is no doubt we can have object identification (I prefer this expression) without any subjective experience by a machine. The data generated by such a process can be used for reactions of an automaton. This may be enough for many situations in industrial purposes or assistance systems in cars.

Higher biological organisms do not have qualia for fun but for the reason of higher fitness in the struggle to survive. If a machine is in the situation of an organism it would have some trouble with the visual image processing. The process of object identification generates a long list of objects which have to be qualified one by one. If there is one dangerous object it may happen that this procedure endures too long.

The massive parallel processing brain recognizes the objects and in this recognition the qualification is enclosed by qualia.

Wilfried,

I agree that we should assume that the ability to have qualitative experiences is evolutionarily selected for. But the best evidence for this might just be that every systematically inheritable trait is selected for in one way or another (or at least a byproduct of such selection). What we lack is an explanation of how exactly qualitative experience is adaptive. But even if we had one, this would still not imply that every function which is adaptive by way of involving qualitative experience necessarily has to involve qualitative experience in any possible implementation (e.g. when implemented artificially). And therefore, the function in question (such as object recognition) cannot be equated with qualia.

Apart from these points, I am not sure I follow your latest reply. I certainly agree that natural organisms are equipped with 'fast' object recognition which involves affective states (which in turn involve qualitative experience) instead of merely 'rational' ones (e.g. conscious logical inference). But then the question turns on: Why do the structures which implement 'fast' affective states necessarily involve qualitative experience? Surely, a kind of processing that is analogous to neural affective processing could be built and/or simulated in a non-qualitative way? But that's just Chalmer's "hard problem" restated: Why does any process, if it can proceed in a purely functional way, also have to be accompanied by qualitative experience? I think it does not have to be, even though it certainly often is (e.g. in humans).

Joachim,

" Why does any process, if it can proceed in a purely functional way, also have to be accompanied by qualitative experience? " ist a really interesting question but I deny that it can proceed purely functional. Rational decisions need too much time in the computing step by step.

A sensory stimulus existentially concern the observer. There is no time to reflect pro and contra in dangers situations. A qualitative summary analysis has to take place instead, with possible responses already prepared.

The perception of the situation and the possible reaction have an existential self-reference. Here is a reason for the intensity of perception. There is no functional way possible.

Wilfried,

why do you associate functional processes with speed? I am not aware of any argument to the effect that fast cognitive processes cannot be functionally described. The usual fast/slow theories (in Kahnemann's wake) assume that fast processes involve less "conscious" reflection than fast ones, and that fast ones are more "hardwired" into the brain. But why should this exclude functional descriptions.

In fact - and this is a reason for why there is a hard problem of consciousness after all - we can assume that for everything that goes on in the brain at all there is a functional description. It's just interactions of neurons and the like -- very complex interactions, of course; but why shouldn't we be able to functionally describe some specific processes occuring in the brain?

Hence, I assume that there is a functional description of object recognition in terms of some brain processes, and if there is, then we need not (necessarily) invoke qualia.

qualia is not the result of a conscious reflection. So I do not understand your first reply in your last post. Speed is therefore a quite normal result. I don't deny the qualitative character but the logical computation of the summary scenario.

You are right for the functional description of object identification without qualia. An AI program does not require qualia for algorithmic decisions. This may be a solution for simple purposes like look-and-pick problems in the industrial process. To drive a car autonomously

is not such a simple process so we need better solutions.

Wilfried,

you claimed that it is the speed of the recognition process which requires qualia. That is why I replied that speed of cognitive processes is usually associated with the distinction between conscious reflective thought (slow) and hardwired, intuitive, affective responses (quick). So, I'm not claiming that "qualia is the result of a conscious reflection", but rather that I do not yet see the necessity for thinking quick processes to include qualia.

Your second paragraph in your last reply insinuates that qualia help with the complexity problem. But more complex problems simply require more complex functions, I believe. Why should they require qualia?

sorry,

the auto correction played me in the last post a few small pranks ...

Joachim,

you wrote:" ... I do not yet see the necessity for thinking quick processes to include qualia. " and "... more complex problems simply require more complex functions, I believe. Why should they require qualia? "

I think both questions have one answer: qualia.

The complexity problem is a hard problem for computers. They have a long list with identified objects which have to be calculated for evaluation. If you have a quale for each object it is already evaluated and a reaction is instantaneously possible.

Wilfried,

even if that were true, it would only mean that qualia make object recognition somehow easier or more efficient. It would not mean that object recognition and qualia are somehow equal.

Hence, I do not think that qualia are necessary for object recognition.

However, I also doubt that qualia are sufficient for object recognition. Imagine feeling an intense heat at your right finger tip. Undoubtedly, that is a quale; but do we thereby recognize the object causing the quale? Is it fire, a hot stove, an electric stimulus? It may be all sorts of things, and hence not suffice for object recognition.

Joachim,

I agree in all points you wrote in your last post.

May be we have different intentions: I am looking for understanding qualia and you are looking for effective object identification in machines. That doesn't matter. Such discussions will bring clearness and effort on both sides.

My initial question was a spontaneous idea. Qualia have object identification as part of the recognition process because we do not see colored shapes but objects with colored surface. In this discussion, it became clear that this is not enough, but more aspects have to be added to build a quale.

Your are right in machine recognition. Computers are very quick and list processing is typical for them because there are very effective algorithms for.

On the other side are there biological systems with extremely effective brains which have qualia and consciousness. I think a quale is part of consciousness and if we do understand qualia we do understand consciousness too. The existence of subjective experiences is a challenge I like to understand.

Your example with the feeling of heat at the finger tip is not so simple as it looks like. My finger is part of my body and I have some information where my finger is in space. If I feel something like heat and I know where the danger is. I also know what kind of heat it may be (with closed eyes) because my finger has not only temperature sensors. If I feel a hard object, then this object is hot and if I only feel heat, it's probably fire. Naturally if I have a look to the location where I feel something hot I can see what it is. This is a reaction to update my data set of the environment. The reaction is initialized by an emotion (surprise) which caused my attention to focus the location I feel the unexpected heat.

Object recognition is very complex and we realize not only the object but also its properties and we are able to anticipate effects or intentions if the object is moving. This is very much to calculate and time is not available as much as you like. I think a quale instantiate an object with all its aspects in our mental representation of our surrounding so we can handle the object adequatiously to the situation we are in. A quale is highlighting an object for conscious decisions .

Wilfried,

thank you very much for the clarification! I think I understand your approach better now. Replying to your original question, and assuming your point of view, I think object recognition may be supported by qualia, but it is not "equal".

Further, I think it depends on a lot of additional research to really say whether object recognition is in fact supported by qualitative experience - even though I concede that your hypothesis, given your further explanations, does sound plausible.

Also, we have seen through our discussion that a lot of effort must go into making clear what exactly we mean by "qualia" in such cases. Some clarification regarding their role (and the associated role of meaning) in object recognition, can - at least to some degree - proceed independently of empirical research.

Further, even if said hypothesis is true, object recognition will depend on BOTH qualitative and non-qualitative cognitive processes. Teasing these processes apart will also play an important role in such research.

Joachim,

thank you very much for your kind reply. Our discussion was fruitful and I am honored for the received approval . The question was answerd.

I look forward to future discussions.

It was my pleasure, Wilfried. Thank you for the stimulating discussion!

No. A machine can detect without any subjective experience. To have a quale you must be able to feel something.

Karl,

may be we use different expressions. A quale is IMHO my subjective experience of an object with my sensors but there is no feeling at all.

Feelings will occur if the object has emotional relations I remember or it will be surprising or dangerous etc. I do not think that feelings are necessary to a quale therefore I like to soften your hard "no".

A quale is related to the identification of an object like the identification in deep learning algorithms with ANN. I agree, this kind of identification is not the same as our quale but it is quite near of it. A quale has meaning, not feelings. Meaning is a weak expression and it is not part of the identification process in machines but in living creatures.

Wilfried,

your notion seems to be at odds with the standard definition of qualia, which says that a quale is NOTHING BUT a phenomenal experience or feeling (https://plato.stanford.edu/entries/qualia/).

Joachim,

thank you for your clarification but is it really a clarification?

Maybe you complain because the word 'phenomenal' in the expression 'subjective experience' was missing. Excuse me, you are right.

What does it mean "phenomenal experience"? This expression has no relation to any process. If it would have we should not discuss here. The present discussion has the aim to make cloudy expressions transparent for understanding complex processes of the mind and 'phenomenal experience' (quale) is not understood yet.

@wilfried, there is something hugely important you indeed I agree are recognizing even though the words would need to be adjusted, something others have not. "Emotional" feelings are very different from a red quale or a sound quale and have something they don't. Not only that, the difference is the key to understanding the whole consciousness mystery. I think you will like my papers.

Wilfried,

perhaps you mean something different by "feeling"? Because, while I do think that emotions come with feelings, a feeling can occur independently of an emotion, don't you think? Such as a feeling of warmth etc.? Of course, "feeling" is a notoriously fuzzy concept... much fuzzier than qualia, I would think. What adds to my condusion is your way of explaining qualia by using the term "meaning"... to me, mental meanings are representations or contents, i.e. functional things, whereas qualia are completely independent of function. Or maybe that is what you mean by "phenomenal experience has no relation to any process"? Imagine the taste of chocolate, which, to me, is the paradigmatic quale. It has no meaning, but pure, genuine, qualitative feeling.

Joachim,

I like chocolate too ... (smile)

You wrote the " quale (of chocolate) ... has no meaning, but pure, genuine, qualitative feeling. " I have some problems with his simple sentence. Maybe because of the complexity of chocolate taste.

If I open my eyes I am seeing not colored shapes but things I know: the neighbors red car on the street, a brass mortar on my desk, my blue pencil case or the colored curtain beside the window and so on. The curtain and my pencil case are good examples and there is no feeling at all. I do not mean "emotion"! What feeling do you have looking to a gravel stone on the road? This is why I think not a feeling is part of a quale but meaning. I am aware of the meaning of the item. In case the object has some closer relations to me I realize feelings too because I remember situations in the past with smart bodily reactions.

Chocolate evokes pleasant feelings. It is sweet, but not too sweet, creamy and supple, slightly fat and aromatic. Lots of features that signal the high nutritional value and create a gentle pleasure and I know: It is chocolate. This mental item "chocolate" is linked with many associations which are activated in the moment I realize what it is. This represents the meaning of chocolate and in this moment I am aware of chocolate.

Karls assumption is that a quale evoke one or more feelings. This is a good idea. My proposal is the addition of meaning (in relation to the subject) which trigger feelings but we are not aware of the items because of the feelings but of the meaning. The curtain or a gravel stone does not have significant relations to me and therefore I do not have feelings at all.

Perhaps I misinterpret this quale but chocolate is your favorite example so you are expert.

Wilfried,

I generally agree with you, which is why I think the only disagreement is in what we take "qualia" to mean.

For instance, when you say "If I open my eyes I am seeing not colored shapes but things I know", I would say you're completely right; but what role do quales play in this? They are the experience you have when perceiving, nothing more - whether the perceived things are "shapes" or "things you know" does not seem to matter, right? So, just as you correctly point out that perception need not involve emotion, perception need not involve qualia either. (Notwithstanding the fact that emotions and feelings are not the same.)

The judgment "this is chocolate" and "I am experiencing a chocolate-quale" seem similarly independent to me - they are contingently linked, because we rarely, if ever, experience a chocolate-sensation when not eating chocolate; but they must not be conflated. Hence, meanings are independent (and only ever contingently linked) with quales.

You might of course play a terminological trick and speak of something like "phenomenal meaning", which would either point to the phenomenal content ("experiencing so-and-so") or perhaps even the thing a quale is contingently linked to. But, again, that would be a trick; when you speak of meaning, I take you to mean some kind of mental content which is individuated by genuine objects (such as "chocolate"), not by quale ("experiencing chocolate-eating").

Joachim,

I see...

'They ask: what is this what philosophers called 'qualitative states'? And I answer, only half joking: As Louis Armstrong said when asked what jazz is: if you have to ask first, you'll never understand. '

- Ned Block: Troubles with Functionalism

Wilfried,

very apt ;-)

Frohe Ostern

Happy Easter to all

Understanding jazz I have to develop a feeling for this kind of music. Karl Sipfle will like the word "feeling". This kind of understanding is different to an understanding which is suitable to construct understanding.