It is important to avoid the critical mistake of conflating our modern notions of
complex life (even the most basic
single-celled life being relatively complex so far as
molecular machinery goes) with the sort of 'life' which is invoked by this question. Indeed, the question to be asked first is, what is 'life' at all? It seems upon examination that we ought to count as a living thing any entity which, having come into existence, is capable of making a reproduction of itself prior to the end of its existence, so long as such reproduction is capable of in turn making another generation of reproduction, with no inherent end to this process incumbent in the composition of the original thing or its progeny. And so we then must ask, what is the simplest possible arrangement of atoms into a molecule capable of either
self-replication, or of replication of other molecules?
The reason the 'either/or' must be asked is that there are two possible paths to answering the self replication question. If, in nature, a molecule can be formed which is capable of replicating itself -- that is, of attracting atoms of the same kind of which it is composed in such a way that they order themselves along the same alignment as the atoms of the original -- then no second molecule is needed at all. But if there is a molecule capable by its structure of copying molecules around it, and it bumps into a second molecule, the same as itself or different from it, but equally capable of copying molecules around it, then each could make copies of the other ad infinitum; and if these copies stay essentially paired up, then you've got your standard two-molecule self-replicating machine. So forget the unnecessary notion of hundreds of molecules comprising chains of hundreds of atoms each, what is needed is either one self-replicating molecule, or two molecules capable of replicating something else around them.
Now, these sorts of molecules are hardly as hard to come by as one might
imagine. The simplest self-replicating molecule is, naturally, a
crystal; and the simplest crystal is that of frozen
water. With water, the amazing complexity of the crystalization process is displayed in the formation of
snowflakes, which take on any of a seemingly infinite array of tri-symmetrical
hexagonal patterns. But though water must freeze to crystalize, a great many crystals of somewhat (though not too much) greater molecular complexity can be observed to form and grow through the same molecular self-replication process at higher temperatures in rocks and caves, under the ground, sometimes right out there in the open, some even at quite high temperatures.
But, naturally, we don't consider snowflakes and other crystals of that nature to be alive. Snowflakes form from circumstances, and are not borne from other snowflakes. Crystals do not reproduce beyond the environment of their initial formation and growth; there are no new generations to consume the remains of the old and adapt to the conditions around them. But, in truth, it is quite likely that there is at the basest level no sharp division between what we can consider to be alive and what we can not. After all, a sufficiently complex self-replicating molecule which quite fortuitously is in a position to find materials with which to replicate itself indefinitely could be considered alive. Or perhaps life could require a slight step up in complexity even above that -- perhaps a central molecule which not only replicates itself, but replicates (or causes to be replicated) some other structure. Indeed, what we seek, then, would be something much like
this example of biologists creating a self-replicating RNA molecule, which not only replicates itself, but additionally generates an enzyme, a separate molecule capable of assisting in the replication process.
And, really, that's all it takes, for once
such a thing as this exists in an environment sufficiently soupy with raw materials, it not only will replicate itself, but some copies can and will generate copying errors (perhaps even reduplication of entire strings) which can with surprising quickness lead to better and better reproduction models. And that's a process powerful enough to bring about limitless variation, just as we experience in the most complex of molecular machines, ourselves.
----
locke baron says re spontaneous creation of life on Earth: Things like viroids and prions really muddy up this mess, don't they? Prions are pretty much just rather complex crystals, and viroids aren't a heck of a lot more than that....
DTal says re spontaneous creation of life on Earth: .... I think your closing paragraph is overdrawn, however - a molecule that copies itself need not copy itself in such a way that copying errors leading to increased complexity, or that the mutated molecule will *pass on* its deformity. Obviously this did in fact happen, but the interesting part of abiogenesis is it's not at all easy to see how....