Hi Velina!

I've just posted a reply to Vladimir Farber that, to my understanding, has basic relevance to your question.

Following from it, there's the implication that more clearly-defined representation of the individual episodes of a continuous sequence should increase a subject's ability to recall the representation of each episode in the form of a deliberately prescribed chunk (i.e. a chunk consisting of a predeternined number of episodes, and perhaps paced in time). With mental rehearsal,  the chunked low-level individual representations would entail the forming of a high-level cognitive representation that included the chunk's additional, deliberately imposed parameters, allowing automatized top-down retrieval of the entire set on demand.

Hope this is of some help - please let me know if it needs clarifying or elaborating.

Hi Richard,

Thank you for being the first that provides an answer to this question. I agree with what you say: "the chunked low-level individual representations would entail the forming of a high-level cognitive representation that included the chunk's additional, deliberately imposed parameters, allowing automatized top-down retrieval of the entire set on demand." Yes, of course, I agree.

Your point seems also being related to episodic memory no? I think that this is the same with motor memory (I guess you deal with this too when looking to your profile).

My interest is mostly concentrated to the interaction between the Top-Down and the Bottom-up, which interaction creates let's say  "a meaningful" picture of the perceived. I mean, before the what ever kind of "semy-treated" up-going units reach the state of  "being *understood", they are compared with the already existing Top knowledge of the corresponding type. So, I guess that if one doesn't have a model of how should Am sound like, she/he could possibly capture that these are sounds in minor, but not more. That would however mean that she/he has already a model for how does a minor sound. That Top-down conduction of the existing model to the Working memory treatment and it's comparison with the conduced Up signal, gives as result "understanding". Of course, Learning is based very strongly on this, as the perceived  features that are compared with the existing models (and their entire structure in the  knowledge space) will little by little form the concept of Am precisely (a model that, of course, all music professional have).

This is in fact a general scheme of what seems to happen. But my concern is how to model it in more formal terms... and namely - the interaction between the conceptual knowledge related lo language units which are mapped to the perceived phonological stream. As the Top-Down seems to play a decisive role in speech recognition and in learning, including the formation of novel concepts.  There are chunks, yes, during the processing. For example one chunks the phonological content that has been captured (and partially guessed) into meaningful units on a higher level of representation, for example a word-form. That happens quickly and is not a process that we are aware of.  I guess that it uses WM resources.

I have tried to follow what have you posted posted as reply to Vladimir Farber, I saw the questions that he asks, but I don's see you r post there. Is that because of the system or is that because you have mailed to Vladimir Farber your comment?

So, excuse me if my reply is not taking into account the main reference of your reasoning, but I think that however you understood better my point and you will correct me, discuss further et cetera.

Velina

Here is a link to my reply to Vladimir Farber.

My usage of the word "episodic" refers to any kind of event that an observer perceives, including the experienced consequence of a self-initiated, goal-seeking motor act., such as a certain musical chord and its attended to timbral qualities. So yes, my answer is relevant to motor learning (especially to TOTE models)

https://www.researchgate.net/post/Is_there_a_relationship_between_our_memory_level_and_between_the_level_of_how_we_organized

Thank you, Richard, I saw the question/answer.

You refer to the Miler's magic number 7 and to his basic three-component model that I also try to use.

Hi Velina,

To answer your original question, may I suggest looking into the work of William T. Powers and his protegés who developed Perceptual Control Theory. If you are not already acquainted with that, PCT is directly concerned with cybernetic modelling of living systems, and particularly with hierarchical control of behaviour. Decision trees, such as those used to represent troubleshooting strategies, are especially well suited for such modelling. For an introductory synopsis of PCT, by Powers and his associates, go to this page: http://www.pctweb.org/PCTUnderstanding.pdf,There's an excellent, freely downloadable book of readings on aspects, concerns and methodology of PCT available from www.livingcontrolsystems.com. Essential  reading also is Miller, Galanter & Pribram's book, Plans and the Structure of Behavior (1960), in which the TOTE model was originally described and developed, and  which Power's work continued. Go to https://archive.org/details/plansstructureof00mill to read or download it.

Hi Richard,

Yes, right, Perceptual Control Theory is one very good step to do something that resembles a bit more human behaiviour, yes! Thank you. Yes, aware of this work.

I have not read this Miller, Galanter & Pribram's book! 

I just took a look now and downloaded. It concerns mostly behavior yes...plans, motor skills ooh, even instincts! Instincts are of course a source of an intelligent behavior, but one has to explain that to AI specialists and to philosophers... Excellent!  I will read this all very attentively, it seems to me that could give me some ideas about from where to start a model of initial conceptual knowledge.

Thank you!

My initial question stays.

It is related to the calculus, to the procedure that runs on more micro-level. The "understanding"  procedure....  top-down bottom up  "getting meaning" procedure. 

If we processed at time the entire set of stimuli that reaches our sensation, the perception would brake down, to leave us in a sort of a Very Dark Chinese room. So, the W.Memo bottom-up top-down  "calculus"" is done on only what is "important". And this is, according to me, a key mechanism in human perception - conceptualization and even thinking.

"If we processed at time the entire set of stimuli that reaches our sensation, the perception would brake down, to leave us in a sort of a Very Dark Chinese room. So, the W.Memo bottom-up top-down "calculus"" is done on only what is "important". And this is, according to me, a key mechanism in human perception - conceptualization and even thinking."

Absolutely so! Cognitive representations link situational contexts with goals, and more particularly, with information singled out from overall perceptual data as values of the qualitative variables that, for a particular individual, define its contextually related goal and subgoals. Cognitive representations, which may be situationally specific or generic, thereby drive selective search and attention, optimizing activity in sub-perceptual neural assemblies that are specialized for processing the goal-relevant informational modalities for any  presented situation.

Thinking back to theoretical work that might inform your project, possibly the model of memory representation and retrieval architecture proposed by Damasio in Ch. 6 of his Self Comes to Mind might also be useful to you. It has considerable explanatory power, I think.

Best regards,

Richard.

Damasio! Fascinating! Yes! And I agree with everything that is said there, in Ch. 6 of his Self Comes to Mind.

THANK YOU, RICHARD! I'll of course read from the beginning.

Best Regards,

Velina

Velina, and Richard,

I count myself only as a relatively well, but broadly, informed dilettante in your field. My question - which I often have to ask disciplinary specialists - is about the dangers of operational metaphors and axiomatic assumptions. I believe I know the origins of the bottom up and top down terminology. I am dubious about metaphors applied to brain/mind. Metaphors, generally, are useful until they are not. But the point at which they become not-useful - and take researchers down what transpire to be blind alleys - is usually unclear. Until there is some kind of Kuhnian shift. And then it all seems obvious.

When such metaphors become institutionalised - when grants and degrees and departments become structured around them - it can be very difficult for an embedded researcher to even conceive of questions which are impossible to ask within the disciplinary - or subdisciplinary -paradigm.

I mean no disrespect to you, I know from direct experience as well as literature, that such questioning can appear incomprehensible, or offensive. But even if researchers feel that they are pursuing objective knowledge through experimentation, all research programs (I think) rest on some theory, some set of axioms, which by definition are untestable. So at the root of this fact finding is belief. I suppose this has to be, but I think it is a reality we all ought to be vigilant about, and so must continually try to ask 'out of the box' questions, in order to test those underlying assumptions.

Simon Penny

Simon, thank you a lot. Yes, good advice.

Well... Let's say that in my ignorance I have no any paradigm in the head.  So, metaphorically speaking, a don't have any strong axiomatic base.

Do you think that the information flows and their processing are already well closed and even locked in the box?

Velina

I would say " information flows and their processing " constitutes a paradigm :) A computationalist paradigm. I am cautious (as Shannon was) of the mis-application of his theory. The notion of 'flow' implies directionality. And 'processing' is an industrial if not computational metaphor. I'm not comfortable about the value of any of them with respect to brain function, or any biological function.

Yes, of course... I was limited without being much aware of that.

But what ever terms I could use it should be a limitation. So, have we to mute? Do you propose something?

My advice is: 

1. to try to be aware when a metaphor becomes integral to an explanation and to scrutinise it.

2. try to be aware of axiomatic assumptions which underly methodologies, disciplinary formations etc. Engaging in interdisciplinary interactions is one way to reveal these.

None of this is easy of course. Discovering one's own implicit assumptions, or those of one's discipline, is difficult intellectual work. It is much easier to shelter in an academic silo where colleagues engage in mutual reinforcement of all kinds.

Yes, Simon, thank you. That's what I try to do... interdisciplinary interactions. Apparently without any success.

The metaphor of top-down bottom- up seemed to me basic and large. Or that doesn't fit some Lakoff or others understandings?

The metaphor of top-down bottom- up seemed to me basic and large. 

As it does for me also. As an approach to understanding how, in principle, the brain accomplishes what it does - and (importantly for me) what it could feasibly learn to do - I've found top-down/ bottom up modelling to be immensely productive, elucidating and intellectually liberating for identifying and clarifying the fundamental problems of my field.

While recognizing the importance of re-examining one's axiomatic assumptions in order to assess their correlatedness with demonstrably real, materially embodied entities, processes and systems, the point of that is surely not to rid oneself of all axioms in a Baconian spirit, but to arrive at better, more realistic axioms. Axioms of some kind are essential for suggesting paradigms, which may or may not lead to profiting outcomes. Those outcomes decide the validity and acceptability of axioms, rather than some general prejudicial distrust of axioms in principle - which is intellectual Ludditism. Without paradigms, there can be no orderly research, but only an arbitrary, "anything goes" dabbling of the sort proposed by Feyerabend, Without paradigms, and the axioms that may suggest them, there could be no accounting for progress in science in terms of a Popperian-Kuhnian model. Should we reject that model, too, aware that it is grounded in axiomatic assumptions?

i thoroughly agree that the kinds of axiomatic assumptions that we should challenge are ones that are nothing but metaphors invented for convenient  pseudo-scientific explanation without knowledge of actual, real causal entities and the structures emergent from them. The top-down/ bottom-up paradigm is not founded upon a metaphor; it refers to a self-explanatory characteristic of hierarchically organized systems. In the aftermath of more than a century of advancing neurobiological research, are there any still any reasonable grounds for asserting that the central nervous system is not a hierarchically organized system? As axiomatic assumptions go, it is as good as can be had, surely. Importantly, its hierarchical organization informs us of its structure, and nothing can be modelled without intent to represent how its components are structurally related.

Another argument in defence of pursuing a particular paradigm is that one's intellect, natural predispositions and background knowledge may be especially attuned for exploring its substance - the venture knocks sparks off one, stimulates engagement in difficult complex thinking that. for those so attuned, may prove productive - provided one keeps on. There can be no long-term pursuit of a paradigm and its possibilities without a sense of faith that it will lead to advancement, and that is fuelled and inspired by the sparks resulting from one's compelling curiosity.

Keep on as you have been, Velina, despite the dangers Simon has reminded us of!

Thank you a lot, Richard.

I have not such a high capacity to explain things as you do, thank you.

Richard

thankyou for your lucid notes. My call for vigilance with respect to axiomatic assumptions is just that. It is based on years of experience working with colleges and researchers who, it appears, he never been encouraged to consider the sets of assumptions that define methodologies, experimental programs etc. I also know on several cases historically where such uninterrogated assumptions have driven research in directions which have later transpired to be faulty. I am persuaded by Philip Agre's call for 'Critical Technical Practice' and would recommend that essay - to anyone. I am not quite so derisive of Feyerabend as you are.

"The top-down/ bottom-up paradigm is not founded upon a metaphor; it refers to a self-explanatory characteristic of hierarchically organized systems"

In certain quarters in cognitive neuroscience, the notion of " hierarchically organized systems" is axiomatic. Yet in other areas it is not nearly so. It seems the assumption of the existence ' hierarchically organized systems' is not unlike the assumption of 'representation'. It is assumed to exist, and it is assumed that everyone means the same thing when they use the term. Neither are true in the case of representation and it might not be the case with  notions of heirarchies.

In general I find functionalist and internalist explanations of mind/cognition etc dubious.

Simon, thank you a lot.

I start seeing the problem, or at least a part of it.

Do you suggest to not look at cognitive processes as running within an hierarchically organized system?

If we try to represent the issue in Hegelian form, we should this time start by defining the purpose. The purpose already fixes more or less (according to me) a corresponding level of abstract representation. One can call that a metaphor and point out the "axioms" on it's basis in order to obtain the global scheme of reasoning.  I have tried to determine this in my question -  "formal models, ideas... flows interact... using working memory ...deals with LTM representations...  sensory input ...  let's say "understanding"... learning too"

When the trial is to put together high level cognitive processes such as learning and memory (of several functional types) and sensation as a kind of input, I propose to present that relation via some hierarchy.  That fits the best my personal understandings about what the systemic approach should give as a result - an hierarchy. Perhaps I'm wrong.

Memory is already something that is to be taken on the corresponding level of abstraction and within one of the numerous existing paradigms (I don't propose for example the quantum level... or the one based on sodium-potassium pumps.)

So, can we escape? Perhaps a paradigm shifting is more dangerous?

In any case, I didn't think that there are not different approaches to the same global problem of  "'how does this function". You mention the Mind, yes, why not. But my question was not in formal models of Mind or in models of the neurological level of codding and transferring the signals. Perhaps the results of this models will be join-able one day,  in some points.

Or you mean cognitive processes do not exist?

Velina

in another thread I have become aquainted with the work of Timo Järvilehto. I attach a paper he forwarded to me. I offer it as an exqample of someone who is thinking outside conventional paradigms.

Thank you Simon, I'll read it and come back to your suggestion later.

In any case I don't think that I try to escape the conventional paradigms, but perhaps I'm wrong.

As regards entrenched terminology and assumptions: if there's a culprit, it's very many of us! There's quite an interesting, philosophical analysis by Francken and Slors (2014). From the abstract:

'Commonsense cognitive concepts (CCCs) are the concepts used in daily life to explain, predict and interpret behaviour. CCCs are also used to convey neuroscientific results, not only to wider audiences but also to the scientific inner circle. We show that translations from CCCs to brain activity, and from brain data to CCCs are made in implicit, loose and unsystematic ways. This results in hard to connect data as well as possibly unwarranted extrapolations. We argue that the cause of these problems is a covert adherence to a position known in philosophy of mind as ‘mental realism’. The most fruitful way forward to a clearer and more systematic employment of CCCs in cognitive neuroscience, we argue, is to explicitly adopt interpretivism as an alternative for mental realism. An interpretative stance will help to avoid conceptual confusion in cognitive science and implies caution when it comes to big conclusions about CCCs.'

Article From commonsense to science, and back: The use of cognitive ...