Six Degrees of Unsophistication
Today I saw a web page devoted to word association. It offers you a word and asks you to respond with another. Once you do, it gives you another word, and the process continues as long as you like. Whenever you get tired of it, you can move on over to the list of the most recent words you’ve seen, along with your responses and a count of how many people responded the same way. Clicking on any of these words (queries or responses) moves you to an interactive network of bubbles connected by dark lines, a visual representation of the tiny fragment of the network surrounding that particular word. The bigger the bubble, the more common the word; the thicker the connecting line, the more responses received associating two words. You can click any of the associates to see what it connects to in turn, and so on, although the bubbles fade out over time, to keep the screen legible.
The network is necessarily connected because it started with the single word “volcano” and built itself up recursively. Every word (other than “volcano”) to which you are asked to reply is somebody else’s response to a word already in the network.
I’m pretty sure this has been done before, probably several times. It’s easy to implement, and builds itself up in no time. This particular example could use some more depth of graphic representation; after three steps, your starting word has faded away, which doesn’t leave much room for intricacy, or for surprising loops. Some of the associations are surprising, but mostly in the context of how few steps it takes to get from one word to a wholly unrelated word. “Donut” to “Submarine” in two steps isn’t very interesting—donut to sandwich to submarine is an obvious route—but doing it in three steps through “Homer Simpson” and “yellow” is. The five-step round trip among words that have nothing to do with words two steps away is cool, but since only a very short tree is shown at once, you have to rely on your memory to discover them.
Unfortunately, interesting links are few and far between. This is partly because most links are prosaic: asked to respond to “cow,” you might respond “bull,” or “pasture,” but probably not “purple,” even though “purple” and “cow” do have a direct connection, thanks to the poem, and ice cream fountains. It is also partly because the engine is not sufficiently selective. Whole phrases don’t belong here, nor do misspellings. But the sparseness of interesting links is due primarily to the paucity of the participants’ imaginations.
As I learned by going to the top page, the number one entry is “sex.” This does not surprise me, as, two steps from my first entry to the graphic on the word “whey,” I hit “fuck.” “Fuck” has a lot of connections. A lot a lot. And that’s not counting variations like “fuk,” “fux,” and “fuxxor.” Random key bashings were common, too: “asdf,” “giojsl,” and “feije.” At times, it seems like the bulk of the free association word web is the product of a Tourette’s Syndrome support group. The query-response engine includes a button to eliminate non-words and bad spelling, but enterprising editors can’t keep up with the potty-mouths who think typing “fuck” into the box is hilarious. When it comes to studying human intelligence, even informally, somebody’s always willing to spoil the party. And that’s a shame.
I’d very much like to see a more representative description of the words people thought of. In one sense, of course, lots of people were thinking the words “sex” and “fuck,” but that’s not what I mean; I mean I’d like to see what thinking people were thinking. This isn’t it.