Re: Naturalistic conlang spelling reform!

Shreyas Sampat wrote:
> Herman Miller wrote:
>
>> And so on. Of course, this means keeping track separately of the
>> pronunciation and the spelling, but that part should be easy enough to
>> figure out. One thing I haven't been successful with is figuring out
>> how to keep track of the history of a language.
>
>
> I keep Sein.'s vocabulary in an Excel database with its reflexes in
> earlier and different stages of the language.
>
Hmm, OpenOffice can deal with Unicode, so this is a possibility. Back
when the only spreadsheet I had was Microsoft Works, it wouldn't have
been usable. The advantages over a word processor are the larger number
of columns and the ability for a cell to store a word longer than a
column. So it could have its uses.

But in the initial stages of starting with a modern language and filling
the details of the earlier history of the language, I'm not sure this
would be very helpful. I don't even know the sound changes that occurred
in previous languages, the order of the sound changes, which words may
have been borrowed from other languages, or how the meanings of the
words may have changed. In most cases I don't have any records of
earlier or related languages, so pretty much anything goes at first. But
as soon as there are two or more related languages or historical stages
of a language, maintaining consistency becomes an issue.

It's usually not that hard to get started. Take Yasaro for an example: I
know that the distinct rising and falling tones arose from shifting
word-final stress to the preceding syllable. So I can reconstruct an
earlier form of "vélir" (line, row) as "velír", while "tèling" (narrow)
would have been "téling". The Yasaro sound "o" came from a nasalized "a"
at the end of a word, pronounced [O~], which lost its nasalization. So
"thùno" (to sleep) was once "thúnã", and before that may have been
something like "thúnam".

But for every sound that I have some idea of its history for, there are
many others that I don't know where they came from. Take the "ch" [tS]
in "chânga" (to eat) for instance; sounds like that typically would have
come from something else, but what? /kj/ ? /tr/ ? Lots of possibilities
there. Yasaro has a dental/alveolar distinction in nasals -- where did
that come from? Then once I figure out where every sound in the language
came from, I need to keep track of all the rules.

I think it must be easier to start with an older language and work
forwards. The few times I've tried to create related families of
languages, one of them is always a modern language, then I have to
reconstruct an older language and push it forward in a different direction.