Time Depths and Comparisons

[Koontz John E]

On Fri, 30 Mar 2001, ird wrote:

> Has anyone found a good way to handle time in presenting
> reconstructions along with dated language data, especially across
> languages?  Dates of transcriptions, accurately interpreted and
> updated, must reflect language states over a period before and after
> recording, say two or three generations, maybe 75-100 years.
> Sometimes the beginnings of continuing changes have been documented,
> like */s/ > /h/ observed in IOM by Whitman.  I am also aware that
> change accelerates during periods of rapid change (e.g., the striking
> differences betweeen the English in my faqmily Civil War letters and
> contemporary English in the same locale).

I'm not quite sure I understand the question.  An example of problems
you're facing might help.

Typically, e.g., in the CSD, the database representation is something
like (highly idealized, with partially invented data):

PMV *s^uNk(e) | dog

DA s^uN'ka | s'oN'ka | dog, horse | Foo 1970:198a

OP s^aN'ge | shoN'ge | horse | Bar 1884:101.2

which could be represented alternatively as

PMV-Reconstruction s^uNk(e)
PMV-Gloss dog

DA-Phonemic s^uN'ka
DA-SourceForm s'oN'ka
DA-Gloss dog, horse
DA-Source Foo 1970:198a

OP-Phonemic s^aN'ge
OP-SourceForm shoN'ge
OP-Gloss horse
OP-Source Bar 1884:101.2

This would be rendered in a published dictionary something like

PMV *s^uNk(e) 'dog' DA s^uN'ka "s'oN'ka" 'dog, horse' (Foo 1970:198a); OP
s^aN'ge "shoN'ge" 'horse' (Bar 1884:101.2).

Of course, heavier use of alternate type faces would be made, etc.

In explaining matters putting a few well chosen examples into a table
makes it easier to make a point, but publication of volumes of material
requires the paragraphed format.  In collecting the data you need
something that is convenient for your database system, and for data entry.

JEK