So here’s something I found interesting while procrastinating:
A blogpost by Martin Haspelmath about a talk of his (“Towards an IPA of morphosyntax”).
One of the main topics of the talk was how to identify the ‘natural kinds’ of morphosyntax, that is whether we could find a nice overview of the fundamental features and categories that are relevant for syntax. He argues that this turns out to be highly problematic. Definitely go read the post if that interests you!
What I personally found most interesting was his comparison of different approaches to the problem where he contrasts ‘an essentialist approach to a measurement approach: Mendeleev vs. Passy’:
“But in linguistics, as in other social sciences such as anthropology, a measurement approach seems to be more promising (as first argued forcefully by Bickel 2007). So instead of
– looking for the “correct analysis” of a language-particular pattern
– debating the “status” of a phenomenon
– proposing “underlying” entities that allow a more elegant description
general linguists should “measure” the differences between languages and use these measurements to find general pattern.
Units of measurements are not claimed to correspond to anything in nature – they merely serve as yardsticks to allow scientists to compare related phenomena and to state possible generalizations. Science should “cut nature at its joints”, but when nature appears to be continuous, a measurement approach may be fruitful.”
That is, we either have a more physics-y approach to language, looking for deeper and ever more abstract concepts that underlie a phenomenon (i.e. Standard Model of particle physics or periodic table approach),
or we take serious the possibility that language simply does not behave that way, that is the unreasonable success of physics doesn’t repeat itself here; so instead of a priori imposing on the subject matter how in the end it will be solved (what Haspelmath therefore calls aprioristic accounts) we just measure what’s there and try to find ever more precise ways to compare the data we have.
Haspelmath also concedes that there might not even be any natural kinds to find.
I can kind of see both points. The first possibility is undeniably more fascinating if it turns out to be true and it might be more able to push research into a more penetrating direction, even if not everything turns out as well as one has hoped it would;
the other approach on the other hand is more honest and comes with way less philosophical and ‘presuppositional’ baggage, and the danger is smaller that in the end one always only finds what they were looking for since the own preconceptions are not questioned.
Somehow this reminds me a lot of these Bobaljik triplets and alleged impossibility of an ABA pattern: It is an amazing find that in apparently no language you find something like good – better – goodest, and this seems to carry over to other triplets. On the other hand there is the ensuing industry of explaining away every counterexample there is, e.g. by either claiming that the bad guy actually doesn’t belong in this triplet, or by changing the order to yield an AAB pattern.
That is, one might want a deeper explanation for why this is the case, maybe even one in terms of UG but it might just be a strong statistical tendency of ‘disorder’ (i.e. suppletion) to only increase. So the most relevant question is what a counterexample would have to look like that leads these *ABA people to stop trying to explain it away.
But before I get even more sidetracked, I’ll end the post.