I was entering random names in Google Ngram when I noticed that in the late 80s to early 90s, there is a bump in the occurrence of lower-case European country names. See for example:

united kingdom: http://bit.ly/16srJju

italy: http://bit.ly/1fGCAHG

Something similar happens for most West-European country names. A search for "usa" gives a slight bump too, but for china or russia it's absent.

Upper-case names of countries show no bump.

Does anyone have an explanation? Also: how reliable is Google Ngram?

  • Hmm it's kinda interesting but I'm not sure it's a linguistics question. It might be better asked on webapps.SE Sep 26 '13 at 10:44
  • 4
    The frequencies involved are very low. Perhaps it's due to a single book where upper-case letters were not recognised correctly or that for some reason didn't use them? @hippietrail It's a question about corpus usage and how to interpret result based on corpus queries, so I think it's on-topic.
    – robert
    Sep 26 '13 at 12:22
  • I do think corpus questions are on topic but this is more like a corpus meta question. But I have the same theory as you. One or two thick items published that year with lowercase country names for some reason. Perhaps even scanned by Google more than once by mistake. Anyway the data is noisy and we can probably only speculate on the reasons. Sep 26 '13 at 12:26
  • I see, thanks for the comments. I'm not a linguist and I don't even know what "corpus" means as a field/category, so apologies if the post was somewhat off-topic! So a follow-up, perhaps even more off-topic: Is it possible to extract the sources of hits in Google Ngram?
    – matimo2
    Sep 27 '13 at 11:10

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.