Open wcroft opened 6 months ago
I have no problem with adding an optional field of "dispreferred aliases", or even as a flag that could be set for each alias. What would make this interesting is making the search functionality respect that. That is, dispreferred aliases should appear in the search, but maybe as a less relevant match.
(transferring this issue from my email to GitHub): should we allow dispreferred aliases to be ambiguous between CCs?
I would rather that people mostly stop using dispreferred aliases: they are sometimes misleading, and often ambiguous in a bad way (that is, they are used to denote two different concepts which are useful CCs but not the same and hence potentially confusing). The reason I included them in the book is because many linguists including descriptive linguists use them fairly frequently. So when you’re reading a reference grammar, you should know what that term is used to “really” mean, including possible ambiguities.
But if that is the purpose of dispreferred aliases, then there is also no reason why a dispreferred alias can be an alias of multiple CCs, because those ambiguities are real in the “data” (= the text of reference grammars). Still, I feel uncomfortable about sanctioning terminological ambiguity in the database.