Poetry Forum

Full Version: Chess and poetry
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
The Cuban grandmaster Jose Raul Capablanca predicted in the 1920s that chess would be dead in no time, and advocated for an expanded board, with new pieces.
In the early 20th century, Max Born told a group of physicists visiting Gottingen than 'physics would be over in a few years' (I got this from A Brief History of Time)
Neither Capablanca nor Born were correct, though Capablanca was closer to the truth.
Today, physics is no closer to being 'solved', given the onion like nature of reality, but chess is indeed closer to being 'solved'. At least, we have computers that play better than the best humans.

Coming to poetry - it is, after all, a craft following rules of intuition. Rules of intuition can be mocked through machine learning. It should be possible to feed a computer the best poems of the past 50 years and see it spit out a poem in the style of John Ashberry.
Or maybe it has happened already.
There's a difference between "closed" and "solved."  Tic-tac-toe is closed and solved; I believe checkers is, as well.  The number of possible chess games is large enough that it hasn't yet been solved, only better-played by machines which can look ahead further than humans.  But chess is closed.

Baseball, however, is open.  Hmm.

Poetry isn't closed, I think, partly because there's no scoring system possible:  what constitutes a poem is arbitrary, and once you decide what's a "poem" and what isn't you're left with judging which poem is better.  And the judgment is made by the reader, not the poet, which turns poetry into a sport... or possibly an economic transaction.

It may not always be possible to determine which poems are machine-generated and which human-created.  But there will always be a ghost outside the machine which says, "This one's good."  And the ghost won't care who or what produced it.
(05-31-2021, 12:12 AM)dukealien Wrote: [ -> ]  Tic-tac-toe is closed and solved

When I discovered the secret of 'always block the central square' as a 7 year old, it was like finding the solution to the Reimann hypothesis...
I can't imagine a reader being indifferent to whether a poem was produced by a machine or a human.  Unless the reader is also a machine.
It can be "mocked", but at a certain point you gotta wonder -- what if the computer actually has intuition?

Another way to look at it is that the computer will still end up having been designed by a person, and so that person becomes the poet.

Part of the value in poetry seems to be that it's made by people. It's a human connection. Computers can and often do mock humanity nowadays, and plenty of people are fooled by them, but that hasn't stopped people from interacting with each other online or publicizing themselves and their art.

On the other hand, a lot of art is increasingly being valued without consideration for its humanity, not so much with whether or not it's made by a human as whether or not it satisfies this or that basic need. Art valued purely for spectacle, for melodrama, for political expediency -- art whose value is measured by dollars. Computers already satisfy those individual elements. So, too, lesser or still-improving artists. And in an age of semiliteracy and squandered democracy, fewer and fewer people are willing to consider the sum rather than the parts.

Then again, the people who do consider the sum have always been few. Great artists recognized as great during their time more likely survived because one or the other element seemed to shine brighter than everything else. Most people worry about base survival too much to consider art more deeply, while the rich have always been representative when it come to number of rubes. It seems to me more a function of history than any individual age to recognize those artists whose work is manifestly human. Though with anthropogenic climate change, rapidly developing diseases, and continuing access to nuclear weapons, who knows if we'll still have history in the future.
Might be getting off track here, but thinking about strategies, tactics, or tropes from other art forms...

A machine can certainly write bad Hemmingway.  Could a machine write good Hemmingway?

A machine could paint uninspired but accurate sfumato.  Could a machine paint good sfumato?

(The machine would not be in danger of making mistakes in the identified effects so long as its programming was correct.  A human would.  Is that the human advantage, or just why human effort touches us?  The spice of danger, appreciated by the viewer/reader?)

And beyond that, could a machine invent a style at the level of Hemmingway or da Vinci?  Even if its explanatory/defining examples of the style were not very good, but could be surpassed excellently by others?
Good is subjective. If the standard of good changes to something a computer can generate then it will have generated something good.