Monday, February 14, 2011

Ngram Viewer sucks, true dat

proof positive....


true dat...

3 comments:

Brandon C. Loudermilk said...

First image link is broken... don't know anything about Ngram viewer - but you can't diss a tool based on the data it is accessing; And unfortunately, although corpora are much better (bigger, more specialized, utilizing spoken, not only written text) than they used to be (e.g., Kucera & Frances 1967), they still pretty much are lacking in many regards.

Chris said...

Hehe, I love the Ngram Viewer, of course. It's a great PR move for corpus linguistics (similar to Resnik's recent Language Log comments about IBM's Watson competing on Jeopardy: it helps get the incentives right).

But this was more or less a nod to a Twitter exchange with a grad student, not meant to be serious. I should have made that more clear.

Joe Kessler said...

Heh. Chances are this is just reflecting occurrences of a construction like "It is true that..." But thanks for investigating the issue! =P