Concerning WALS - Bees, Bats, Butterflies

[Salinas17 at aol.com]

In a message dated 11/7/08 12:50:44 PM, dryer at buffalo.edu writes:
> <<For another, even if the explanandum is something that we can
> be fairly confident of, hypotheses about why languages are the way they are
> ultimately just that, hypotheses.  All too often, they are untestable and
> unfalsifiable and always will be.>> 
> 
A Quick Comment.   Early natural histories organized groups of animals 
according to categories that we find odd these days.   Birds and bats and 
butterflies were all things that fly, so they were grouped together.   Now this was not 
so much a theory, or even a hypothesis, as it was a way of giving some kind of 
organization to that part of the natural world.   A random organization makes 
a presentation of data difficult for any audience.

Modern biology textbooks often poo-poo the categories of the old natural 
historians because modern biology views them from the point of view of modern 
theories of temporal evolution and relateness.   That birds, bats and butterflies 
all fly is a coincidence from that perspective.

But, of course, in terms of physics, birds and bats and butterflies do indeed 
have something in common -- in the sense of their main method of motion.   
These three not-closely-related animal forms do share a common solution to the 
problem of motion, even if each particular solution arose independently.   What 
appears to be an erroneous grouping actually does inform us about how 
external contingencies can lead to a common solution, DESPITE a lack of direct 
relatedness.   An important aspect of evolution.   And possibly of language.

So, is WALS the equivalent of the old naive natural histories, slapping 
together linguistic forms on some superficial sense of typology?   I don't think 
so.   The real question is how the groupings end up working with everything 
else.   How does one explain all these "typologies?"

The worse possible reaction is to belittle the process as a mere 
"hypothesis."   Hypotheses are the meat of any REAL science.   Any theory that does not 
yield testable hypotheses, by long and hallowed tradition, sucks.  It seems to 
me that the organization of WALS constitutes some sort of hypothesis.   It is 
just a matter of someone somewhere actually stating it, and making some 
testable predictions based on it.   If A, B, C, are here, then where would we expect 
to find D?

Stating that all the observable language structures are in reality one 
structure does not help, because it doesn't explain why there is more than one human 
language.   We need a sensible description of the process, the mechanism -- 
given the data. Which means we need a Copernicus, not a Chomsky or a Greenberg. 
  We need a way to say all this makes sense, if we just put the sun in the 
center.   (At this point, I'm afraid a Darwin or Einstein are down the road a 
bit, for linguistics.)

In the mean time, WALS is a collection of data organized from some general 
point of view.   If there is a Copernican solution to the data, then please have 
at it.   In the meantime, we need to be satisfied with bats and bees and 
butterfies living together, all under the same chapter heading.

Regards,
Steve Long


















**************
AOL Search: Your one stop for directions, recipes and all other 
Holiday needs. Search Now. 
(http://pr.atwola.com/promoclk/100000075x1212792382x1200798498/aol?redir=http://searchblog.aol.com/2008/11/04/happy-holidays-from
-aol-search/?ncid=emlcntussear00000001)