--- On Thu, 1/17/13, Herman Miller <[log in to unmask]> wrote:

> In my normal Romanization of Tirelat
> (and other recent languages), I tend to use diacritics so
> that each phoneme of the language is represented by a single
> letter. E.g. [...]
> But recently I've been working on a map of the world, with
> lots of tentative names for places where I don't even know
> what languages are spoken. I decided to use a consistent
> spelling for all place names, rather than trying to figure
> out the phonology for all the languages ahead of time. E.g.,
> there's a name "Lanyamets", but I don't have a clue whether
> "ny" represents two phonemes /nj/ or a single phoneme /ɲ/
> in whatever language is spoken there, or whether "ts" is
> considered as one or two phonemes.

You don't necessarily need to know, either. Just keep in the back of your
mind who the cartographer is: if it's the semi-omniscient you, then you
will at least have a clue that the NY is something to consider when it
comes to making the language that goes with that country. If the map was
drawn from the perspective of a Tirelatian cartographer, then I'd suggest
that the names will be in good Tirelat orthography, but perhaps mangled
with respect to their native names, e.g., Bangkok v. Krung Thep.

I tend to make maps from the semi-omniscient perspective, and thus place
names will be more or less native in form. When I make maps from the
perspective of one country or another, the names might be different, being
in the native language of the cartographer, not the people depicted in the