Monkeys, Typewriters, and Interpretation

It’s a common statement. “Enough monkeys with typewriters, given enough time, could eventually write Hamlet.” The idea being that all human artistic accomplishment, and Creation in general, is essentially just random noise, and inherently meaningless.

Well, firstly, the claim is false… give typewriters to monkeys and they probably won’t even try to type with them, or if they do they’ll just hit the same key over and over and over. It won’t be the string of random letters and symbols needed to “eventually write Hamlet”. And even if they did just write random symbols, the experiment would have to be run for an impossibly long time before they produced anything. If every atom in the universe were a typewriter spewing out a random character every second, it would take longer than the universe has been around to write just a sonnet of Shakespeare’s, let alone Hamlet, which is much longer.

Not that that proves much; I never saw the monkey analogy as a very good one for random chance bringing about human art in the first place.

The issue isn’t whether monkeys could really write Shakespeare. Rather, we are meant to wonder why, if Hamlet is a finitely long work that can be converted into a number (take a text file with the content of Hamlet in it; that file is just a string of 1s and 0s, i.e. a really long number), call it H, that a random number generator would eventually spew out if it ran for long enough, why should we look at it as in any way transcendent? Why should we look at human thought as in any sense transcendent if everything it produces is finite?

The answer, I think, is that even if the random number H could be generated by a random number generator, it can’t be interpreted by the random number generator. There needs to be someone out there who picks out H from the other random numbers our RNG spews out, says “this is Hamlet”, reads it, and gets from it what there is to be gotten from the play Hamlet. That requires language, something that doesn’t seem to have a very good finite representation. Without a way of translating those 1s and 0s not just into letters (which can be done with a computer program), but into words, H is no more meaningful than H*1.1 or H*0.9.

I think it’s interesting that H, as a number, is meaningless in and of itself. The computer needs pre-written rules for how to translate H into a series of letters and symbols. If it has a dictionary installed it could then try (if it were told to) to analyze the strings of letters it sees and paraphrase the entire thing. But the computer would never look at the strings of letters and see words, by which I mean things that have meanings that we can try to approximate with other words, but cannot define exactly.

It seems to me that language is really what separates humans from computers, RNGs, or monkeys with typewriters – those all manipulate symbols, but humans actually use words, language. So the fact that humans speak a language, rather than just manipulate symbols, is what makes the number H not equivalent to Shakespeare’s Hamlet. They are the same quantitatively, but not qualitatively.


Comments are closed.

%d bloggers like this: