Hear, hear, Martin. On Dec 17, 2012, at 8:24 AM, Martin Haspelmath wrote: > > … > But I think one should be careful to avoid the impression that the grammatical terms we use are anything other than convenient tools used by linguists. Categories of languages are language-specific entities, and they cannot be "polysemous" or "homonymous". > > Greetings, > Martin >