Levels in the evolution of (X)
This somewhat rambling post is an attempt to think about the two things in the history of the universe which have in some sense evolved: life, and thought.
Walter J. Ong was an English professor and theorist of information technology, that is, of the different modes of consciousness brought about by orality (speaking), chirality (handwriting), printing (which really should have a Latin word!), and, now, the various electronic media. As I read, I found myself persuaded of the existence of these differences—for example, orality emphasizes memorability, while printing emphasizes visual organization—but skeptical of the claim that these modes followed one another in a teleological sequence. Why should that be? I was particularly skeptical of his claim that electronic media are the next step in said sequence. To be sure, it is difficult to envision writing being invented before speaking, or printing before writing, or computers before printing. But it’s also easy to imagine any of these being invented in a slightly different way and so leading to an entirely different future. “Electronic media,” in particular, is a black hole: why assume that telegraphy, telephony, radio, television, computers, and the internet share some common essence, rather than being disparate inventions with disparate implications for the human psyche?
Meanwhile, I’ve been mulling over Eliezer Yudkowsky’s thoughts about natural selection (which are intended as an analogy for how artificial intelligence will bring about a new paradigm of self-optimization, and so are also, in a way, about that vague category “electronic media”). I’m still not sure what I think of AI GO FOOM!, but I am attracted to the more restrained form of teleology his approach suggests. Each innovation, he says, happens by accident, but each innovation also makes possible new innovations by opening up a new “search space,” so it begins to look teleological, when really it is just a question of dependencies.
The other development Eliezer thinks comparable to AI GO FOOM is human rationality, i.e. civilization. Here we come to the reason for this post: though Eliezer doesn’t talk about it, there exists a noteworthy homology between the various innovations which contributed to evolution, and those which contributed to civilization. The (roughly) three evolutionary developments Eliezer thinks “notable” are:
- Cellular integrity (and DNA is basically an extension of this): The point, Eliezer says, is to “Force a set of genes, RNA strands, or catalytic chemicals to share a common reproductive fate”
- Multicellular organisms: Of course this depends on cellular integrity. Eliezer says that “the key here is the controlled specialization of cells with an identical genetic heritage”; in other words, the cells aren’t bound together in the organism the way genes are bound together in cells, but rather the the cell itself comes to contain more genes which now express themselves only in special circumstances; the point is again to allow for more complex blueprints
- Sexual selection: Eliezer doesn’t go into detail here, and it’s difficult to sum up what sexual selection does. The Red Queen hypothesis suggests that, once sexual reproduction comes into being, sexual replicators can change more quickly and so outcompete non-sexual replicators; hence it being advantageous even for some single-celled organisms. But true sexual selection can exist only when the individual replicator has something like a “desire” to reproduce, which leads to competition between possible mating partners. Best-case scenario, this accelerates natural selection by having organisms be attracted to what will make them more fit; worst-case scenario, it leads to weird feedback loops and the peacock’s tail.
Evolution, per Dawkins, optimizes genes; human reason, meanwhile, optimizes memes. If memes exist wherever there is learning, then the memetic equivalent of the first time a molecule copied itself (i.e. of the first gene), was the first time an animal got the idea to do something by looking at what another animal was doing—at that moment was born the first meme. Of course, like a free-floating molecular replicator, such memetic replication is quite weak. It takes further developments to get to things that seem really interesting. And these developments are in many ways analogous to those that took place in the history of genetics:
- Language seems a lot like cellular integrity. The point of language is to tie together a bunch of different memes in the vocabulary of the language itself, allowing them to be transmitted together (you can’t learn just one word of a language), and so allowing more complex memes to form than otherwise would have been possible.
- Writing is like multicellular organisms. It allows for the controlled specialization of thinkers who share an identical memetic heritage. Without writing you cannot have a division of labor between domains of knowledge, you can only have the received wisdom of your culture. Put differently, without writing, it’s difficult to develop jargon.
- Printing is like sexual selection. Writing did, of course, allow for communication between disparate regions, but it’s primary purpose was to pass on information within a society. Printing allowed for the development of vast communication networks, and so to the cross-pollination of ideas. It also seems reasonable to say that it edges out writing-sans-printing through something like the Red Queen hypothesis. Finally, printing can, in fact, exist without writing—printing just means pressing a single pattern onto multiple receptables. But it only gains it’s innovative power when it is combined with writing, ie with specialized thinking, at which point it produces something like the scientific revolution.
What, if anything, is “electronic media” like? Of course we can’t answer this question until we know what “electronic media” means: is Ong right that the most important aspect of it is that it allows for instant communication, or is Eliezer right to emphasize the way it allows a feedback loop in the designing of machines? If Eliezer is right that AI GO FOOM, then it doing so resembles nothing so much as the origin of life and thought themselves. In which case it’s silly to think about electronics as a stage in human consciousness, just as it’s silly to think about the origin of human consciousness as a stage in the history of evolution, rather than the stage at which biological evolution ceases to be the most interesting thing going on.
I’m suspicous of this way of thinking, of course, but not so suspicious that I’m willing to dismiss it entirely.