Submited on: 13 Oct 2010 10:17:43 AM GMT
Published on: 31 Oct 2010 12:01:54 AM GMT
 
Development vs. Implementation
Posted by Dr. Sergey Petrov on 23 Nov 2011 12:49:43 AM GMT

1 Is the subject of the article within the scope of the subject category? Yes
2 Are the interpretations / conclusions sound and justified by the data? Yes
3 Is this a new and original contribution? Yes
4 Does this paper exemplify an awareness of other research on the topic? Yes
5 Are structure and length satisfactory? Yes
6 Can you suggest brief additions or amendments or an introductory statement that will increase the value of this paper for an international audience? No
7 Can you suggest any reductions in the paper, or deletions of parts? No
8 Is the quality of the diction satisfactory? Yes
9 Are the illustrations and tables necessary and acceptable? Yes
10 Are the references adequate and are they all necessary? Yes
11 Are the keywords and abstract or summary informative? Yes
  • Other Comments:

    The research underlines known weakenesses of AI approach to knowledge representation and modeling, in  a sense "Implementation Approach" offering a new point of view - or philosphical position - to the issue, kind of "Development Approach". Naturally it creates a huge set of questions, for example, about innner representation of object model per se. Some of them are answered, others require more research - and thinking.  

  • Invited by the author to review this article? :
    No
  • Have you previously published on this or a similar topic?:
    Yes
  • References:
    Petrov S.V. Logical Study of Systems Structures // System Science, 1987. Vol.13. No. 3 4. pp. 107 114. Petrov S.V. Finite Axiomatization of Languages for Representation of System Properties Dependencies Axiomatization // Information Science, 1989. Vol. 47. No. 3. pp. 267 286.
  • Experience and credentials in the specific area of science:

    About 10 years of AI research, including knowledge representation and modeing. 

  • How to cite:  Petrov S .Development vs. Implementation[Review of the article 'Joint Acquisition Of Language And Cognition ' by Perlovsky L].WebmedCentral 2010;2(11):WMCRW001153
1 2 3 4 5 6 7 8 9
Report abuse
 
Untitled
Posted by Dr. Anatoly Temkin on 17 Feb 2011 12:11:14 PM GMT

1 Is the subject of the article within the scope of the subject category? Yes
2 Are the interpretations / conclusions sound and justified by the data? Yes
3 Is this a new and original contribution? Yes
4 Does this paper exemplify an awareness of other research on the topic? Yes
5 Are structure and length satisfactory? Yes
6 Can you suggest brief additions or amendments or an introductory statement that will increase the value of this paper for an international audience? No
7 Can you suggest any reductions in the paper, or deletions of parts? No
8 Is the quality of the diction satisfactory? Yes
9 Are the illustrations and tables necessary and acceptable? Yes
10 Are the references adequate and are they all necessary? Yes
11 Are the keywords and abstract or summary informative? Yes
  • Other Comments:

    The author advances a provocative hypothesis about interface between language and cognition. Little is known about neural mechanisms of interaction between language and cognition. Since Chomsky’s hypothesis that language and cognition are independent, despite incredibility of this statement, little progress was made toward understanding of related mechanisms.

    Why children learn language early in life, can talk about virtually anything, but do not understand the world like adults do? Why no animals possess human type cognition or language separately? Only humans possess these abilities, and possess both of them. Is cognition the same as language or different, and if different, then what exactly is similar and different?

    Do people really understand what they say? If they do, what does it mean? What is the difference between those that think well, and those that talk well?

    All these seem to be mysteries

    Perlovsky proposes that cognition and language are different abilities, he proposes a concrete mechanism of their interaction (dual model), and this single assumption seems to explain all these puzzles about the language and thinking. His suggestion that adults might be much like children, that adults also might be talking without thinking most of the time, that we are different from children only quantitatively – we understand more often than kids what we say, still, not too often – this suggestion seems incredible! Yet, it made us think about ourselves. How often do we understand what are we saying? What might save these incredible statements is a detailed list of experiments proposed by Perlovsky to test his theory, experiments that seem easy to conduct.

  • Invited by the author to review this article? :
    Yes
  • Have you previously published on this or a similar topic?:
    No
  • References:
    None
  • Experience and credentials in the specific area of science:
    None
  • How to cite:  Temkin A .Untitled[Review of the article 'Joint Acquisition Of Language And Cognition ' by Perlovsky L].WebmedCentral 2010;2(2):WMCRW00480
Report abuse
 
Some contributions to understanding big issues
Posted by Dr. Daniel S Levine on 18 Nov 2010 07:50:22 PM GMT

1 Is the subject of the article within the scope of the subject category? Yes
2 Are the interpretations / conclusions sound and justified by the data? Yes
3 Is this a new and original contribution? Yes
4 Does this paper exemplify an awareness of other research on the topic? Yes
5 Are structure and length satisfactory? Yes
6 Can you suggest brief additions or amendments or an introductory statement that will increase the value of this paper for an international audience? Yes
7 Can you suggest any reductions in the paper, or deletions of parts? Yes
8 Is the quality of the diction satisfactory? Yes
9 Are the illustrations and tables necessary and acceptable? Yes
10 Are the references adequate and are they all necessary? Yes
11 Are the keywords and abstract or summary informative? Yes
  • Other Comments:

    Perlovsky has put together a provocative theory that makes major contributions to the dialogue on at least two major long-standing issues in the cognitive sciences.

     

    One of these issues concerns how much our thought is structured by our language.  While not embracing the extreme Whorfian "language of thought" hypothesis, Perlovsky demonstrates how the generic capacity of language is at the foundation of our ability to categorize our mental space in a sophisticated fashion.  In other articles he has discussed implications of this thesis for how differences among spoken languages influence cultural differences, now and at other times in history.

     

    The other issue concerns how much human behavior is rational and how much is primarily driven by either emotion or habit.  While positing the importance of a knowledge instinct that is universal in humans (although some people act on this instinct to a greater degree than others), he avoids the "reason versus emotion" dichotomy that has retarded the progress both of cognitive neuroscience and of society.  He ties the knowledge instinct to a sense of the beautiful or the sublime, in other words to our highest and most generous emotions.  Hence his work contributes to a more nuanced understanding of the interrelations among reason, emotion, and habit.

     

    While much of Perlovsky's theory is abstract, and guided by general mathematical principles, it gives a great promise for driving hypotheses about what to look for in the brain.

  • Competing interests:
    I have published jointly with Dr. Perlovsky and been funded by his branch of the Air Force.
  • Invited by the author to review this article? :
    Yes
  • Have you previously published on this or a similar topic?:
    Yes
  • References:
    None
  • Experience and credentials in the specific area of science:

    I have been doing research in neural networks since 1975.  I am a former President of the International Neural Network Society and author of a textbook on neural and cognitive modeling (Erlbaum, 2nd edition, 2000).

  • How to cite:  Levine D S.Some contributions to understanding big issues[Review of the article 'Joint Acquisition Of Language And Cognition ' by Perlovsky L].WebmedCentral 2010;1(11):WMCRW00147
1 2 3 4 5 6 7 8 9
Report abuse
 
Mathematical Description of Language and Cognition
Posted by Dr. Ross Deming on 05 Nov 2010 04:10:42 PM GMT

  • Other Comments: This article is unique and interesting. Perlovsky offers a mathematical description for the acquisition of language and cognition, how they jointly evolved, and how they interact with each other. The article places elements of psychology and neurological science on firm mathematical ground using simply and elegant equations. In the article the mathematics are described on a summary level, while the references provide the interested reader with a more in-depth development.
  • Competing interests:
    no
  • Invited by the author to review this article? :
    No
  • Have you previously published on this or a similar topic?:
    No
  • References:
    None
  • Experience and credentials in the specific area of science:
    I am a researcher who is active in the areas of machine vision, automatic pattern analysis, and adaptive signal processing.
  • How to cite:  Deming R .Mathematical Description of Language and Cognition[Review of the article 'Joint Acquisition Of Language And Cognition ' by Perlovsky L].WebmedCentral 2010;1(11):WMCRW00118
Report abuse
 
Thank you Ross Best Leonid
Responded by Dr. Leonid Perlovsky on 06 Nov 2010 01:53:46 AM

  • Other Comments: In this most welcome contribution, Dr. Perlovsky shows how his seminal ideas on the mechanisms of the mind, which have been developed over the last 20 years, can be used to address the relation between cognition and language. In fact, the dichotomy between vague and crisp which permeates his Neural Modeling Fields/Dynamic Logic Theory finds in language and cognition a remarkable realization, which was foreseen more than a century ago by Ferdinand de Saussure: ?Without language, thought is a vague, uncharted nebula. There are no pre-existing ideas, and nothing is distinct before the appearance of language.? Today, de Saussure?s insight is being confirmed by functional Magnetic Resonance Imaging (fMRI) and magneto-encephalography (MEG) studies as pointed out in Perlovsky?s paper. Of course, language can be vague, as poetry remind us, and it may not be a coincidence that our senses of beauty and enlightenment are risen when language is used in that vague mode. The paper offers an authoritative overview on the bearings of recent fMRI and MEG studies on the language-cognition issue, and presents a novel mathematical framework aimed at integrating these two faculties which define the human species. It is definitely a unique contribution of its kind which will certainly motivate further research on this critical interdisciplinary research topic. In what follows I draw attention to some secondary points the author may wish to clarify in a revised version of the paper. The author repeatedly asks ?Why animals do not talk and think like people??. The answer is simply because they don?t need it! Their communication systems are sufficient for their needs (actually they evolved to satisfy those needs) and that?s the end of the story. Language is completely useless for any animal but us. In fact, a still unsolved question is what is the selective pressure acting on the early hominids that prompted the appearance of language, i.e., the difficult question and the question one should ask is ?Why do people talk??. Another question posed by the author, ?why animals without human language cannot think like humans?? seems to link language (only humans have language) and human-like cognition. So I wonder whether the language-cognition issue is not an artificial problem, alike the famous ?brain-mind? problem, and language is indissoluble from human cognition. In fact, recent evidences from neuroimaging show that communication and language abilities are dealt with by different parts of the brain [Willems RM and Varley RA (2010). Neural insights into the relation between language and communication. Front. Hum. Neurosci. 4:203. doi: 10.3389/fnhum.2010.00203], which seems to vindicate the claim by the young Chomsky that language is an internal representation system which has nothing to do with communication. That being sad, I see no reason to separate (human) language from (human) cognition. Another point the author should be more cautious is about the claim that the association between words and objects is a combinatorial complex problem. The fact that the number of possible combinations is astronomically large does not imply that the problem cannot be solved easily. In fact, it is a mathematically challenging problem to prove that a given combinatorial problem is hard to solve. Most of them are not and those which are are listed in the 300 pages book by Garey & Johnston. In particular, learning word-object associations using cross-situational learning (Locke?s associationism) can easily be achieved using statistical learning: for the Bayes algorithm the learning time scales only linearly with the number of objects and words (Fontanari, unpublished).
  • Competing interests:
    NO
  • Invited by the author to review this article? :
    Yes
  • Have you previously published on this or a similar topic?:
    Yes
  • References:
    None
  • Experience and credentials in the specific area of science:
    Computational Language Acquisition & Evolution
  • How to cite:  Fontanari J F.The language-cognition issue: a real or artificial problem?[Review of the article 'Joint Acquisition Of Language And Cognition ' by Perlovsky L].WebmedCentral 2010;1(11):WMCRW00112
Report abuse
 
Dear Fernando I modified as suggested Thank you Best I am thankful to Prof. Fontanari for an excellent review. His comparison of the developed theory to F. de Saussure's ideas on the role of language in cognition and the difference between the two is of course most welcome. His critique is clear and important and it is being addressed in the new version of the paper. Prof. Fontanari emphasizes that the paper is not clear on three issues, the difference between human and animal cognition, the role of language in cognition, and the role of combinatorial complexity in associating words and objects (or meanings in general). These are the most important points since all of these issues are at the center of this paper. According to Prof. Fontanari critiques, the new version of the paper addresses these issues in more details, and hopefully clarifies the points previously uncertain.techniques register brain signals originating from multiple concurrent processes in the brain. Identifying NCCC events related to specific experimental conditions has been difficult, because all unrelated processes in the brain appear as noise, masking signals of interest. I appreciate Prof. Kozma suggestions to combine the developed technique with theories developed by Prof. Freeman and himself. The developed technique enables identifying multiple processes and separating them from “other” processes irrelevant to experimental conditions of interest. I agree with Prof. Kozma suggestions, that of greate interest would be collecting and analyzing experimental data related to predictions made in this paper. These include neural processes connecting cognitive and language brain areas; vague nature of cognitive representations as related to crisp language representations; increase in vagueness of cognitive representations at higher levels of cognitive hierarchy, and other predictions in the paper. Whereas the publication concept of WebmedCentral journal seems revolutionary and highly desirable, it is an unfortunate deficiency that the current production creates un-predictable reformatting of the original files, and despite several attempts, still some equations are difficult to understand and some figures are not of sufficient quality. I hope that the journal management would take care of these deficiencies. Meanwhile, interested readers could ask me for the pdf file of the paper, either through the commentary in the journal or by finding my email on the web, my webpage and email is easy to find.
Responded by Dr. Leonid Perlovsky on 06 Nov 2010 01:53:25 AM

  • Other Comments: This paper summarizes the author's prominent view on the close interrelation between cognitive development and language. The work is based on the author's decades-long pursuit of developing advanced pattern recognition methods using Model Field theory (MFT) in a wide range of areas, including image processing, bioinformatics, economics, human cognition and consciousness. My specific relevant expertise relates to the area of neurophysiology, specifically the analysis and interpretation of EEG and ECoG experimental data, identification of neural correlates of higher cognition, and the development of complex neurodynamical models of cognition. Therefore, I focus on the part of the target article dealing with experimental aspects of perception. The introduced Model Field Theory (MFT) is a very elegant framework to pose extremely complex identification problems in a transparent form. The present work elaborates on the consequences of MFT to the operation of cortical information processing, and the potential identification of neural correlates of cognition and consciousness (NCCC). Using powerful fMRI, MEG techniques, significant headways have been made in finding NCCCs. However, in everyday life one would desire having a more flexible nonintrusive method, which allows monitoring cognitive functions using scalp EEGs. This is an extremely complex task, and several leading exponents in the field declared that it is unsolvable. The use of MFT is especially beneficial in this context, as it shows that it is indeed possible to extract cognitively relevant information from extremely noisy data, containing high level of clutter and dynamical effects. Combination of MFT and recent advances in neurophysiology by Freeman and his colleagues provides a unique opportunity to gain an insight into cognitive processing in humans, using nonintrusive techniques. The target article refers to the detection of phase cones in EEG data. Phase cones are the manifestations of cortical phase transitions observed in the spatio-temporal brain dynamics. Freeman and colleagues developed the cinematic theory of cognition, in which human cognition is viewed as a sequence of frames manifested in cortical amplitude-modulation patterns, while the transitions signify the shutter operating by means of rapidly expanding phase synchronization patters in a conical shape. The target article describes the potential of MFT and the related Dynamic Logic (DL) approach to identify phase cones as NCCC. This by itself indicates a critical breakthrough in the field of cognitive monitoring. However, the author goes even further, and provides a comprehensive theory linking language and cognitive development. It would be very interesting to see how this theory could be manifested in neural correlates, which will be the objective of significant research in many years to come. Several technical/typographic comments: - The equations are not properly shown in the version I have access to. The same is true for some of the figures, especially Fig. 4. In a revised version of this work, these problems should be eliminated. - The reference list should be extended with works by prominent authors in the field. The author does mention L.A. Zadeh, W.J. Freeman, M. Arbib in the text; but the corresponding basic references should be added as well.
  • Competing interests:
    No
  • Invited by the author to review this article? :
    Yes
  • Have you previously published on this or a similar topic?:
    No
  • References:
    None
  • Experience and credentials in the specific area of science:
    Neurodynamics, EEG analysis, Cognitive modeling
  • How to cite:  Kozma R .Joint Acquisition Of Language And Cognition - A Review[Review of the article 'Joint Acquisition Of Language And Cognition ' by Perlovsky L].WebmedCentral 2010;1(10):WMCRW0093
Report abuse
 
Dear Robert, I am corresponding with them, asking to improve their procedure. Thank you Leonid I added recommended refs, and other mods. I am thankful to Prof. Kozma for emphasizing the fundamental issues addressed and solved in this paper, and for outlining an entire new research field that would be opened up by combining the developed technique with the currently available EEG and ECoG experimental data. Especially interesting is the vast area of studying neural correlates of cognition and consciousness. Finding these events in the past was difficult because EEG and other brain imaging techniques register brain signals originating from multiple concurrent processes in the brain. Identifying NCCC events related to specific experimental conditions has been difficult, because all unrelated processes in the brain appear as noise, masking signals of interest. I appreciate Prof. Kozma suggestions to combine the developed technique with theories developed by Prof. Freeman and himself. The developed technique enables identifying multiple processes and separating them from “other” processes irrelevant to experimental conditions of interest. I agree with Prof. Kozma suggestions, that of greate interest would be collecting and analyzing experimental data related to predictions made in this paper. These include neural processes connecting cognitive and language brain areas; vague nature of cognitive representations as related to crisp language representations; increase in vagueness of cognitive representations at higher levels of cognitive hierarchy, and other predictions in the paper. Whereas the publication concept of WebmedCentral journal seems revolutionary and highly desirable, it is an unfortunate deficiency that the current production creates un-predictable reformatting of the original files, and despite several attempts, still some equations are difficult to understand and some figures are not of sufficient quality. I hope that the journal management would take care of these deficiencies. Meanwhile, interested readers could ask me for the pdf file of the paper, either through the commentary in the journal or by finding my email on the web, my webpage and email is easy to find.
Responded by Dr. Leonid Perlovsky on 06 Nov 2010 01:52:17 AM

  • Other Comments: This paper presents an interesting computational framework for the modelling and investigation of the link between language and cognition. As such the work is consistent with neuroscientific and psychological literature on embodied cognition, as in the seminal examples of the perceptual symbol system (Barsalou 1999, Behavioural and Brain Sciences) and of language grounding in action (Glenberg & Kaschak 2002, Psychonomic Bulletin & Review). One potential extension of this hierarchical modelling framework for language/cognition interaction is that of modelling the relationship between hierarchies in language (e.g. compositional phonetics, compositionality in syntax) and hierarchies in motor knowledge (e.g. compositionality of motor primitives for complex actions). Given the compositional nature of Pelovsky?s approach, one obvious extension of this work is the modelling of action and language integration in embodied cognitive robot platforms. Initial developments in this field have focuses on the developmental perspective in action and language learning in humanoid robots (Cangelosi et al. 2010, IEEE Transactions on Autonomous Mental Development ? see also www.italkproject.org). A fruitful combination of MFT-DL with cognitive robotics methodologies can lead to advances in both our scientific understanding of language/cognition/action relationship, and in technological development for human-robot communication systems.
  • Invited by the author to review this article? :
    Yes
  • Have you previously published on this or a similar topic?:
    Yes
  • References:
    None
  • Experience and credentials in the specific area of science:
    None
  • How to cite:  Cangelosi A .From language to action hierarchy: a proposal for robotics implementation[Review of the article 'Joint Acquisition Of Language And Cognition ' by Perlovsky L].WebmedCentral 2010;1(10):WMCRW0068
Report abuse
 
Dear Angelo, I am corresponding with them, asking to improve their procedure. Thank you Leonid I am thankful to Prof. Cangelosi for a thoughtful review of my paper. Indeed an extension of this work to mathematical modeling of the perceptual symbol system (Barsalou 1999) would be an important continuation of this work, and this would be an interesting direction for future publications. Equally important would be extending it toward grounding language and cognition in actions, which could be achieved by combining the current approach with those developed by Prof. Cangelosi in a number of publications. Indeed several new directions of research would emerge from combining the developed method with cognitive robotics. I am thankful to Prof. Cangelosi for outlining these fields of future research.
Responded by Dr. Leonid Perlovsky on 06 Nov 2010 01:51:27 AM