Read *The Cognitive Brain* (MIT Press 1991).

hi

you need to look at the whole picture.

Our brain is a complex structure and its research is an a hot topic at present.

There are several issues to that.

First, look at the structure of the brain - look at the neurons structure- look at the connections between them.

Second, look at the evolutionary aspect of the brain - at present human brain is not the same as the previous homo sapiens species.

Third, The senses to which the brain is connected for collecting the informations.

Fourth, the way the brain handles the numerous amount of information- the way it processes it= For ex: the unnecessary past information it forgets automatically . Is it not true?

Like wise there are other several issues to the brain.

What tools we have at hand now is not sufficient. Still research is going on in the same direction.

Hope this answer helps you to some extent.

Thank you.

Best,

INDRAJIT MANDAL

The fundamental question about our cognition is this: How can we possibly represent the world around us when we have no sensory transducers that can detect the volumetric space in which we live?

Joachim: "As tuples of symbols. A symbol is a set of bits or trits or

whatever base you like, with the ability to address

anything thinkable."

This just seems to evade the problem of how our brain is able to create the referent -- the thought of being in the center of a surrounding volumetric space -- that we are able to represent/address by a symbolic system of tuples.

I am quite intrigued by the discussion so far and it is all very interesting. However, when reading the question first I had the nagging urge to ask: "What do you mean by cognition?" But the part of the question ... of how cognition works ... already seemed to imply a process-directed how-question. That somehow finds its reflection in the answers provided so far - they mostly seem to focus upon a mere representational picture, not taking into account the possibility that we as humans create the world in which we find ourselves.

But such a phenomenological perspective might not be what Devis is looking for?

Tom: "... they mostly seem to focus upon a mere representational picture, not taking into account the possibility that we as humans create the world in which we find ourselves."

Yes. For a treatment of this view see "Space, self, and the theater of consciousness" and "Where Am I? Redux" on my RG page.

What is life?

What is qualia?

Another question might be, "What would a 200cm3 larger brain feel like? A loss of about 40,000 generations of evolutionary increase in ~2000 generations or less and still counting is written in the fragment that remains of the original Homo genepool. And written in the fossilized stone of our own genepool's products, our skulls. What then might cognition actually be if our psychopathic form of Hominin, the last Homo standing, has suffered so much loss in the deviant evolutionary path our line has taken? We have been and are currently selecting out the more 'intelligent' of our members, generation by generation, because they are the least enslaved to the extant collective delusion of the time and place and either appear insane or as prophets of some form of wisdom and, in either case, the psychopathic rulers of the group suppress all deviance from Authority and/or threat to their ability to parasitize their group, typically with methods which remove not only the capacious individual's collection of alleles but relatives' as well, ensuring that positive alleles are thoroughly selected out of future versions of our genepool over hundreds of generations. It is observable today, and drones make it even more thorough, preferentially removing the 'smart' people and their families without taking out a chunk of more 'normal' alleles as well. Drones seem almost like an 'evolutionary enzyme' which has appeared to speed up our idiotification.

But, the point made by Dr. Wes Raikowski reflects another thinker by the name of Kurt Gödel in his incompleteness theorems which point out that a system cannot be completely understood from within itself. As this goes for mathematics, it would also seem to go for us. And, it is in a Universe described by what Dr Raikowshi calls "physics" that one can stand and view us from the outside. We have reached a point in our accretion of consistent properties of our Universe where we may see ourselves completely, our limits and our importance, as we must be, and see us as an interstellar visitor might view us, obligate communal cannibalistic monkeys which parasitize our own collective genepool, driven by a species specific universal hierarchical group structure of a small psychopathic parasitic clique, a larger buffer layer of psychopath sycophants who manage and directly feed upon a much larger herd glued together with altruism, empathy, sympathy, cooperation, and passing the best bits from the herd on to the pure power clique. What is the single collective external behavior of any extant human identity group? Trying to eat the group next to it as motivated by its psychopathic clique of parasites.

But, seeing this way requires a knowledge of our universe which we now have accreted but is broad and can be difficult to individually acquire, at least motivationally. It's the only way, though, with which we may see beyond our developmentally implanted common collective delusion (CCD), our Homosap 'reality'. The downside is that within our perceptual CCD, Homosap is so much more significant and 'advanced' than in the view from the Universe. We cannot think small enough to really describe our place in this Universe nor large enough to imagine complexity much greater than ourselves even though it already exists within each and every one of our own cells. Saying this, we are smaller than we think, was why they burned Giordano and trying to bring this physical perception of us into our own personal version of the CCD, our own identity, invites extreme cognitive dissonance and a strong desire to not think that way, a.k.a. violent denial. But the joy of physical reality is that we must all agree with it whether we want to or not. And, for the religiously minded, albeit The Book insists it is G-d, there can be no question that The Work is, in fact, your G-d and should be studied as a direct expression of the Divinity which you seek. Another joy of the physical worldview is that, within it, we may actually have a 'purpose'.

Within the above view, there is a straightforward description of cognition which removes much of its 'mystery' and points out its inevitability and much of its function, albeit, not simple enough in its underpinnings to be succinctly described here.

There are a lot of good and high leveled questions but where are the answers ...?

As I have already raised the question about the 'assumed' universal validity of representational accounts in relation to constitutive accounts. I would like to point to the rather pre-reflective way in which 'information' is used explain what it does this or that within the various attempts to explain consciousness.

A random set of bars, as they are to be found on my workbench at home may not have any information-value for anyone, but for me they indicate how often I used a specific tool and when it would be time to service it. - Stupid example, but it may just serve to highlight the point I wish to make. Signs of any kind are just that - or not even that - as long as no 'sense' or 'meaning' is attached to them. This sense or meaning is what makes signs to become information for someone who can read the signs and interpret them within a horizon of sense, shared with others.

In this respect I do miss a bit more focus in the discussion and some acknowledgement of the fundamental difference between computational information processing and human brain activity to create sense or meaning for the owner of the brain. If the sense-providing horizon is not to be located within the brain - or at least dependent upon a cultural background against which it is individually acquired - then the chances of a purely naturalistic explanation seem bleak to me. The only thing which could be revealed by naturalistic accounts would be the physical conditions of the possibility for sense-processing, but the emergence of sense and it's individual significance for the one undergoing the relevant states seems to be beyond the reach of naturalistic models.

Ever better accounts of how information (sense) is processed, once it had been constituted, may reveal more and more of the physical structures, but seem to neglect the underlying question of how consciousness brings it about that some aspects of my environment appear to me, that they have significance and importance to me?

I like to hear your thoughts on this - and I have to admit, even this critical entry does not provide the answers Wilfried was looking for.

"How is information encoded within the mind (in the brain)?"

It is not "encoded", i.e, there is no code, no symbol standing for the things around us. So I would say the main challenge in cognition is explaining it with arepresentational models, instead of proposing fancy chit-chat descriptive models of this and that (which in  fact explain nothing).

I might be wrong but this is how I see it.

Serban, 

I agree, there is no code. Each brain is unique and has its own "code" so it is hard to find a common code translation. Nevertheless there is a representation of environmental objects, so we should ask what kind of representation is necessary to work like our brain works to create a mind.

@Wilfried: Please note I'm quite passionate on what I say below, so please do not take it personally :)

"Nevertheless there is a representation of environmental objects" - nope, there is no need for such a thing. Positing representations is handy, it "explains" things the same way positing Zeus "explained" thunders. We are left with the same phenomena to explain, but by a sleight of hand we introduced the concept of representation (or the choleric god Zeus) and from there on all seems to be explained (but is not). Now, look again, after introducing representations, nothing is explained. And I mean explain as in coming up with the mechanism whereby a cognitive activity is carried out (on, this, see Searle, 1990, ref. below). Representations do not and cannot explain anything. This however is not an interesting point to make, because it cannot be demonstrated. Note that the belief that representations are useful and explain cognition is just as much impossible to demonstrate. Without being pedantic, I suggest anybody who believe that representations are the key to cognition should read at least the first chapter of Chemero (2011) -- see ref. below. (My bet is once you've read the first chapter you'll carry on reading and down that road learn how non-representational cognition may work -- there are some really nifty examples in that book!).

"...so we should ask what kind of representation is necessary to work like our brain works to create a mind" Our brain does not create our mind. Firstly, the brain without the body it is attached to and which interacts with the environment has no sense at all. Secondly, a body-brain unit that does not interact with the environment has no mind; a body-brain unit that *does* interact with the environment is a body-brain-mind unit. Or to put it in other words, there is nothing more to the mind than a body-brain unit interacting with its natural environment.

References cited above:

"Radical Embodied Cognitive Science", a book by Anthony Chemero (MIT Press, 2011)

"Consciousness, explanatory inversion and cognitive science", a 1990 Behavioral and Brain Sciences paper (BBS 13 (1):585-642)

@Wes

I disagree with two things you say, but I disagree really a lot on those two matters. Let me explain myself. When you start out with "Take an apple as an example of a representation", either you are talking of the idea of an apple (thus needlessly reifying a perception into an idea) or you are talking of a real apple. The former is not necessarily a representation, the latter is clearly not a representation.

The second disagreement is about what the segregation between "sense spaces" (as you call them) entails. Sense spaces can be conceived as the different parts of the input layer of an artificial neural network (henceforth ANN), the fact that they are segregated bears no consequences on what happens at the next, "upper" level, where those inputs are integrated. What matters is how these inputs are "put together" in other parts of the brain, which would correspond (rather loosely, it is only a shorthand for my argument) to the hidden layer of an ANN. The nonlinearities that occur at that higher level makes it possible to end up with a space that is completely different from that of the input space(s) and, importantly, that also integrates (possibly in a nonlinear way) all the sensory inputs from the different senses. I'm no neuroscientist but such integration areas in the brain are documented.

As to "The ability to sense selectively makes representations in the brain feasible but only if they are defined as spatial relations between activated sensory cells of a particular modality.", this sounds as gibberish to me. I don't understand why there is an "only" in the phrase, and above all I couldn't care less of the conditions necessary to representations, since I think representations are useless in a theory of how the brain causes intelligent behavior (or cognition, if you like).

Again, not that my view is necessarily correct, but I don't see how your arguments undermine it...

Cheers,

SCM

Starting with arepresentational models of cognition: the book by A. Chemero I cited in my previous post is about those models. I cannot say it better than Chemero does, and I clearly cannot summarize about a hundred pages in a couple of paragraphs, so if you are interested in finding out more about a-representational models get that book and do read it.

As to my modeling with ANN, the thing that is important is the fact that the ANNs I use behave as nonlinear systems. The fact that is electrons flow through my computer of for what matters toothpaste flow or whatever would make up a nonlinear  system is irrelevant to my simulation purpose. Not that I think it is irrelevant to human cognition, I know how information is passed from a neuron to another, that is via neurotransmitters (and not via toothpaste flow) but my modelling is not at that level, it is rather at the level of the system. And as far as from my modeling I get predictions that nobody has thought before (and I would not have either without this precise type of modeling) and can then check these predictions in behavioral experiments, there is at least a heuristic value you cannot deny and of which I am happy. You can check this in this paper

https://www.researchgate.net/publication/215527067_The_role_of_cue_information_in_the_outcome-density_effect_Evidence_from_neural_network_simulations_and_a_causal_learning_experiment?ev=prf_pub

and in this one

https://www.researchgate.net/publication/220233644_Self-refreshing_memory_in_artificial_neural_networks_learning_temporal_sequences_without_catastrophic_forgetting?ev=prf_pub

About the second paper, and above all if you have a strong background in psychology of memory/learning, tell me sincerely if you would have predicted the behavioral results depicted in Fig. 9.

Anyways, not sure I answered your questions. Let me know.

Cheers,

SCM

Article The role of cue information in the outcome-density effect: E...

Article Self-refreshing memory in artificial neural networks: Learni...

@Wes: Thank your for the nice things you say on my papers.

As a matter of fact, I agree that "its [ANN] operation cannot be equated with the operational principles of the brain". However, my view on ANNs (and my work this far) is that they can be used to i) derive ideas on how the brain works, because both the brain and ANNs are (for ANNs, on some conditions, i.e., not all ANNs are) complex nonlinear systems, and ii) to put constraints on the theory we make about the brain or, more modestly, about this and that cognitive phenomena found in humans. On the contrary, starting from the idea that there are representations in the brain and representations are cause-effective not only poses no constraint on a theory of cognition, but allow for "explaining" everything (and it contrary).

Now, as you seem to ask for simulation tools (say, ANN) that work as the brain does (i.e., that are related to neurotransmitters and the like), I would like to point out that there is a formal mathematical equivalence between an algorithm used to train the ANNs I have been using and CHL (contrastive hebbian learning) algorithm, which is closely based on what is known about LTP (long-term potentiation), and, crucially, depends only on the pre- and post-synaptic activity available locally. The reference is

Xie, X., & Seung, H. S. (2003). Equivalence of backpropagation and contrastive hebbian learning in a layered network. Neural Computation, 15, 441-454.

Cheers,

SCM

I made my decision (as to the key process) quite a while ago: inductive processes. Helmholtz and other very distinguished scientists made the same decision. Moreover, I don't see any other candidate process that can compete with the inductive process in that respect.

However, it goes without saying that all possible candidates should be very carefully considered.