Review of Shadows of the Mind by Roger Penrose

Improbably, "Books for Buds" continues. This one is for TenaciousK, and evidently not a moment too soon. Evidence for TK can still be found here, and although he ventured into the Outland some months ago, you can still catch him around once in a while. The rationale for Shadows of the Mind is that it's about consciousness and the brain, which is something that TK writes about frequently and, evidently, is involved in professionally. Sounds good as far as it goes, but the author of Shadows of the Mind is a mathematician, and approaches the subject from the edges of math philosophy and quantum theory, whereas our subject would be better described as a behaviorist, looking at the various identified or speculated physical and chemical processes in the brain, and assessing how they affect people's actions and thoughts. Does this book really suit TenaciousK? Well, let's put it this way: Roger Penrose opens up with a fair-minded 3000-word discussion about gendered pronoun conventions--I think it works out just fine.

#

I don't mean to imply that the author's verbosity (or TenaciousK's for that matter) is a handicap. For his abundance of words, Penrose reads is quite a pleasant read. (Careful observers may note that it took me the better part of a month to plow through this anyway. I blame free time.) He writes with an enthusiasm that can be charming in professors, an animated spasticity that translates, with a a lot of exclamatory asides and italicized profundity, to the written equivalent of popping out of his chair with and gesticulating excitedly when the ideas frequently take him. You get the feeling that here's a guy who'd miss meals in the throes of inspiration, and who'd whip about the front of a classroom like a sprite--four chalkboards filled with wild illegible scribbles--as he teaches students. It's infectious: I'd absolutely love to see one of this guy's lectures.

I do wonder who his purported students of Shadows of the Mind are supposed to be. I imagine them to be people like me, with an adequate technical education, but outside of the field. He writes generically to lay people, but without the various slapdash theoretical concepts I've gathered over the years (my quantum has always been pragmatic, and still inadequate for my job), I'm not sure I'd follow very well what the hell he was talking about at all. He does a good job of keeping the language easy to follow, but in the process, he necessarily obscures mathematical details. It's cutting right to the hanging philosophical questions, but glossing over how they arise, and while his practical examples are cute, they're a bit confusing. He doesn't leave the reader in a good situation: deferring the details of his ideas, as he must, it's not easy to challenge them, and well, they're the proverbial extraordinary claims. A lazy dolt like me is stuck going after his logic.

Here's what he's basically done: he has taken the weird mysteries of cognitive science (where does consciousness come from?), and looked for the answers in the difficult fringes of mathematical philosophy (a sound algorithmic system can not observe certain mathematical facts about itself), quantum theory (at what point does quantum superposition give rise to classical observation, and for that matter, how the hell does gravity fit in?), and cellular biology (is the cytoskeleton involved in information processing?). When he finally gets to them, his propositions are pretty wild, but Penrose is too smart and too honest not to recognize counterarguments, and the problem is that they sound more convincing (to me), and no less weird, than what he's advocating. Why must we insist, for example, that human thought is a sound algorithm? That it is algorithmic (that is, it follows rules to change its internal state and output, which in this case would be physical rules), certainly appears to be the physical case, and considering the mathematics, a quote from Alan Turing claiming that some level of fallibility is (perhaps) a necessary condition for intelligence resonated better than a chapter's worth of disclaiming that very thing. He needed to ride on a sophisticated appeal to incredulity that such a machine as ours cannot contradict itself internally. The other issue is that for his pondering on quantum mechanics, I have difficulty accepting that brain physics are special in any way that device physics aren't. Yes, there are quantum state reductions that must occur for information to propagate in neurons, but this is also true for electrons in transistors (or you know, in almost every physics we exploit for anything), and, for that matter, more obviously the case.

I don't want to sell his speculations short--they're fascinating hypotheses--and it would be thrilling to have them proven right. But Penrose is also interesting as hell, and expertly grounded, as he points out the disconnects in the theories of AI and quantum mechanics. These discussions make up (perhaps tellingly) the bulk of the book, and I'd probably recommend it for the background more than for his particular conclusions.