Time Depths and Comparisons

[Rankin, Robert L]

> Has anyone found a good way to handle time in presenting
> reconstructions along with dated language data, especially
> across languages?

No. :-)

Having said that, I do have a paper in which I attempt to date subgroup
splits in Siouan using the largely-borrowed cultigen terminology (squash,
corn, beans, etc.) in which I display the results in two ways. (1) a
Stammbaum with the nodes labeled and dated (tentatively of course) which
conveys the temporal distinctions and (2) a map (which should be a *series*
of maps) to convey the spatial distinctions that need to be made.

Beyond that, I must say that I've never been happy with all the ink that has
been spilled debating whether the family tree or wave diagrams are "true" --
or, worse, whether one is true and the other isn't. Obviously both are
"true" but neither is complete. It's what happens when we try to collapse a
stack of comparative grammars and a stack of dialect atlases to a single
page. There is no substitute for a discursive treatment.

But, like John, I'm not certain this is the kind of question you're asking
about.

> The farther back we go, the harder to date a reconstruction,
> and if we attempt comparisons among related languages widely
> separated across space as well, whether in their lexicons or
> structures, we need to observe the speech communities'
> changed environments and interactions with neighbors and
> trading partners, and should date acquisitions of loans if
> possible.

Yes. And that's why most comparativists don't have much use for Labovian
dialectology. It isn't because there is any serious disagreement between the
two groups (at least using Labov's '94 book as a benchmark); it's because
it's SO hard to determine precisely those communities, interactions and
environments at any great time depth. Sociolinguistic observations are often
possible at very shallow time depths, but the returns diminish quickly at
any serious time depth. This often just leaves us stuck with
uniformatarianism, sound change regularity and concommitant relative
chronology as tools. Plus whatever shallow archaeological correlations we
can make.

But there wouldn't be any sport to it if it were easy.

>
> I'm trying several things in my works-in-progress.  First, I
> use visual alignment as pioneered by Haas in listing items to
> be compared, phoneme by phoneme, morpheme by morpheme, but
> the order of languages, and thus of data cited, is given as a
> continuum of similarities, which de-emphasizes fixed
> boundaries for the subgroups we have been calling Mississippi
> Valley (Winn.-IOM, Dakotan and Dhegihan), Missouri River, and
> Ohio Valley.  Mandan isn't isolated, and Catawba is right in
> there, not separated out.  A (genetic) continuum line can
> also be shown with curves or loops to suggest spatial
> arrangements, and thus likely closeness among speech
> communities who were neighbors at different time periods.  I
> haven't seen how to combine a tree and a continuum.

Although when divergence begins, there is normally a dialect continuum, and
that continuum may persist for awhile (or in some cases "forever"), my own
experience with Siouan tends to cause me to de-emphasize the continuum. One
can't always do that with language families like Muskogean, northern
Athabaskan or others for a variety of reasons, but a unique combination of
circumstances (e.g. migration into areas of reduced annual rainfall) caused
numerous of the Siouan-speaking tribes to move along major rivers into the
eastern plains virtually forming a Stammbaum on the face of the land. The
analogy isn't perfect, but the subgroups seem very well defined to me
(compared with, say, Germanic or Romance).

> Does the Comparative Siouian Dictionary (archived at U. of
> Colorado I think) handle comparative time depths?

It's pretty primitive. It's presented subgroup by subgroup, those groupings
having been determined in advance and pretty much agreed upon. It hasn't
been generally accessible, but it needs to become so.

Again, I'm not sure this addresses your points.

Bob