| Welcome to The New Coffee Room. We hope you enjoy your visit. You're currently viewing our forum as a guest. This means you are limited to certain areas of the board and there are some features you can't use. If you join our community, you'll be able to access member-only sections, and use many member-only features such as customizing your profile, sending personal messages, and voting in polls. Registration is simple, fast, and completely free. Join our community! If you're already a member please log in to your account to access all of our features: |
| Hey Moonbat | |
|---|---|
| Tweet Topic Started: Sep 6 2017, 07:49 AM (107 Views) | |
| Klaus | Sep 6 2017, 07:49 AM Post #1 |
![]()
HOLY CARP!!!
|
A question that has been on my mind recently. Maybe you have any insights: I see evolution as a kind of graph like this: Consider an (infinite) graph whose nodes are all potential genomes. There is an edge from A to B if B can potentially evolve from A in one step (one generation). Let's assume there was a first life form. Let's call this node "S" for "start". Now I wonder whether this graph has the following properties: 1) Is there a path from S to any other node in the graph that corresponds to a viable lifeforms? That is, are there potential lifeforms that would never be "found" by evolution, not even given infinite time? (In CS, search algorithm are often stuck in local minima) 2) For any two nodes in the graph reachable from S, is there a path between those nodes? That is, is evolution sometimes a one-way road or is it invertible? Edited by Klaus, Sep 6 2017, 08:49 AM.
|
| Trifonov Fleisher Klaus Sokolov Zimmerman | |
![]() |
|
| Copper | Sep 6 2017, 08:03 AM Post #2 |
|
Shortstop
|
Like a hard working, dedicated, sincere Congressman? |
|
The Confederate soldier was peculiar in that he was ever ready to fight, but never ready to submit to the routine duty and discipline of the camp or the march. The soldiers were determined to be soldiers after their own notions, and do their duty, for the love of it, as they thought best. Carlton McCarthy | |
![]() |
|
| Horace | Sep 6 2017, 08:03 AM Post #3 |
|
HOLY CARP!!!
|
Given that evolution is based on random mutations, the chances are non-zero that a sheep will give birth to a human. I wouldn't bet on it, though. Except in Great Britain. |
| As a good person, I implore you to do as I, a good person, do. Be good. Do NOT be bad. If you see bad, end bad. End it in yourself, and end it in others. By any means necessary, the good must conquer the bad. Good people know this. Do you know this? Are you good? | |
![]() |
|
| jon-nyc | Sep 6 2017, 08:08 AM Post #4 |
|
Cheers
|
For number one are you asking if the mutation could occur? Or if the resultant organism could survive? If the first it seems tautologically true given your definition of nodes as 'potential genomes'. If the second it seems obviously false. |
| In my defense, I was left unsupervised. | |
![]() |
|
| Klaus | Sep 6 2017, 08:37 AM Post #5 |
![]()
HOLY CARP!!!
|
If a resultant organism could not survive, it could not have any outgoing edges because it could not reproduce. The vast majority of nodes in that graph would be isolated because they don't correspond to viable life forms. But it is conceivable that there are potential life forms that could survive and be successful but where there is no path of incremental adaptations that leads to that life form. Or not? I guess it depends on the granularity of DNA changes. For instance, is there a non-zero probability that the offspring of a fruit fly is an elephant? (edit: just saw that Horace already posted the same idea) |
| Trifonov Fleisher Klaus Sokolov Zimmerman | |
![]() |
|
| Mikhailoh | Sep 6 2017, 09:25 AM Post #6 |
|
If you want trouble, find yourself a redhead
|
FIFY |
|
Once in his life, every man is entitled to fall madly in love with a gorgeous redhead - Lucille Ball | |
![]() |
|
| Moonbat | Sep 6 2017, 09:41 AM Post #7 |
![]()
Pisa-Carp
|
I'm finding it hard to construct an abstraction that allows me to think about such questions whilst remaining (relatively) faithful to nature. Infinite time is usually taken to mean any event with finite probability occurs which will mean I think we explore the entire space of genomes or it would if the space of genomes were finite. The fact your space of genomes is also infinite though means that even without anything about natural selection I suspect the problem is indeterminate. We also have the problem of sexual (or other non-binary) reproduction, which means you can't think of one node linked to one other node. We also have the problem that an organism's fitness depends both on the details of which other organisms are around and on the entire of history of which organisms were around (because life changes the environment). I think we can solve some of these problems whilst preserving the spirit of your question e.g. we could restrict ourselves to some finite subset of genomes (e.g. genomes up to the size of the Paris japonica genome). Instead of infinite time, we could assume any mutation whose probability is greater than some threshold value occurs, similarly, any genome whose probability of reproducing is greater than some threshold reproduces. We will further restrict ourselves to female/asexual genomes (sexual reproduction thus simply constitutes a change in the kind of mutations that are likely). Most possible genomes will still be non-viable so we will restrict our nodes to those genomes that if they occurred would have a chance of reproducing greater than the threshold. We're still left with life changing the environment (i.e. when we claim we will keep nodes that constitute organisms that have a decent chance of reproducing, a decent chance in which environment??) The above all assumes the environment is fixed and I'm not sure how to reason about your graph if it's not. If we pretend it is fixed then we end up with a genetic algorithm with a fixed fitness function and that is a well-defined problem (though we have thrown away a very important part of evolution by natural selection). For 'reasonable' threshold values the answer in the genetic algorithm case, will I think always be 'no' to question 2 because the population will get fitter as generations pass and less fit solutions will not come up again. (it will also be 'no' in real life - organisms well adapted to life on Earth 3 billion years ago are not likely to turn up again because the environment on Earth 3 billion years ago is not likely to turn up again.). In the formal case, I think the answer to question 1 is usually 'no', unless the landscape we are searching is very smooth there will be good solutions that we never find. In real life the answer will also be 'no' because (amongst other reasons) the events that occurred in the evolutionary past get frozen in, I expect there to be numerous organisms that are perfectly viable but have slightly different biochemistry (whereas basic biochemistry is highly conserved on Earth) they can't emerge because they would have to undo and then redo the many adaptations that have happened in the past three and half billion years. |
| Entia non sunt multiplicanda praeter necessitatem | |
![]() |
|
| « Previous Topic · The New Coffee Room · Next Topic » |








5:03 PM Jul 10