[Lingtyp] Structural congruence

[Matthew Dryer]

Peter,

I can’t speak for Martin, but since my views on this are very similar to 
his, let me respond from my point of view. When I say "describing or 
analyzing a particular language is a completely different enterprise 
from classifying the language typologically", I do not intend to deny 
that the two enterprises are very tightly linked. You cannot classify 
languages typologically without “taking into account how the systems of 
particular languages work”.

Consider your example “If, based on the TMA questionnaire, one says that 
"language A has a past perfective", this is a statement which makes 
sense from the point of view of the system of this language, not only 
for the purposes of comparison.” I would argue that “language A has a 
past perfective” does NOT make sense from the point of view of the 
system of the language. The language will have a particular aspectual 
category with specific semantic and morphosyntactic properties and it 
may satisfy the definition of a comparative concept “past perfective”. 
But what defines that particular aspectual category is the specific 
semantic and morphosyntactic properties it has in the system of the 
language. If you spell out those semantic and morphosyntactic properties 
in great detail, then you’ve described that category. The statement 
“language A has a past perfective” doesn’t add anything to what you have 
already said. It only says something about how that language-particular 
category resembles categories in other languages, but it tells us 
nothing about the system in language A.

The point is clearest if the language you are describing has an 
aspectual category that bears some resemblance to what people have 
called past perfectives in other languages but differs somewhat.If you 
have fully described the aspectual category in your language, then 
answering the question of whether that category is a past perfective 
cannot add anything to what you have said. The aspectual category in 
your language is defined by its properties in the system of your 
language. The question of whether that aspectual category is an instance 
of a past perfective thus has no bearing on the system of your language 
since once you have fully described that aspectual category in your 
language, you have said all there is to be said about that category as 
far as the system of your language is concerned.

Matthew

On 1/20/16 5:34 PM, Peter Arkadiev wrote:
> Martin, I agree that doing typology is of course different from language description, but I am reluctant to subscribe to the view that these are two fully disjoint enterprises with entirely different logical and epistemological bases. You give Dahl's "Tense and aspect systems" as an example, but in my view this work is notably not only about comparative concepts and typology, but also about ways to adequately describe particular languages, which has been proved by successful application of TMA questionnaire to many languages beyond the initial sample. If, based on the TMA questionnaire, one says that "language A has a past perfective", this is a statement which makes sense from the point of view of the system of this language, not only for the purposes of comparison. Moreover, I am wondering how typology of grammatical systems can be achieved without taking into account how the systems of particular languages work.
>
> Best,
>
> Peter
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/lingtyp/attachments/20160120/a45f66fb/attachment.htm>