Abraham Moles's newly translated Information Theory and Esthetic Perception is a book with one interesting idea, which emerges, crablike, from the murky burrow of Chapter Five. The idea is that realizations of a work of art--the performances of a symphony or play--are not uniquely determined by the output of the composer or author, but vary from one occasion to another, so that people like to hear Beethoven's Fifth Symphony or see Hamlet even when they know the notes or lines by heart.
In Information Theory and Esthetic Perception, the exposition of this idea requires 45 pages. The ratio of verbiage to original thought in the rest of the book is vastly higher. Moles, like many before him, has failed to see that a relabelling of old problems does not necessarily increase our knowledge or understanding of them.
Let us consider perception by an individual human being as communication from the external world to that human, says Moles, now a professor of philosophy in Strasbourg. Let us consider in detail artistic communications, since it is particularly easy to isolate them. Then esthetic perception, as a special kind of communication, should be amenable to analysis by information theory, Moles concludes, since information theory is a mathematical theory of communication.
This reasoning is an example of what philosophers call the fallacy of equivocation: what Shannon and Wiener, inventors of information theory, meant by "communication" is not what Moles has in mind. However, trying to apply a theory in new domains is fair game, so we proceed.
Moles first presents a derivation of Shannon's H-formula for information content which does not live up to its claim to being non-mathematical. (How could one derive a mathematical expression non-mathematically?) He then applies the formula to a few examples. He ignores recent statistical findings that in estimating H from relative frequencies some correction must be made for sample size; but that doesn't really matter because of his relative frequencies are drawn out of thin air. Instead of actually calculating H with his few real relative frequencies, he rounds them off and lumps them together, saving half an hour's work with a table of logarithms.
After presenting information theory in a way that should attract none but the credulous ("intuitively"), he rehashes and relabels the audiology of Stevens and Davis, date 1938. Psychophysics has advanced since then. The translator has inserted one wistful reference to the sone and mel scales of loudness and pitch, but that is no substitute for the complete reworking need. In the entire discussion of the sonic message, the "information theory" so painfully presented before plays no practical role.
Again and again the skeleton of a theory is trundled out dressed as a theory. Page after page of factually empty diagrams, consisting of boxes and arrows, indicates the way the results, if they existed, would be organized. Inside the boxes are only labels of categories that have not been measured or understood. Great is the intellectual debt of these diagrams to medieval analyses of the faculties of the mind or the virtues of God.
A number of smaller inconsistencies are also irksome. A few examples: Dr. Moles says the range of loudness in music is from 30 decibels to 100 decibels. On the next line he says Stokowski performed triple pianissimo at 20 decibels; was that not music? In Chapter 1, from concocted statistics about a "typical" musical score, he calculates the redundancy of certain aspects of musical notation to be 15 percent. In the rest of the book he refers to the great redundancy of the musical score in comparison with the slight redundancy, perhaps 20 percent, of musical performances. He should have rigged his example to support his later (equally unfounded) conclusions. There is a continuing confusion between physical frequency and pitch. Why he numbered some of his figures and not others is a mystery wrapped in an enigma. His index would be laughably inadequate if it were not so unlikely that anyone would ever want to use it.
The translator, who is or was a Harvard undergraduate in mathematics, has sliced his way through most of Moles's opaque French and has rewritten most of Moles's numerical example so that they are at least internally consistent and comprehensible, if not worth the trouble in the first place. But his performance too is less than ideal. The top of page 200 is as obscure in English as the original was in French again "modulated transmission of coded impulses" is a ludicrous, if more grammatical, transformation of PCM, of "pulse-code-modulation" transmission.
Finally, unless this translation was intended as a document of historical curiosity, why did it take eight years to bring it to an eagerly waiting public? The translator admits taking four, which is difficult to understand; this leaves four years to the University of Illinois Press, which seems unprofitable.
If it is any consolation to Moles and his obviously devoted translator, it may be admitted that, when the original French edition appeared in 1958, it was quite fashionable to dash cybernetically about. After a sufficiently loud and long obeisance to information theory (sufficiently loud and long to attract financial support and publication), it seemed not to matter what profit came from applying Shannon's H-formula to any kind of relative frequencies, real or imaginary. It was no fault of Shannon's that, by the end of the Fabulous Fifties, the H of his formula stood most often for Hooey.
In the translator's preface, Moles is quoted as saying that he intends his book to serve "as an introduction to an informational theory of psychology which would not require a too extensive knowledge of mathematics, and could consequently be suitable for students in psychology." For the reasons given above, I would not recommend it as a text. Neither would I recommend it as a reference. I would recommend it as speculative entertainment to those who have a vast capacity to forgive carelessness, loquacity, and dullness.