Can the theory of evolution be applied to non-living things?

Lam, could you please respond to Message 22?Tsjok, English isn't my native language either, but that doesn't prevent me from fashioning my texts in a readable format. As it stands now, your text didn't invite me to actually read it. You could also improve things by cutting down on your liberal use of brackets.WK, you said: Semantically, 'simulation of evolution' isn't the same as 'evolution', but in a computer programme the effect of a simulation of evolution is the same as that of the real thing.

---

"It's amazing what you can learn from DNA." - Desdamona.

Damaged Monarch,1. A simulation is by definition not the real thing. So a simulation of a living thing is not a living thing. So all simulated living things are non-living things. (But not all non-living things are simulated living things.)2. If data in a computer is changing via an evolutionary algorithm, then the question of whether the process should be called a simulation or not, is irrelevant, because the fact remains that the data is really changing. Judging from the result, a simulated evolution is indistinguisable from a real evolution.So what the example shows is that at least some non-living things can evolve.

---

"It's amazing what you can learn from DNA." - Desdamona.

It seems that your definition of evolution is based firmly on the one example that we observe in earth's living nature, and it seems you think that any form of change in a population of things must work exactly like that, in order to be entitled to the name of 'evolution'.When I made my programme, I did not intend to make an exact simile of nature's evolution. How could I, it would have been a foolish endeavor to even try. Instead, I only borrowed two principles: those of random mutation and natural selection. The rest of the machinery was all my own invention. Yet I observed a striking similarity between what happened to my population of solutions to the Travelling Salesman problem on the one hand, and what is continuously happening to populations of living creatures in the natural world, on the other hand.It is of no consequence that there are differences between the two processes of change. What is important is that two simple principles have a similar effect, regardless of the way in which those principles are implemented. Evolution is an iterative process. Let me put it this way: evolution and simulated evolution both produce real changes, whereas a simulated living thing produces simulated shit and an actual living thing produces the smelly kind. They are most assuredly not the same.I will think about a formal definition of life and of evolution, although I'm sure I would be reinventing the proverbial wheel. If you could build nanomachines (you know, solid old-fashioned Borg technology) that exactly mimick the behaviour of the cellular machinery that deals with all things DNA, and you would inject them into a cel, then, yes, why not?{edited to correct a typo}This message has been edited by Parasomnium, 05-06-2004 05:32 PM

---

"It's amazing what you can learn from DNA." - Desdamona.

That is a very interesting question, jar. If we could devise an algorithm that changes information via evolutionary principles, and we implemented the algoritm in two different substrata, say a computer and a room full of people with abacusses, then which of the two is a simulation of the other? In such a case, we might not want to speak of a simulation, but rather of a mapping between two congruous processes.

---

"It's amazing what you can learn from DNA." - Desdamona.

Lam, I apologize for giving you the impression that I was belittling you. It was not my intention to do so. But you yourself might want to try to control you emotions a little better. When discussing with Desdamona, your head nearly exploded, you said. I would hate to see the same thing happening here. I never mentioned ‘beings’ and ‘artificial lifeforms’ in the description of my programme. It deals with information, random mutation of it, and selection of it. It’s funny that you mention this because a simulated disaster is exactly what I used to overcome the problem of ‘local minima’. A simulated disaster would decimate the population and the local minimum situation would be resolved. We can observe the same phenomenon in biological evolution. If you stop doing it, I will not be saying it. Excuse me? Are you serious? Isn’t that what the poor tend to do, proportionately more so than the rich? Can you tell me what a simulation of information looks like? Can you distinguish between a sell-by date and a simulation of a sell-by date? I’m talking about evolution by random mutation and non-random selection, of which biological evolution is one example, and memetic evolution another. It can. Or do you think that only apples can fall from trees? When I’m using the word ‘evolution’, whether applied to genes or memes, I mean exactly the same process in both cases. You, however, are equivocating on the word attracted. That is a fallacy.

---

"It's amazing what you can learn from DNA." - Desdamona.

Thank you, Ned. You said very succinctly what I was trying to get across. Your definition of evolution certainly works for me.

---

"It's amazing what you can learn from DNA." - Desdamona.

Lam, I apologized. And I said I wouldn't want to see your head explode. I'm trying to be friendly, but you're not being very helpful.Can we put this to rest?

---

"It's amazing what you can learn from DNA." - Desdamona.

So what? Ned's definition has the necessary elements needed for evolution as a process to take place, regardless of what exactly it is that evolves. I think that biological evolution is just that process; it is not important for the process of evolution we see happening in nature whether the things evolving are alive or not.The important thing is that the process takes place at all, simply on the basis of random mutation and natural selection. In nature's evolution, random mutation takes place during a process that could be called one of the defining elements of life, but for my purpose of showing that random mutation and non-random selection can produce any evolving complex system, it is not important how the random mutations come about.I think that anything, whether alive or not (by whatever definition of life we'd agree upon), when subjected to those two principles of random mutation and non-random selection, would evolve, i.e. change in such a way as to become better adapted to the circumstances. The problem here is that you didn't get my point. All I was trying to tell you is that there is a difference between simulated life and real life and there isn't a difference between simulated evolution and real evolution. The excrement was just an example and a joke (which you didn't get either). I never did not allow this, I just never got around to it, because I was too busy trying to make another point clear. As should be clear by now, I'm not interested in artificial life. (Well, I am interested in it, but not as a part of the problem we are discussing now.) So I am not going to discuss those publications. My programme is not about artificial life, it is about evolution. And about a quick solution to the Travelling Salesman problem, which falls into the category of the so-called N-P complete problems, meaning their solution takes an inordinate amount of time when approached with conventional techniques. The fact that an evolutionary approach yields such quick results, should also be an answer to those who argue that it takes too much time to arrive at the levels of complexity that we observe in nature. The example shows that the time argument is not a problem for evolution.{edited to correct a typo}This message has been edited by Parasomnium, 05-09-2004 03:58 AM

---

"It's amazing what you can learn from DNA." - Desdamona.

WK, thank you for your reaction, I was beginning to fear for this thread. There's no need to answer every point I make in what follows, but feel free to do so if you want to. Alas, your assumption is wrong. If my evolutionary algorithm would certifiably produce an optimal solution for an NP-complete problem, I would probably become fabulously famous, because that's one of the rather annoying aspects of that class of problems, it is almost certainly impossible to produce the optimal solution in an acceptable amount of time. And if anyone ever finds a quick way to the optimal solution of a particular problem of this type, they would have solved all problems of this type, hence the fame. This is the result of another characteristic of this class of problems: they are all interchangeable, there is a one to one mapping possible from each one of them to every other one of them. If you need it, there is an excellent explanation of NP-completeness here: NP-completeness - Wikipedia. For now, suffice it to say that finding the optimal solution for the Travelling Salesman problem for any non-trivial number of cities requires an inordinate amount of time (like 'longer that the universe exists').So, as you may have guessed by now, an evolutionary algorithm does not produce the optimal solution. But it can produce a 99-percent solution, and in a very short span of time. And that is also what we see happening in nature. The solutions nature's evolution comes up with are generally very good, but hardly ever optimal, and sometimes even a bit strange. That they are very good solutions accounts for the fact that creationists often proclaim to see design in them. But, for example, our eye having a blind spot because the nerve layer is on the wrong side of the retina, or our oesophagus and windpipe crossing each other, sometimes resulting in death by choking, can hardly be called optimal solutions, can they? But optimal or not, for such sophisticated, seemingly designed solutions, they were arrived at in remarkably short, 'reasonable' time spans, contrary to creationists' often heard objection.In short, the parallels between an evolutionary algorithm running in a computer and nature's evolution have led me to conclude that the process of evolution, stripped down to its bare essentials, can be applied to non-living things. You have a very valid point there. I will talk about the life matter shortly. But first, let me stress that when I mention the term 'evolution' (in my statement that "there isn't a difference between simulated evolution and real evolution"), I don't stricktly mean the biological version of evolution as we see it happening in nature. We can never hope to simulate nature's evolution to such detail that viable conclusions about the details can be drawn from the simulation, it is much too complex for that.What I would prefer 'evolution' to denote in this discussion is a general, more abstract principle, more or less like what Ned proposed: For me, in simple terms, the three pivotal elements of a useful definition of evolution would be:

• change in the outward appearance or behaviour of whatever is under scrutiny
• random mutation of whatever it is that (wholly or in part) determines the outward appearance of the thing under scrutiny
• non-random selection of more successful members of the 'population'

Since this definition does not single out nature's evolution as the only possible example, I should have said "there isn't a difference between a simulated evolution and a real evolution", or perhaps even better "a simulated evolution is a real evolution", in the sense that the essence of evolution (those three pivotal elements I mentioned above) would be present in the simulation. And, above all, there would be real changes noticeable, in that the memory of the computer (where all the information about the simulated evolving 'population' is stored) would really change under the unfluence of this evolution.Now, about life. I concede that, for a given definition of life, a simulated life-form could fulfill the necessary criteria demanded by that definition. So, I agree a definition of life is needed in the end if we want to discuss the possibility of evolution of non-life. But if you look into the example of the Travelling Salesman problem, its solutions (itineraries!) certainly do not qualify as 'life' for any meaningful definition of life, yet one can use an evolutionary algorithm to arrive at a reasonable solution. To put it in simple terms: you start with a 'primordial soup' of mostly very poor solutions (random itineraries) and, through an evolutionary process, very good solutions (very efficient itineraries) bubble to the surface.I realise I still haven't given you a definition of life (I'm still thinking about it), but for my purposes the exact boundary between life and non-life is not very critical because I think I've shown that even very clear non-life (meaning it does not fulfill any plausible criteria of 'life' at all) is susceptible to the essence of evolution.I hope this sheds some light on how I think about this matter.This message has been edited by Parasomnium, 05-11-2004 07:33 AM

---

"It's amazing what you can learn from DNA." - Desdamona.