Messing around with some computational language tools, I generated this list of words which are more frequent on this blog relative to a standard corpus (some misspellings removed), in order from most overused. Many of these are unsurprising, but I had no idea I used "cannot" more than is normal. Or "parasitical", which is more worrying.
cannot simpleminded parasitical excoriate delegitimize kvetching temperamentally treacly politcs cosmopolitans authoritarians twitter rightwingers inexpert constructivists constructionists entertainingly clathrate undesireable frenzies mystifies wastefulness repurpose gintis wobblies kunstler turmoils bukovsky bankrolls laitin smidgeon sociopaths scienceblogs cleavon oddsmaker vegetating reifying situationists doper yecs popularizer nobels cultish solidary arduino militarist prolixity congealing proft larded atran nixonian seatmate appeaser rationalists leftish libertarianism literalist materialist vitalism rejoinders schuon fusty facebook torahs arduously hugeness universalizing tinkerers factuality autoworkers parasitize rationalist dominionism physicalist incarnating idiocies axiomatically ferreted gourevitch glaringly symbiote averagely incisively shitheads skimped netzach appall metonymic onrush chokehold halldor churchy scampers starkest agentive dalliances emet mistimed ceasefires hallucinated reimagined overplaying bioethicist copleston disempower flippancy oversimplifies outrageousness indvidual ginned douchebags explicates plumbs mencius metaphysically schelling foregrounding polarizes outlives subtexts acquiesces nostrums undescribable malkuth marketeer analagous preeminently remediable flamers slipperiness bunraku proles burkean peaceniks materialists unaccountably athwart mcworld petraeus romanticizing unnamable huffpo ineffectually commonsensical interoperating empathizing wingnut supplicants hypostasis inchoate obama transhumanists fulminate affordance nonviolently geneological gashed mussed chuppah charnel felin reconstructionism verbalizing tegmark crabbed armys shalizi dehumanization hoohah vannevar copyable bungler unlikeliest preindustrial legitimated downscale fugs bilin slavering egomania naveh determinedly oligarchies chasten reappropriated bekki taleb bioethicists valdis ultraconservative wahabi straussian rewatch anthropomorphism ecstasies libertarians ruination exceptionalism vacillate overreach forthrightness informationally bushites rottenness biomorphic parceled twittering sorley parapsychological irreligious statists maddeningly selfing militarists bushite infuriates deconstructionist dallying harrows glutted worths misplacement engross jewishness hearkens girdled zombified prohibitionist braf sniggering positivists prostrating doomy schmaltzy yesod hewing philosophize doomsayers unconcern conflate jibes misappropriate convulse constructionist relabeled cavalierly mesmeric phantasms atrophied nattering reductionist personhood asocial placating incuding amorality incontestable weida greybeard inescapably scrabbling foreordained puthoff antiabortion commandeering iphone reinterpreting fudges minsky spluttering obsessional explicating rovian subdues ascription graeber counterargument plops
Now I'm playing the Burroughs-ish game of trying to find meaning in this shredded language. "physicalist incarnating idiocies axiomatically" sounds applicable to a number of discussions I've been having lately.
3 comments:
This is interesting. I clicked on 'reconstructionism' due to something half-remembered and found much more interesting than average (by my peculiar standards) articles on L'Affair Forrest/Beckwith and one starting off with "Ron Paul is outed as a racist, anti-semitic, homophobic, conspiracy-mongering piece of shit. No surprise really, but a lot of nerds are hurt and confused. Add on his support for medical quackery, denial of evolution, and ties to Christian Reconstructionism*"
Would you mind saying a little about "computational language tools" and how used? This really interests me. One thing I'd really like is a computational language tool that would take in Von Mises _Human Action_ and output a cogent argument something "this doesn't smell right, and has certain definable (and worrying) things in common with Das Kapital.
If you have a favorite place to show for open source computational language tools I'd be much obliged.
Let's see, 'parasitic' ==> 1st item'
"Hostile AI: You’re soaking in it!" Certainly one of my favorite titles...
"Reconstructionism" confusingly means both one of the most right-wing forms of Christianity and one of the more left-wing forms of Judaism.
These were very simple tools that I wrote myself, in Clojure. I can share the source if that is useful, but it's pretty basic stuff (so far). More interesting NLP tools can be found here, the base frequencies I used here.
I take it, based on description at http://googleresearch.blogspot.com/2006/08/all-our-n-gram-are-belong-to-you.html
that counts in count_1w.txt are of how many times in 1,024,908,267,229 words a given word appeared.
Post a Comment