Hate Speech Dataset

There's nothing that comes to mind that exactly fits what you're looking for.

Construction of a corpus of hate speech would force a researcher to walk on thin political ice. It requires a firm empirical definition of Hate Speech. If this definition can be construed to include what some could interpret as political disagreement, unnecessary and necessary fire and ire could be brought in the direction of a well-meaning researcher. Not to mention, the odious nature of the task of collecting this linguistic material.

Take for example, the very high profile debate over the location of Park51 (or Cordoba House). Some of the opposition to the placement of the Islamic center and mosque could possibly be construed as "Hate Speech". However, other opposition was more measured and (while I disagree with the position) I wouldn't go so far as to call it "Hate Speech". I, for one, wouldn't want to be in a position to draw that line.

That said, there seems to be some research on "Hate Speech" or at least hateful speech. Here's a paper on swearing and abuse in British English which suggests that the Lancaster Corpus of Abuse exists. Then there are a few more linguistic oriented papers which don't necessarily have associated corpora, including this and this.

This may not be the most satisfying answer, but maybe it'll get you in the right direction.

i have hate speech data sets, but they're proprietary, with ordinal labels corresponding to the demands and restrictions of brand advertisers. We can talk about sharing, but we'll need to talk about what it is exactly that you're doing to do with it. please email me [email protected]

Hi Josh,

I tried to e-mail your stern address, but it came back as undeliverable. Is there another way I might contact you? Thank you for your help.

M Holtz

oops, my email is [email protected]