Probably this site is familiar to most of you, but I've only just come across it. Great idea, eh?

Hmm...

I wasn't suggesting it's a good substitute for the usual method of writing poetry, of course. But if the method were ever to produce a reasonable poem, which seems not unlikely give the results so far, the implications would be rather interesting, for poetry as well as genetics. All those ideas of the poem as something that happens to you (the muse, inspiration, Eliot's idea of the poet as catalyst, postmodern ideas of the language speaking the speaker instead of the other way round) rather than something you consciously create wouldn't look too far off the mark. But of course, these poems are in a sense being composed, by thousands of people working in collaboration. It reminds me of a whimsical statement by John Ashbery:

"Research has shown that ballads were produced by all of society
working as a team. They didn't just happen." ('Hotel Lautreamont)

That's why I said "hmmm...," rather than "what crap." But of the examples I read on there, at least one seemed to be worse than both the ones it was made of. It's a funny idea, though. Love the Ashbery quote; I'm kind of warming to him in my old age...

Performing poetry from memory has had a slightly Darwinian consequence for me. Sometimes I recite a poem and some smartarse in the audience follows along by reading a book of mine at the same time. They point out afterwards that I didn't recite the same poem "word for word". I'm always surprised to hear this because I feel that I've just performed the version that I've always performed, sure there might be the odd bit of intentional tweaking that I can immediately recall and justify.
It's only when I look at the printed version, usually a few years old, that I see that the poem has changed over subsequent "generations", usually into a better version. I guess one could look at each recital as a generation of the poem, and the odd accidental adjustment as a copying error. Any obvious errors will be noticed and never repeated, therefore not successfully passed down to the next generation, neutral errors might be passed down or not without me really noticing, and positive copying errors will be noticed as an improvement and intentionally carried down into the next generation. This is probably closer to the "unconscious" or "artificial" selection that Darwin mentions in the beginning of the Origin rather than full-on natural selection, but I still find that there are plenty of interesting parallels.

Am I the only one who seems to automatically read "cellardeller" as if it were "cellard weller"?

I was that smartarse!

But I have to agree, after reading in public I always go back and 'edit' the written poem to reflect the latest subtle changes, which in many cases are suggested by the interaction with the audience.

I wouldn't mind just changing it to Niall O'Sullivan, but I don't know how to do it. :unsure:

I'd love to know where you got the "w" from though...

One of many, Dom!

What makes it particularly Darwinian is not so much that I notice faults when performing and go back and change it, but that the mental facsimile of the poem seems to change a little each performance without me noticing.