Review: Black Mischief: Language, Life, Logic, Luck

Black Mischief: Language, Life, Logic, Luck by David Berlinski
My rating: 3 of 5 stars

I wanted to like this more than I did... Berlinski weaves together linguistics, computer science, mathematics, statistics, and molecular genetics to come to reveal enlightening connections between them on the nature of life and success of evolution. For instance:

“Let us contemplate the following experiment,” Schutzenberger said, “the Weizenbaum Experiment.”

Joseph Weizenbaum of MIT had just published a book on the misuse of computer theory. Largely because he wished that he had said many of the things that Weizenbaum had said, Schutzenberger had taken an immense boisterous dislike to Weizenbaum and hence thought it appropriate that our experiment be named after him.

Now mention thus of an experiment may suggest that something or other is about to be executed or performed; in fact, the Weizenbaum Experiment is one of those purely imaginary affairs in which trains of thought are allowed to meander and then merge by anastomosis.

“In life,” Schutzenberger went on thoughtfully, “there are two mathematical structures or spaces. There is a space made up of DNA or the proteins. This is an alphabetic space. Its objects are words. And there is a zoological space. Its objects are organisms. This is a space of alphabetic representatives. We say that in both spaces a natural metric exists – the natural metric – and that evolution proceeds in both spaces according to this natural metric. What is more, there is a mapping between the two spaces. It is this mapping that establishes that my DNA serves to express me.

I nodded and began to take notes.

“Furthermore,” Schuztenberger said, “there is in the space of the nucleic acids or the proteins a probability transition system.”

“This probability transition system – do you have in mind a finite-state Markov process?”

“Yes,” said Schutzenberger dreamily, “a finite-state Markov process.”

Schutzenberger stopped pacing and folded his hands in front of him, the long, curled, tobacco-stained fingers locked.

“We now observe,” Schutzenberger said, “that the probability transition system is roughly in accord with the natural metric. We are speaking only of the alphabetic space, remember.”

I looked up, for the moment unconvinced. “Why?”

“We say that the probability transition system is in accord with the natural metric because the most likely changes in the system are those that transform strings into nearby strings.”

“I don’t see this principle follows from the idea that alphabetic changes are independent.”

“It does not,” Schutzenberger agreed. “It follows from the observation that the probabilistic structure of the alphabetic space is not uniform. Indeed, it is this observation that shows ultimately that life does not comprise an ergodic system.”

This was the sort of lovely lunatic leap that Schutzenberger was forever taking in conversation. I must have looked up with an expression of radiant confusion; Schutzenberger directed a warm, beaming smile into the ambient atmosphere, and went on, untroubled by my lack of confidence.

“Life,” he said, “is conservative. Not everything that can change does change. For the most part, biological strings do not change at all. When they do change, they change in only one position. It is highly unlikely that a given string will change in respect to every position. We do not in life see a strand of DNA change its character at every possible codon in one sudden mutation.”

I caught Schutzenberger’s point.

“Now we need something more,” he said, with the air of a man constructing a wonderful instrument, “a mirror, so to speak. We wish to say with regard to arbitrary strings not only how far apart they are under the natural metric, but how far apart their representatives in the real world are. For this we require an induced metric. Very common in mathematics.”

I stopped writing to look up, and shook my hand to release its cramp.

“So when we talk of strings of DNA or strands of protein,” Schutzenberger went on, “we can talk of the natural distance between them or their induced distance. Two sets of strings may be close under the natural metric and far apart under the induced metric. You know, there is the famous experiment in which chimpanzee and human polypeptides were compared. Simply considered as strings there is virtually no difference between them. Evidently there is some difference between a chimpanzee and a human being.”

We were for the moment both quiet.

“In fact, zoologists often assume that the chimpanzee and the human being are closer than they really are.” Schutzenberger held his own hand in the air, palm outward.

Having grown impatient with his own digressions, Schutzenberger finally said, “Let us now perform the Weizenbaum Experiment. We suppose that we have certain strings of alphabetic objects, and that there is some initial probability distribution defined upon them. That is to say, at the beginning of the experiment, the strings are most likely to be in a certain initial configuration. We also have – are you writing? Good – a probability transition system, one that tells which changes in the strings are probable and which are not.”

What detracts from this is Berlinski is enamored of his airy, peripatetic academic life and insists on inserting vignettes of his career ranging from name dropping to trite scenes to a recurring focus on breasts ("curiously bustless coeds", "humungous knockers", etc.). Chomsky gets special focus for his work in linguistics. Apparently, many of his peers are not impressed and more so since he only knowns one language. These criticisms are presented in a way that seems catty.

View all my reviews