pseudo-generalizations

[Tom Payne]

Yes, Matt, you are right. But, still I think one should always be
striving to motivate statistical universals, even if it's hard. Such
motivations will always be "stories," i.e., hypotheses with more or less
evidence in their favor, but I don't think this is really qualitatively
different from what other sciences do. For example, I have asked
advanced physicists whether the atom is a proven fact, or a theoretical
construct. They always say it is a theoretical construct that explains a
heck of a lot of data. (Don't ask an undergraduate physics major this
because the lower level courses tend to make it sound like a proven
fact). Someone just "told a story" something like: "What if matter were
made up of little particles called atoms, that have such and such
properties. What would the world be like if that were true?" Then
experiments were designed to test whether the world is the way it should
be, if the story were true. Many things about the world make it seem
like atoms are real, and a few things are still problematic, but the
science goes on trying to refine the story.

So when we say "What if speakers intend their messages to be
interpretable? And what if iconicity would help messages be more
interpretable in such and such a way? How would that affect the
structure of human communicative systems?" I think we are doing pretty
much what the physicists are doing -- telling a story that has more or
less value depending on how much observable data are explained by it.
Statistical universals are (part of) the observable data.

Tom 

-----Original Message-----
From: Discussion List for ALT [mailto:LINGTYP at LISTSERV.LINGUISTLIST.ORG]
On Behalf Of Matthew Dryer
Sent: Tuesday, September 30, 2003 11:07 AM
To: LINGTYP at LISTSERV.LINGUISTLIST.ORG
Subject: Re: pseudo-generalizations


I have a bit of a problem with Tom's claim "A statistical tendency is
just
an intriguing pattern until you come up with a principled explanation
for
it."

There is a danger in viewing crosslinguistic generalizations as lacking
value if they are not explained. Whatever the problems of reliably
testing
crosslinguistic generalizations, these problems are not nearly as severe
as
the problems of testing hypotheses for why these patterns exist.  In
fact,
the problems of testing explanatory hypotheses are so severe that they
are
hardly worth discussing.  Because of this, we can always be much more
confident of the crosslinguistic generalizations than we can be
confident
of our explanations for them.  For that reason, I see the
crosslinguistic
generalizations as more valuable than their explanations.

This is not to say that a crosslinguistic generalization with an
explanation is not more valuable than a crosslinguistic generalization
without an explanation.  It just means, on my view, that having an
untested
and probably untestable hypothesis for why the generalization exists
doesn't add a whole lot to the value of the generalization itself.

This is a reason why we should be wary of placing too much value on the
existence of an explanation for believing crosslinguistic
generalizations.
If there is a problem with being confident that an crosslinguistic
pattern
is real or not, it would be a mistake to think that having an untested
hypothesized explanation for the pattern adds much reason to believe
that
the pattern is real.

Tom's comments seem to ignore completely the difference between coming
up
with a hypothesis for why a crosslinguistic pattern exists and finding
relatively convincing reasons for believing that that explanation is
right.
I think relatively few - if any - explanations of crosslinguistic
patterns
have achieved the latter status.

Matthew Dryer

--On Saturday, September 27, 2003 10:01 PM -0700 Tom Payne
<tpayne at OREGON.UOREGON.EDU> wrote:

> Right! A statistical tendency is just an intriguing pattern until you
> come up with a principled explanation for it. And if the principled
> explanation explains more than one statistical tendency, all the
better.
> For example, "iconicity" is one general principle that accounts for
> several structural statistical tendencies in the world's languages. If
> you posit a line of reasoning something like 1) speakers intend their
> messages to be interpretable, 2) iconicity makes messages more
> interpretable, and 3) language structure evolves under these
pressures,
> then you can explain (or perhaps "motivate" is a better term) many
> statistical universals, such as: the tendency for phonologically
larger
> structures to express semantically more complex concepts (I'm thinking
> of causatives, reflexives, etc.), certain kinds of split ergativity,
and
> so forth.
>
> This is why I like to say I do Typological/_Functional_ linguistics.
> Typology is the descriptive part -- "What are the intriguing
structural
> patterns out there?" But the job isn't finished until the patterns are
> explained according to some coherent and general principle, anchored
in
> cognition, neurology, communication or some other domain independent
of
> language structure itself.
>
> Tom Payne