How Capitalism Changed American Literature | Hacker News Reader

How Capitalism Changed American Literature

publicbooks.org

93 pointssamclemens16 comments

This is fascinating. For those that want to save time reading, the author used a text classifier and had a 70% accuracy rate predicting if a book was published by a conglomerate or an independent publisher.

Which means there is a distinct difference between the type of content that indies will publish vs conglomerates.

Whether this has an effect on literature is what the rest of the article is about.

Was also surprised by the machine learning twist. A good read, with a promising ending for new research.

All his "conglomerate" texts are from Random House. He may have simply measured Random House's editorial style. Other publishers may differ.

Random House has a house style, and guidelines. Here are the guidelines for their Penguin division.[1] Random House accepts submissions only through established literary agents.[2] So they have an organized filtering process.

[1] http://authornews.penguinrandomhouse.com/what-our-editors-lo... [2] http://www.penguinrandomhouse.biz/manuscripts/

> the author used a text classifier and had a 70% accuracy rate predicting if a book was published by a conglomerate or an independent publisher

Genuine question: is 70% actually a very successful rate? It's hard for me to tell without more information; if 70% (or 30%) of books are published by independent publishers, then 70% would be the same rate as just guessing the more common one every time.

70% feels pretty awful to me when you don't even know what it is you're measuring.

It depends on your definition of success but 70% accuracy for classification problems with 50/50 odds is at least useful. In this case, "guessing the more common one" would not be a problem as the data were evenly split between indie and RH.

Could it be something like a classifier having success at telling if someone is a professional NBA player vs a club basketball player, or is there something more to it?

Or we could just say that any system of economic relations in order to finanace books and produce will cause distortion in litrature. Its a comparison to a nirvana that never existed and will never exists. And the difference he explains can be argued about in many different ways, and how it differs from the 60s can also be explained in a number of different ways.

Seems to me the authors was trying to find something that confirmed his priors and he did and then used that as confirmation for a whole larger theory about how he sees the world.

Article cities this essay by the brilliant Zadie Smith -- though sadly behind a pay wall:

https://www.nybooks.com/articles/2008/11/20/two-paths-for-th...

Smith has read aloud several of her short stories in free podcast form at The New Yorker. All are cunning social satire, wielded against on a wide range of deserving targets.

https://www.newyorker.com/podcast/fiction

Interesting read, I have a hunch we'll get similar results if these models were applied to contemporary music as well.

The author is basically taking a situation open to interpretation and using it to slam capitalism as producing mechanical literature?

His conglomerate/nonprofit dichotomy is more accurately free-market vs. state-sponsored. The "conglomerates" which come off as evil in the story have to prove their value to their customers who voluntarily fund them. The "non-profits" are free from the constraint of needing people to actually want to read their books because the population is forced by the government to fund them via taxes.

There are very interesting questions of how economic forces affect art and culture but this author isn't asking them.

If anything the machine learning model showed that when people voluntarily and intentionally fund the publication of books they prefer a certain type of language, but how this author sees this language as "mechanical" is beyond me.

Yeah - sadly it seemed "experiment tailored towards conclusions" as opposed to analyzing the reasons behind it. Especially if they had ones published by one author in one framework and one in another to compare.

The supply side and logistics could make it "mechanical" because it is known to sell well but much of it seems to be anti-popular rhetoric that anything widely liked must be flawed.

While there is a lowest common denominator drive from sheer weight of numbers (most money per person, targetting mean and below average getting good money per effort - frankly the higher brow crowd demand more effort) that doesn't mean they are automatically inferior nor those outside of its framework neccessarily advantaged. The converse doesn't mean "unpopular" are automatically better. Really popularity isn't that great of a heuristic either way.

To get to a larger point qbout writing everyone produces to some sort of motivation or incentive. Even "pure" writers doing it not for making money have at least an emotional connection - let alone attempts at gaining status or promoting ideas.

>If anything the machine learning model showed that when people voluntarily and intentionally fund the publication of books they prefer a certain type of language

This is really the springboard into what you just mentioned - economic forces affect art and culture but they also affect our desires. Exactly these "voluntary" and "intentional" purchases come about through what Patrick Murray calls imaginary subsumption under capital. This is a consequence of what the author addresses - real subsumption under capital. You need to ask why people prefer books that use a certain type of language and how the culture which prefers this type of language arose. Of course the answer isn't wholly economically deterministic, but it seems short sighted (and missing the author's point) to stop the back at "well people did it voluntarily, what more do you want?" - culture is a dialactical process.

Why state-sponsored? Lots of non-profits are funded by voluntary and intentional private contributions.

>how this author sees this language as "mechanical" is beyond me

Are you saying that you see no difference between the two sides of the word table, that the left side is not more perceptual, or that mechanical is not the opposite of perceptual?

Reply on news.ycombinator.com