Slashdot Mirror
Darwinian Poetry: From Bad to Verse
For those who say design cannot take place through the process of selection, behold:
Darwinian Poetry.
Cull the prosaic or nonsensical snippets of text, reinforce the rest, and, slowly... genius? Guess we'll find out. Yes, the poems actually
have sex.
I say interesting and the more I played the more I realized that this is in many ways better than the stuff that I had to read in english class.
Another advantage is that no teacher could ever ask;
What was the authors motivation in writing this particular poem?
I hate that Question
I used to have a cool sig, back when I cared
I rather doubt any good poetry will actually come out of this. It seems to me that two good poems with parts interchanged at random "snip points" will be statistically very likely to become bad poems. A more advanced system is probably necessary before anything worthwhile will be produced.
The idea of having people vote on which poems are best is a good one, though. Maybe the same principle could be applied to other computer-generated word stuff.
To qoute; The goal of this project is to see if non-negotiated collaboration can evolve interesting poetry using (un)natural selection.
Darwinism is all to do with natural selection, while this is un-natural selection. It's about breeding poems, nothing more. That aside, I must say I find the idea interesting, and the end result can't be worse than what a lot of modern poets spew out (these days, it seems like "art" is defined as what the selfproclaimed artist manages to sell).
For a true darwinistic approach thought, it ought to be possible do analyze a heapload of poems written by humans, derive a handfull of rules as to what defines a 'good' poem (lenght, avrage lenght of words etc etc etc) and write a program that 'culls the herd' strickly on basis of those rules, ie: the 50% of the population which come closest to fullfilling the rules (best adapted to their enviromant) are allowed to breed and give rise to the next geneartion, at which point the process repeats.
Everything in the world is controlled by a small, evil group to which, unfortunately, no one you know belongs.
No, it's human interpretation the makes something poetry.
Computers/processes are quite capable of producing works we percieve as art.
Your hybrid is not saving the environment. Its purpose is to make you feel good about buying something.
> Evolution is a religion. It is a set of beliefs.
I believe the sun will rise tomorrow just about the same time it rose today; is that also a religion?
> Most evolutionists say things like "we have reason to believe", or "we believe that foo is x years old". It is still called a "theory", not a proven fact or scientific Law. Actually it is mathematically improbable even.
Actually, if you put imperfect replicators in a rich environment evolution is almost a certainty.
> Just like the early church, the evolution religion changes its views on matters of "fact" and change the timeline and tree of life to fit in with their new findings.
That's a Prime Directive for science: if your model doesn't fit the facts, you have to keep the facts and change the model. That's how science makes progress.
> Those who don't adhere to the beliefs are excommunications and sometimes attacked and discredited. Just ask any creationist with a Ph.D.
That's not excommunication, that's "bullshit walks". Creationists are welcome to submit their articles to the same peer review process that real scientists are. How many do you know of that do so, and what were the reviewers' comments on the rejection notices?
Conspiracy theories are the last refuge of kooks.
Sheesh, evil *and* a jerk. -- Jade
By referring to people who believe one thing as "real scientists" you have just excomunicated them, yourself. This is precisely what he means. Papers referring to intelligent design are mocked in peer review because they do not think of such as "real science." You have just given an excellent case in point.
I'm a scientist. The reason I have a problem with the idea of "Intelligent Design" is that it doesn't have predictive power. Science is all about explaining what's going on in the universe - and the main way we test whether we understand something or not is to use a theory we've developed based one one set of evidence to make predictions about experiments we haven't tried yet.
Evolution has passed the test of time because it explains the existing data and one can make predictions based on it that tend to be surprisingly true.
Intelligent Design is one equally valid way to explain the existing data. But by definition it doesn't give us any idea about what to predict regarding new data. It in effect tells us "stop trying to understand it, it's like this because somebody made it this way".
I'm open-minded enough to entertain the possibility that there might be a Creator who designed all of the creatures on the Earth. But there sure is a lot of evidence that says that the creatures evolved, and are continuing to evolve, and in quite dramatic ways over enormous timescales.
If one could evolve poems, one could do the same with all other forms of written human language, since poetry seems to be the hardest of them all. And you would have AI.
It is obvious that even the code for this evolution would have to be evolved, and program that would do this evolution could be breed too, etc, etc, up to some pretty simple program that will start it all.
This is the missing key: When you evolve programs, they must include both the code that produces results, and code that evaluates those results. This way your results have no ceiling and can surpass humans.
If human intelligence is used to judge the results, then how can we get anything beyond human intelligence (apart from a faster human intelligence)?
Of course this will work, but all it will prove is something we've known for a couple of thousand years: Selective breeding will bring out desired characteristics.
What this study does NOT address:
Irreducible complexity. We already have the groups of words. Well, where did they come from? How do we get the group of words in the first place? We can't do selection on the words until we have the words, so, how do we get the words?
Intelligent design. Intelligence (namely the humans running the model) is determining what we start with and is determining what the desired results (what constitutes acceptable survival),
Cost of mutation. There is no attempt to factor in mutational "drag" if you will. All mutations are either considered neutral, or beneficial. The reality is, most mutations are HARMFUL. Any mutation which does not directly improve the organism, will almost certainly harm the organism, greatly increasing its chance of death. If the mutation rate is too high, the species will die out (known as Haldane's dilemna).
Informational Loss. Nearly all mutations result in a LOSS of information, in this case, the elimination of a word. Once the word is gone, how will it ever come back?
So, this little exercise is nothing more than a cute gimmick that blind adherents to evolution as the source of all life will point to, smile, and say: See you idiot creationists, one more thing to prove your stupid, unthinking mindset wrong.
But the reality is, it won't prove or demonstrate anything other than the time-tested truism that trial and error will eventually get you what you want.
You can tell a great deal about the character of a man by observing those who hate him.