Reasons we might care about definitions of life:
- knowing what to look for - for fun(ding)
- ethical implications if we count simulated life as alive, bearing in mind that simulating Darwinian evolution (if not intelligence) is really trivial.
- other ethical implications
We think that definitions of life really do affect funding for the right kinds of space exploration, and for funding of other good research (e.g. funding for spectrographic analysis of dust clouds, like Dayal's research which he had to piggyback on other grants because it wasn't fundable).
Also, when we get there we don't want to stare right at some alien life and think it's not alive. There's already a danger of this on Mars and it's only going to get worse if we explore lots of other places.
Is this actually an argument for having a definition? A definition might fool us into counting something living as dead.
A definition that we all liked at the beginning of the course is "Life is what we call alive when we see it." But we've agreed that this is not ideal, because, being retrospective, it can't be used to prioritise funding.
Is there a definition that starts from "life is a system keeping itself out of equilibrium"? Maybe, but you need to go a long way from that starting point.
Possible definition: "Life is Darwinian evolution." This solves the funding problem, because Darwinian evolution is the really interesting thing we want to look for.
Problems with that:
1. It doesn't talk about individuals, but maybe that's OK. (See also F&S's biosphere definition.)
2. What about indirect products of Darwinian evolution? Maybe that's OK: because for funding purposes we can say we're looking for life or for products of life.
3. Darwinian evolution is a matter of degree. (This is the main reason we read "The Extended Phenotype".) What about photocopied pieces of paper?
4. If we find things that we want to count as living that don't have separate generations, or don't have individuals, they won't match our theories of Darwinian evolution.
Think about systems analogous to bee hives but without strict divisions between individuals, and trying to extend the concept to genes and mutation within an individual; where an organism is having genetic changes within itself without generations. I think this is still Darwinian or at least pseudodarwinian because the genes inside the thing might have generations; but would we see that to notice it?
Also think about the intelligent gas cloud in Fred Hoyle's "The Black Cloud", where a single intelligence controls itself and its whole environment, so it doesn't have Darwinian evolution even though presumably it did in the past (in the book, IIRC, it doesn't know whether it did).
5. We don't have a totally fabtastic definition of Darwinian evolution. We've identified problems with Dawkins's definition. But at least other people are working on this problem for us!
6. According to Dawkins, it's only genes that have Darwinian evolution, but we count the whole of (e.g.) a human as alive. That's not usually a problem because nobody says we have to count only the replicators as alive. We usually count the vehicles (i.e. organisms) as alive too.
7. If a system starts with Darwinian evolution but then a benign intelligent something gets a lot of control of it, it might want to move away from Darwinian evolution to a different kind of evolution (e.g. without individuals or generations, or whatever) ... but it would still be alive. Note that the organisms in that system could even stay biologically identical to the way they were, perhaps, just changing the way competition works in a way that stops the system being Darwinian.
So:
Life is the interesting components of systems that have Darwinian evolution, and their vehicles (in Dawkins's sense), and maybe some of the most interesting other products of such systems.