I've found that some questions seem to call for descriptivist answers. By that I mean the question is along the lines of "how is this word/phrase typically used?"

I can't think of how to write a proper (i.e. researched) answer for such a question, or even provide much of an explanation, beyond "This is how it's typically used," or "No, it's not used in that way."

Here is a recent example in which a user asks if the word "glee" is typically associated with the concept of "schadenfreude," without requiring additional context. I believe that I am correct in that it is not, but I don't know how one would answer this question authoritatively with research or explanation.

Is there a good way to answer such questions?

Edit: the question I've used as an example has apparently been answered, but I think it still serves as a useful illustration.

  • 1
    You could start by a proof by contradiction? May show that there are very few to no proof that books, literature, and other writings use it in that form? If it's not based on a rule, then the best way to prove/disprove, in my opinion, is with examples. If you can't prove that it is true, you can authoritatively assume it is false.
    – Hank
    Commented May 11, 2017 at 18:49
  • @Hank proving a negative existential ('X is never/hardly ever used') is difficult. Using search engines on corpora has all sorts of difficulties (limitations of the corpus used, having the right search features, creating the appropriate search). Personal authoritative affirmation (a native speaker with demonstrated broad and deep experience), plus voting (to show disagreement or agreement) is not perfect but is somewhat reliable in its own way.
    – Mitch
    Commented May 12, 2017 at 18:28
  • 1
    @Mitch It can be difficult but it's not impossible. I have seen many successful proof by contradictions on this site, most of which were successful largely because of the community support it received. Is it fool-proof? Not at all. A proof by contradiction is still subject to the community, of course, I was simply suggesting something that is a start, not a final say.
    – Hank
    Commented May 12, 2017 at 18:37

1 Answer 1


Three types of tools, if carefully used, may provide evidentiary bases for authoritative answers to the question about 'glee'. Two of those types of tools have already been deployed in the question and its answer, although perhaps they have not been deployed as skilfully as they might be. The two types of tools already deployed are (1) personal experience, and (2) lexical definitions. The third type of tool is (3) the set of available corpora. There again, the tool may be deployed with greater or lesser skill and insight.

Some (of the many) available corpora are accessible through the BYU corpora interface. Using corpora to answer the question as authoritatively as possible within carefully articulated limits is likely to get as near to a "case closed" answer as is feasible. The limits are both external to the corpora, as imposed by the answerer's ambition, ability and resources, and internal to the corpora along with the interface, as imposed by the essential nature (for example, the types and dates of material in the collections) of individual corpora and any artificial restrictions imposed by the interface on the investigation.

The general direction of fruitful investigation has probably already been set by the observed differences between lexical definitions of 'glee' in US English and other Englishs. Weighing against the productivity of going in that general direction are the poorly defined question targets of "in normal use" and the "average native english language user". Those targets might, however, be profitably defined in any worthwhile answer.

See also the answer to a very similar, and recent question, What are some good and authoritative reference/data source for modern usage examples of words?. Note also that questions shown in the "Related" sidebar of this and the linked question may prove informative.

  • Can you add comments on the problems that searching and particular corpora might have? Maybe the user didn't make the right search, maybe the query language for that SW doesn't have the right features to enable a correct search, maybe the corpus doesn't include the appropriate source documents)
    – Mitch
    Commented May 12, 2017 at 18:30
  • @Mitch, I'm sure that's a good suggestion. I'm not sure I have a good handle on the comments you have in mind, though, and I am sure that when I consider adding such comments every lazy bone in my body protests mightily that it's a herculean task. I did add a link to your recent, quite similar answer, which I found when I started looking for ideas to focus any comments. Feel free to change my answer as the whim moves you (or add another).
    – JEL
    Commented May 14, 2017 at 7:52

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .