Randomness is a manifestation of locality. The world is big and anything works at a local level, asynchronously, and randomness ensues.
I want to advance the following hypothesis about the origin of life.
Life is a manifestation of the computational universality of a collection of chemical reactions.
Indeed, there probably are many small collections of chemical reactions which, coupled with a random chemical reduction algorithm, form a universal computing model.
A proof of principle for this is chemlambda. There are still to discover real chemical reactions which implement the (invisible in chemlambda formalism, for the moment) moves shown at the chemlambda moves page.
But they are so simple that there have to be many, many such chemical reaction.
In a system, in a chemical soup, if it happens to appear these chemical reactions, the following is a game of computation and self-multiplication.
Because universality means, in this particular case, that with non-negligible probability, anything can be achieved.
The role of randomness is tricky. On one side randomness selects resilient creatures. That’s a funny thing, for example in chemlambda good candidates for living creatures are quines.
A quine in chemlambda is a molecule which stays the same in a daterministic world. This gives to the quine molecule a certain resilience when faced with randomness, which makes it to have a life: it may grow, it may decrease, for a time it may play around the deterministic state, and it may eventually die.
This is illustrated in the first battle of microbes demo, where several quines are put together and “fed” with enzymes, which appear randomly, but such that if at a moment there are more enzymes for the moves which increase the number of nodes, then the next time the probability of appearance of such enzymes decreases in favour of those which decrease the number of moves.
So globally it appears as if the the quines compete for the moves and those quines having a greater diversity of possible moves thrive, while the other die.
The 9_quine is the most fragile quine, as you see in the demo many of them die (i.e. they transform into a small molecule which is inert to any reduction).
There is a lot to add about this, for example there are other quines which behave like they adopt the strategy of spores, i.e. they regress to an “egg” state and they flourish later, from time to time, when they have to “compete” with bigger quines.
Of course, all this is in the eye of the observer, it is an emergent behaviour, like the intelligence of a Braitenberg vehicle.
But what if quines are a bit too fragile for life? Maybe there are molecules who grow to an approximately stable configuration, in random conditions, for a time, at least until they self-multiply.
[Have you seen the story of the 16 bubble quine, from egg to motherhood?]
Suppose, just suppose that in deterministic conditions such a molecule would grow slowly, instead of being a quine.
This is consistent with the requirement to be resilient in random conditions, there will be a second part of this post when the demos are prepared.
But it has a curious corollary, namely that such a living creature will blow out, like a cancer, in too calm, too deterministic conditions.
The example I have and play with is a molecule made by two 9_quine and a 5 atoms molecule which, if left single, it grown in a regular pattern, but in the deterministic algorithm, when coupled by some bonds with the two quines, it grows very very slow.
This molecule, under random conditions, display an amazing variety of shapes. But all the runs show the same thing, more or less: that the two 9_quines have a role of taming the growth of the molecule, keeping it controlled, but at some moment the 9_quines die, somewhere in the molecule, in some unrecognizable shape, and after that the molecule reverts slowly to the regular growth pattern (which makes it unsustainable if there are phisical limits to the growth).
So not only that randomness select creatures who can survive (and self-multiply) in random conditions, but it may select creatures who live in random conditions, but who die in deterministic conditions.
Maybe that is why life hates when everything is predictable.
I close this post with the comment that however, there are molecules which arrive at a determined state in random conditions.
This may be very useful for computer like computations. The exmple I have is again the remarkable molecule for the Ackermann function.
See it in this video self-reproducing while it computes.
Apparently some molecules display a huge resilience to randomness. The Ackermann function molecule daughters finish the computation at different times, but they do finish it.