I've been meaning to write this one for quite a while now, so here goes. It has to do with the amazing tools available online for scholarly research. For instance, try this search for "occult revival" on Google Books. Do you have any idea how long a search like that used to take when the only tools were paper-based card catalogs in physical libraries? Not to knock libraries, but the answer could be literally: a lifetime.
Note that "occult revival" is not just some random pairing of words. The phrase is a term of art in studies of 19th century intellectual history. I didn't know this a year ago -- even though I'd probably come across the phrase many times in my reading. The way I learned this what you might call "specialized language" was through some really dumb software on Amazon that looks for what the company calls SIPs: statistically improbable phrases.
I should explain that by "really dumb," I don't mean bad. SIPs are incredibly useful. As I said, they're how I learned that "occult revival" was a phrase in common use among certain scholars -- and critically, that those scholars had written books I would very likely want to read. What I mean by dumb is that the software doesn't know squat about occultism, revived or otherwise, or anything whatsoever about the 19th century.
The first clue to how SIPs work lies in the word "statistically." We're not talking some sort of AI here (not, at any rate, the sort that used to make insupportable claims to "natural language understanding"). Instead, these are mindless number crunching algorithms, and in this case the "numbers" they crunch are digital representations of words. What they look for is something computational linguists call co-occurrence.
The second clue lies in the fact that certain words co-occur all the time. Think of clichés. Here's a handful grabbed off a page cleverly titled Clichés: Avoid Them Like The Plague.
Clichés are statistically probable phrases. Which is why they don't convey much. I'll spare you a long digression here on the information theory of entropy, but trust me, it's germane. Now, if you had some way to do it -- say a huge corpus of text data and a super-slick co-occurrence algorithm or two -- you could identify such nearly meaningless phrases "simply" by tallying how many times they... um... occurred. And here's the dumb-but-good part: you could do this without needing clue one about meaning.
Happily for us, giant text corpuses and slick search algorithms are just the sorts of things that companies like Google and Amazon happen to have lying about. And if you can identify clichés, you can also identify non-clichés -- call them Statistically Improbable Phrases.
As it turns out, the phrase "occult revival" doesn't occur anywhere near as frequently as, say, "Hot enough for you?" or "Have a nice day!" But it does occur. And it recurs in certain types of books -- often enough to be statistically significant. In fact, this recurrence defines, in some dimension, what this class of books is "about." More precisely, the algorithm creates that class on the fly. And when you see a part of that class laid out, as below, a lightbulb can go off. It did for me.
This is called -- get ready for a rare technical term here -- learning.
No, not machine learning. Human learning. As in, you come away from the experience with more understanding than you had going in. And as I said at the start of this, it could take a lifetime to acquire this sort of knowledge using pre-digital research methods. Yeah, yeah, maybe that sort of hard work would make us more deeply knowledgeable, but look: some of us are born amateurs and dilettantes (see my discussion in Gonzo Marketing: Winning Through Worst Practices) and we don't have enough lifetimes to waste getting PhDs in all the specialized areas we might want to know something about. Do we? No, we do not. And besides, universities cost too damn much and they're full of arrogant motherjumpers who think they know it all, when the truth is, very damn few know even half of it.
Uh... please to forgive the little rant there. Call it my postmodern manifesto. But perhaps this has all been a bit too abstract. Here, let's have an example. I yanked the following off Amazon and tinkered a bit with the URLs, all of which now should work (they do in my browser, anyway; got Firefox?). Be sure to experiment a bit with the extra cool pull-down references (right after the little arrowheads).
Learn more about Statistically Improbable Phrases (SIPs)
20 references in Gurdjieff; The Key Concepts (Routledge Key Guides) by S. WELLBELOVED
14 references in The New Encyclopedia of the Occult by John Michael Greer
14 references in Spirituality and the Occult; From the Renaissance to the Modern by B.J. GIBBONS
13 references in Occult Underground by James Webb
12 references in Surrealism and the Sacred: Power, Eros, and the Occult in Modern Art by Celia Rabinovitch
10 references in The Place of Enchantment : British Occultism and the Culture of the Modern by Alex Owen
7 references in A Dark Muse : A History of the Occult by Gary Lachman
6 references in Surrealism and the Occult : Shamanism, Magic, Alchemy, and the Birth of an Artistic Movement by Nadia Choucha
6 references in Occultism, Witchcraft, and Cultural Fashions : Essays in Comparative Religion by Mircea Eliade
6 references in Essential Golden Dawn by Chic & Sandra Tabatha Cicero
6 references in Belief Beyond Boundaries: Wicca, Celtic Spirituality and the New Age (Religion Today-Tradition, Modernity & Change) by Joanne Pearson (Editor)
5 references in Blast Your Way to Megabuck$ With My Secret Sex-Power Formula by Ramsey Dukes
4 references in Helena Blavatsky (Western Esoteric Masters Series) by Helena Blavatsky, Nicholas Goodrick-Clarke (Introduction)
1 reference in The Occult by Colin Wilson
1 reference in Modernism and the Celtic Revival by Gregory Castle
1 reference in The Magus: A Complete System of Occult Philosophy by Francis Barrett
Search A9.com for "occult revival"
Before publishers start having heart attacks -- "look at all our stuff this guy is stealing!" -- let me hasten to say that, of the above list, I purchased all the titles you see pictured in the right column (I managed to resist Blast Your Way to Megabuck$ With My Secret Sex-Power Formula). And I purchased them, in many cases, because this search pulled them up.
Finally, this is why I buy from Amazon. I learn there. And I learn there in a way I never could at a walk-in Barnes & Noble store -- or even on their site. As to all the poor Mom & Pop bookstores Amazon has supposedly disenfranchised, give me a break. Amazon has given these stores (any of them that wants it) a global market reach they never could have developed on their own. My only gripe is that this has created a virtual market for used books, the prices of which, as a result, are floating -- some soaring -- ever higher. Bad news for me. Great news for Mom and Pop. And of course, for Jeff Bezos -- to whose laudable vision and maniacal anonymous coding teams this piece is gratefully dedicated. Even at today's <koff> elevated prices, Amazon is still way cheaper than a couple Ivy League sheepskins.
posted by Christopher Locke at # Tuesday, April 11, 2006