automated his and hers word detection

file this one under time waster because you’ve been fighting CSS HTML and CMS macro code for three straight days and you’re siiiick of it

The english language doesn’t have masculine and feminine words like Spanish or English (except for boats and nations I suppose) but alledgedly some words are more likely to be written by men or women. For instance, these words are feminine:

[with]
[if]
[not]
[where]
[be]
[when]
[your]
[her]
[we]
[should]
[she]
[and]
[me]
[myself]
[hers]
[was]

and these are masculine:

[around]
[what]
[more]
[as]
[who]
[below]
[is]
[these]
[a]
[at]
[it]
[said]
[above]
[are]
[the]
[many]

Sceptical? Try it out: the Gender Genie. Enter 500 words of text, and this funky little algorithm developed by some academic types will tell you if the author was male or female. I can’t decide if it’s accurate or not.

Although it was often right, it blew it on a couple of Naomi Klein articles. Is there something Ms. Klein isn’t telling us? On the other hand, it tagged Meg Hourihan talking about baseball. But then juuust lost her when I added in her thoughts on XHTML scripting. Okay, I was pushing the system. On the other hand, it caught Raymi the Minx writing about percodan with no problem, even though I had trouble deciding wether to submit it as fiction or nonfiction.

Here’s an interesting challenge: finding enough 500 word passages, written by females, on the internet, to test the system.

leave a comment