On German Language & Usage we recently had a question about the English word schmuck, which is a false friend of the German word Schmuck.

Out of curiosity I tried the capitalized form vs the non capitalized form in Google ngram.

Being astonished by the high number of hits for the Capitalized form I retried this with the more innocent noun house.

I do not understand the results. I think all the words a normally written in the non capitalized form, except when used in titles or at the beginning of the sentence.

All this looks rather arbitrary to me and I have the impression that I would choose a case insensitive option if only I had a chance to do so.

Is this just another shortcoming of Google ngram or is this a problem for other ngram viewers too, and I have to care to set some options explicitly?

  • 3
    This seems like a more appropriate question for google. They have forums where you can ask them. (i.e. it is off topic here). There's no linguistics in this question.
    – Mitch
    Oct 18, 2011 at 13:41
  • @mitch You mean 'regular' ngrams at least have the option to work case-insensitive?
    – bernd_k
    Oct 18, 2011 at 13:46
  • 4
    I agree with @Mitch. This is a question about how a certain software application works. Incidentally, that application may be used for linguistic purposes. Voting to close. Oct 18, 2011 at 13:58
  • Sounds as if I have to wait for 'Natural Language Processing and Computational Linguistics (NLP/CL)'?
    – bernd_k
    Oct 18, 2011 at 14:00
  • 1
    @bernd_k: You can try asking on webapps.SE but I fear they will only tell you - Google didn't put in such a feature so go ask them for it in their support forums. Oct 18, 2011 at 14:05

1 Answer 1


After doing the search for "Schmuck,schmuck,SCHMUCK,sCHMUCK", which should return a graph that looks like this: enter image description here

Then you should just be able to check yourself by clicking the "download here", which is within the phrase "Run your own experiment! Raw data is available for download here."

On that page you'll find instruction on how to download the data and a link to a total counts file.

After you have those files, it'd be easy enough to do a quick check on the result counts reported versus those within the text provided using something like a GREP command.

That said, believe you'll find that there's not an error in the graph, and would venture a guess that the phrase "Schmuck!" is not all that uncommon.

Not the answer you're looking for? Browse other questions tagged or ask your own question.