Naturalistic conlang spelling reform!

All this talk about oddities in spelling, dialect variations, and so on,
has made me think about the spelling in my own languages. In the old
days I used to create alphabets for all my languages, and when I used
romanization, it was essentially a transcription of the native spelling.
Then around 10 years ago I switched to a more phonemic approach. I
regularized the romanized spelling of most languages, ignoring the
native spelling if I had one.

But there's something attractive about a convoluted spelling system that
reflects a history of change and borrowing. It might be nice to
reconstruct an earlier version of a language that may have been spelled
more regularly, then retain some of the old spellings even after the
language changes. For instance, the spelling of tones in Simik could be
based on the reconstructed old forms:

jûma "dream" < jùma
mêishà "glass" < mèisa
tílma "logic" < tílma
vé:li "good" < vé:li
valâr "finger" < valàzh
zu:rèi "to die" < dhu:zhèi
saré "right" < thazhé
nù:mi "idea" < nu:mí

Jarda spelling could be based on Kianarthal:

seol "fire" < selu
shaf "leaf" < thiasp
lhash "electricity" < liathi
khlar "ground" < skelaj
tis "ice" < theise

Yasaro pitch accent could be marked on the original stressed syllable:

lále "gray" < lalé
kàju "turquoise" < káju

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.

Replies