related texts

My comments indented and in italics

Editor letter
Dear Yehouda Harpaz

Re: Journal of intelligent systems

I have now received reports from three independent referees for your paper entitled "Can the neurons in the brain implement symbolic systems?"

Unfortunately one referee recommend rejection and I am inclined to agree with the comments made. Please find the comments enclosed along with this letter.

What about the other two referees? The editor did not say what they said, and did not send their reviews.

The nature of the referee's comments are such that the objections cannot be met by simple revisions. I am sorry that I cannot therefore accept the paper for publication. I hope that you will find constructive points from the review process which will help you develop your research.

An impressive statement, considering that the review is mainly personal insults and demagoguery (see below).

Your sincerely,

Professor MJ Wright
Editor in Chief
Journal Of Intelligent Systems

Name of paper: Can the neurons in the brain implement symbolic systems?
Author: Yehouda Harpaz

This paper has to be rejected for several reasons.

First of all, the author's knowledge in neurosciences is poor. Although the question he poses is legitimate, he is unaware of the pertinent discussion in progress at least over five decades. As early as in 1943 McCulloch and Pitts developed a neuronal model trying to show that neurons are logical devices operating according to all-or-none logic. This neuronal model requires exactly clocked operations. But the early enthusiasm about its logic and arithmetic capabilities was soon followed by the criticism of its neurophysiological implausibility.

Starts with an insult ("knowledge in neurosciences is poor."), which is probably based on the "unaware of the pertinent discussion". However, the only glimpse of this discussion is a theory from 1943 that nobody takes seriously, and is irrelevant anyway, as it has nothing to say about symbolic systems (it is about logical operations). The reviewer does not give any reference which is actually relevant or newer than 1943, simply because there aren't any.
Until now exact pulse periodicity has not been observed in neural signals. Therefore, the author is right that single neurons cannot implement logical functions, but unfortunately his arguments are not new.
This is a typicial demagogical manuever, based on two implications which are not stated explicitly:
That my claim is that "single neurons cannot implement logical functions"
This implication is almost explicit, and is clearly false. I am not discussing logical functions at all.
That I based my claim on the lack of pure periodicity
this is implied less blatanlty, and is false again. I based my claim on stochastic connectivity.
If the two implications were true, then my arguments were realy not new, but they are false.

Considering the quoted "textbooks" or several papers one has the impression that the author is neither in neuroscience nor in artificial intelligence up to date. Apparently he does not realize the complexity of the problem he dealt with.

Continue with some insults without any justifications. Even if these comments were true, they don't have a place in a review of a paper, which has to be about the content of the paper itself.

The double quotes around "textbooks" is another demagogical trick, intending to imply that the the textbooks that I refer to are not realy textbooks.

From a neurobiological point of view the neural assembly theory based on many experimental finding is one of the favourite approach to discussing the author's topic.

Realy? The "author's topic" is implementation of symbol tokens in a system with stochastic connectivity. Where are the discussions of this in literature of neural assembly?

The recent discovery of stimulus induced cortical synchronization in the visual system now provides a basis for bringing neural mechanisms of local and global stimulus encoding into juxtaposition with connectionist conceptualizations, with a view of accounting for construction of subsymbolic composites out of elementary stimulus features, on the way towards creating symbolic representation.

If this sentence looks like a gobbledigook, it is because it is. Any of the terms
is vague and flexible eough to make this sentence completely meaningless. A possible interpretation is that it means "stimulus induced cortical synchronization is useful in creating symbolic representation", but the reviewer doesn't actually give any hint how "stimulus induced cortical synchronization" can affect the argument in my paper.

It is interesting to think why the reviewer bother to write this gobbledigook, as anybody with expertise in the area of cognitive science would easily see through it. Apparently he thought that the reader of the review (i.e. the editor) will be impressed by it. Apparently it worked.

The author does not refer to this important approach.

Authors who do not see the complexity of a topic or who cannot estimate if there is something new in their argumentation make tendentious statements. So this author does.

Continue with some more insults.

In summary, I see no possibility of proposing a revision of the paper.