Those who know me better will be aware that I keep insisting on careful use of terminology in linguistics, especially in grammar (my main area of research), but also in other areas – for example, I often point out that it’s very problematic to use the term borrowing only for cases of copying (of words and other features from a donor language) that were NOT due to imperfect learning of a second language (i.e. substrate effects). The reason is that we need a general term for all kinds of copying, because in many or most cases we don’t know the circumstances under which the copying took place. But copying as a general term, with borrowing as a special kind of copying, simply won’t work: The terms borrowing and loan have been deeply entrenched in linguistics for over a century, and changing terminology so that a more general term comes to mean something more special simply won’t work for the whole discipline. Those who want to use borrowing in this narrow sense will perpertually have to say “borrowing in the sense of Thomason & Kaufman (1988)”.
I don’t have a good recipe what to do to avoid the constant creation of new confusion due to new meanings of old terms, and maybe not everyone feels it in the same way as I do. Many linguists inhabit only small corners of the discipline, and they have no trouble communicating with the other specialists of their (sub-)subfield. But those of us who are trying to get a bigger picture notice the problems quickly (and of course it gets worse when trying to learn from adjacent fields such as psychology or cultural evolution studies).
If you’ve read till here, then you probably have some interest in problems of terminology, so here are some more relevant considerations:
– One way in which one could improve the situation is by setting up nomenclature committees in linguistics associations, as they exist in fields such as astronomy and biology. Or at least one could organize sessions at big linguistics conferences (such as the SLE or the LSA) to discuss questions of terminology.
– Linguists often react negatively to the creation of new terms, even though they create new concepts all the time. In word-formation studies, this negative reaction to new words has a name: neophobia. But like all pathological fears, this can be cured, and it is clear that a proliferaton of technical terms is much better than a proliferation of meanings of an existing term. Synonyms can be easily mapped onto each other (in some contexts even automatically), while disambiguating polysemous terms requires a lot of knowledge.
– In the age of powerful search engines, terms that consist of unique strings have an edge over terms that just new metaphors. It is thus better to coin a term like quark than a term like string, because the latter has many different meanings in different domains. This extends to the use of hyphens with compounds – it is for this reason that I put hyphens in the terms thing-root, action-root, and property-root (as comparative concepts for nouns, verbs and adjectives, Haspelmath 2012).
– It is not true that terminological confusion is unavoidable, because a century ago the situation was much worse: Almost only terms that people knew were the terms of classical grammar, and it was very difficult to communicate across countries (let alone continents), and to look up earlier usage. Many new terms have been coined in the meantime, and with Wikpedia and internet search engines, we have excellent ways of looking up information.
– One principle of good terminology is that one should not give a new narrow meaning to an old term with very vague meanings. Some terms are so polysemous that it’s better to dispense with them entirely (e.g. markedness, Haspelmath 2006).
– Another principle is that a precise definition of a previously widely used (but ill-defined) term should not diverge too much in its extension from previous usage, because of retro-compatibility. The term aorist may not have a precise and consistent definition in the earlier literature, but if giving it a new definition involving some kind of irrealis modality or illocution would only cause confusion, because previous usages are in the domain of aspect.
– It is important to note that grammatical terminology for comparative concepts is essentially different from terminology for particular languages (Haspelmath 2010). Language-particular terms are best capitalized (e.g. the German Subjunctive mood), but here standardization is much easier anyway. The truly challenging aspect of linguistic terminology is the development of widely understood comparative concepts for grammar.