I have two questions.

If I were to use diachronic corpora to investigate the frequency of usage  for  words within  a word class of English that is mostly archaic today, how do I determine if the archaic words  were  ever considered to be a  part of everyday English?

In other words,  if the pronominal adverbs "therefore" and "whereby"  show up in an Ngram Viewer  analysis as  having a 1/100ths  percentile of usage,  but other pronominal adverbs that typically only appear today as legal jargon have a 1/1000ths  percentile of usage,  how can I determine if they were ever used in everyday English?

My second question is  if it is possible to determine with accuracy when is the frequency of usage of a term so low,  that it is not considered to be apart of everyday language? 

Thank you in advance.

  • 2
    Frequency in written corpora is not a good test of frequency in everyday speech. The written corpora we have from the past are even more heavily skewed towards formal registers than today's are. Jun 18, 2018 at 7:19
  • 1
    You could divide your corpora into formal and less formal, and then project the diff of the frequencies to estimate the frequencies in even less formal corpora which we do not have, but some words might will be completely missing. Jun 18, 2018 at 7:22

1 Answer 1


Any cutoff you use is going to end up being arbitrary, just by its nature. So pick one that seems to work well (based on the hit counts of sample words in a modern-day corpus) and take that as a given.

For the main part of the question, though, the best way is to choose your corpus carefully. If you have a corpus of nothing but transcripts of legal cases, that won't give you a good sample of casual colloquial usage. A selection of letters, on the other hand, or certain magazines, or social media in the last few decades, will give you a much better answer to this question.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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