[nwallin]

I used to think that too. Life could be anything; crystals, pure energy, eddies in the superfluid cores of neutron stars, swimming in seas of metallic hydrogen in Jupiter's core, living clouds floating in Venus' upper atmosphere.

The trouble is that when you start asking yourself about what the fundamental nature of life is, all the fun stuff starts to fall out.

For life to be life, it needs to do work. You need a metabolism. You need to "do stuff". The only way to do work ("work" in the physics sense) is to leverage an energy gradient; you need to have hot/cold, high/low pressure, bright/dark, high/low salt concentrations, high/low energy chemicals, positive/negative charge, etc. And you need to extract work from the differences between them. Therefore, you need to have a physical barrier between your high/low energy states; if you didn't, the system would reach equilibrium without you. So you need a physical barrier, it has to be a solid or liquid: you've eliminated your beings of pure energy, superfluid eddies, living clouds.

It's all well and good to use whatever weird energy gradient you want, but most energy gradients are simply too inconvenient; if you want to get energy, chemistry is ultimately going to be the only one that works. It's fine for plants to break water molecules with relatively high energy photons, but you can't build complicated structures with light. You need chemistry. And for chemistry you need to be in the liquid phase. So you've eliminated your crystal based life.

So everything is chemicals with complex structures in a liquid. It doesn't have to be proteins constructed with amino acids, it doesn't have to be ribosomes constructed from RNA, but you need some sort of 'interesting' chemical. How do you build interesting chemicals? Well you have to use carbon. You have to use carbon, no other element fits the bill. The metals are right out; they only form metallic bonds and ionic bonds which are non-structural. The noble gases are out because they don't do chemistry. The halogens are out because they can form just the one covalent bond. The chalcogens (oxygen, sulfur, etc) are also out for the same reason; with only two covalent bonding points you can form chains and rings, but not a branching structure. Nitrogen is too happy to form a triple bond with itself, producing a gas, so basically all molecules with lots of nitrogen atoms are explosive. Phosphorus is a decent candidate, but its chemistry is very finicky. Carbon is great, because it forms 4 covalent bonds that are not too strong and not too weak. Silicon is ... ok? But it cannot form double bonds, it tends to form weak bonds, and is unstable in most chemistries.

So life has to be carbon based. This puts an upper limit on temperatures, because interesting carbon chemistry breaks down at high temperatures. Venus is out, the depths of Jupiter are out.

So we're looking at other solvents besides water. Methane is a great candidate, and as a result lots of people are clamoring to get a Titan mission going. Ammonia is something people talk about, but it's unlikely to be able to form ponds or oceans on a world; it's probably going to break down into methane and molecular nitrogen, because again, carbon rules. But if we detected a large amount of ammonia in the spectra of a planet that could maintain temperatures consistent with liquid ammonia, you could bet that everyone is going to clamoring to study it in detail.

So what are we looking at? Carbon based life on planets with liquid oceans. Which is basically the same as us. We haven't gotten here because we're biased towards what we're like, we've gotten here by starting from first principals, considered everything, and eliminated the things which are impossible.

[don_esteban]

I don't see how your argument refutes living clouds. Hurricanes, after all, are self-organized heat engines, a proto-life that is hampered on Earth by very marginal conditions and that pesky dry land.

On gas giants, the conditions are much better for sustainable 'life' of long-lived cyclones, you can see huge amounts of them on Jupiter an Saturn, even very long-lived ones (hello The Great Red Spot). Instances of these cyclones merging with ('eating') each other are well-documented.

Yes, they work on very different scales (both space- and time-) then us, and might have problems reaching high-intelligence in reasonable time, but I see no reason a long-lived giant cyclonic storm can't match the complexity of small bacteria. In the environment they 'live' they would interact with similar 'life' on a continuous basis and the more successful ones would survive and might even 'evolve' in a Darwinian fashion.

This shows that the chemistry-based arguments are way too limiting.

Of course, stars themselves are a prime example of self-sustaining energy-extraction 'life', again, absolutely without the need for chemistry - there the problem is that the meaningful interaction with the outside environment is rather limited (on the star's scale) so it would be more difficult to 'evolve'. I would still like to see self-sustaining solitons of magnetic energy thriving in/on the stars, competeing with each other (and their interactions/fights ending up in solar flares), but I am not sure such things can work (unlike hurricanes, that definitely work).