Monday, November 29, 2004

Cosmic Reproduction

Scientific theories make an infinite set of predictions of which only a finite set can ever be tested. (a.k.a. "The problem of induction".) Humean scepticism led us to accept that we can't definitively prove any theory correct, but Karl Popper qualified this by insisting that a single observation can demonstrate that it is incorrect. He therefore insisted that the test of a good theory was whether it could be falsified.

More recently, Alan Sokal's reponse to the anything goes relativism of Feyerabend has been to point out that hardly anybody is systematically sceptical and relativistic in everyday life - so why should we treat scientific knowledge any differently to ordinary knowledge?

A quick summary of theoretical physics' biggest contemporary dustup can help us see how these ideas are being applied in the field. It will also allow me to reiterate my point about how relativism is itself rather relative.

The key protagonists are Leonard Susskind originator of String Theory, and Lee Smolin whose alternative approach goes by the name of Loop Quantum Gravity. They are not so much fighting out of different corners as sliding around different parts of a Möbius strip, occasionally embracing, occasionally catching the other with a glancing blow.

The things they agree on are actually very significant. For example, both men are convinced that universes reproduce giving rise to mutated offspring that differ in the values of the fundamental constants of nature. (Universe clearly isn't a word that was ever supposed to have a plural. Multiverse, coined in 1960 by Andy Nimmo is the term typically employed to overcome this semantic hurdle but Susskind suggests we use Megaverse as the collective noun, given that multiverse has been variously used to refer to both the set and the members of the set.)

If you thought dark matter (the stuff that hypothetically makes up most of the mass of our local universe and yet is as yet un-observable) might pose problems for a discipline that self-consciously progresses by observation, then just imagine the extent of the theoretical conundra posed by the idea that most of reality is behind a cosmic horizon we are unlikely ever to be able to transcend.

The stakes are suddenly very high: "If a large body of our colleagues feels comfortable believing a theory that cannot be proved wrong then the progress of science could get stuck", Smolin warns.

His tetchy exchange with Susskind began over the question of how significant it is that we ourselves are here in this particular neck of the megaverse. Susskind supports the Anthropic Principle, which states that the shape of the particular universe we inhabit is the way it is precisely because we are in it. Smolin says this is unfalsifiable and therefore "outside science". (Though it's clear that as cosmology reaches beyond the visible universe the boundaries of science are being inevitably tested, and rigorous investigation now cohabits uneasily with equally rigorous speculation.)

Smolin is convinced that the primary mechanism for reality reproduction is bouncing singularities. The abundant presence of carbon in any universe is the ideal condition for the formation of stars massive enough to collapse into black holes that go boing. Carbon-based lifeforms such as ourselves are a side-effect.

"Pah!", snorts Susskind - even Stephen Hawking now admits that black holes don't lose information so if they are where new universes are conceived, they would be born in a pristine quantum condition with no memory of the initial state, like offspring with no determined genetic resemblance to their parents. Susskind theorises that the megaverse is instead in a state of eternal inflation, which constantly spawns new pocket universes as it grows. Most are inhospitable to life.

The two would appear to agree that our own universe is untypical, but crucially Smolin asserts that the untypical is probably typical overall. This counterintuitive idea is justified because in any fitness landscape the distribution of variants peaks around small regions of parameter space. A typical universe would therefore appear untypical amidst any randomly selected group. (It's intriguing that both men also seem to assume that maximum reproduction is the only measure of fitness. Could there not be some other selection pressure?)

This argument is occuring at the point where classical determinism is being distorted almost beyond recognition by the infinite potential represented by multiple worlds. Intuitively I side with Smolin in the debate, not because I feel I know that black holes rather than eternal inflation account for the proliferation of multiple realities (although it is a vaguely cooler idea!), but because his theory is the more background independent of the two: "The view of time evolution that Susskind wants to preserve is tied to the existence of a fixed background...Eternal inflation is also a background dependent theory, indeed, some of its proponents have seen it as a return to an eternal, static universe."

In my view there is also a kind of fitness landscape of knowledge in relation to available information - this might perhaps be the fundamental nature of reality. All knowledge is therefore relative, because the fixed background scale is entirely implicit - potential not actual. Knowledge and value are also relative to the environment in which it is constructed. In another world with different constants "something else goes" (to paraphrase Feyerabend), and it remains to be seen whether the scientific methodology can stretch to bring coherence to all possible alternative realities.

Meanwhile, in our world we simply have to judge things by how reasonable they are. There's no fixed background rulebook just for us. Nothing else I believe in makes sense if this is not in fact the case.

No comments: