(I’m just pretending that this place is like Twitter once was.)

Ben Schmidt in particular has illuminated issues different from the ones we uncovered.

Some distillation of things here:

pdodds.w3.uvm.edu/teaching/cou...

More to add in the future.

Some of the (massive) problems in Google Books have been reduced as they’ve delivered new versions. But there’s still real trouble in treating the dataset as an evolving corpus that is year-to-year the same in the way it’s all been put together.

💛

Do you not not want to support this person?

I just came across this myself and had to really concentrate.

Have sent a (quixotic) request to have this remedied internally.