Given that both romanizations are transliterstions of a (supposed) native
script I really see no problem. You won't e.g.  generally se the same
transliteration/transcription of Cyrillic, Devanagari, Japanese or other
scripts on a general map or in a newspaper as in a philological article.
The uniform use of Pinyin to transcribe Mandarin is both recent and
exceptional. In particular you may take an analogy from the way different
transcriptions of Japanese strive to represent either Japanese phonology,
kana spelling or the pronunciation in the terms of the conventions of
(usually) English, or some compromise between these according to the whim
of the scheme designer. In practice the choice of scheme or convention is
mostly guided by the intended or expected audience of the publication -- in
particular how much their bias is linguistic or not, or how desirable or
critical lossless retranscription/retransliteration is. We all know that a
cartographer will generally fall at or near the 'in terms of English/target
audience's language conventions' end of the spectrum.

In fact there does seem to exist a fairly widespread comvention which can
be summarized as "consonants according to or based on English conventions
and vowels based on Italian/Spanish/German conventions" -- an effect of
awareness that English vowel spelling conventions are very out of the
mainstream for the Latin alphabet as well as internally ambiguous. It is
interesting to see that this 'AngloRomance' convention is widely adhered to
even by Swedish writers targetting a Swedish audience. And even where you
will see Swedish-based conventions for consonants like <tj sj zj> (mainly
Cyrillic) you won't see Swedish-like <å o> but Italian-like <o u>. And the
sibilants will be <zj tj sj sjtj> not <zj tsj sj tj> which would better
reflect the fact that Swedish <tj> is [ɕ] for most speakers, not to speak
of the fact that Swedish lacks [z ʒ] and <sj> is [x] or [χ] for most
speakers.  The same goes for Ancient Greek: you may see <Atena> rather than
<Athena> but you won't see <Athäna> and you may see <Akilles> but never

My own shifts in conventions for transcribing Sohlob. I started out over 15
years ago with an ASCII-based system using <tj sj dj zj> for alveopalatals.
This was actually sub-phonemic since I decided very early that [ʑ] was an
allophone of /dʑ/, and <j> was used only in those digraphs. At the same
time I used <ny hl hr> for single phonemes and <ng> ambiguously for /ŋ/ and
/ŋg/, justified by the conventions in the 'native' script which was and is
under-specifying to a high degree. There were also the slightly odd <e> for
/ɨ/ <ae> for /æ/ which was unambiguous because of vowel harmony. I later
switched to a Latin-1 based convention with <æ c ç j j> instead of <ae tj
sj dj zj> and this in spite of the fact that /tɕ dʑ/ were digraphs in the
native script! When Unicode entered the scene I did introduce <ŋ> instead
of <ñ> in the transcription of the (then) protolanguage Kijeb, and I did
consider introducing <ñ ŋ ł> instead of <ny ng hl> in Classical Sohlob but
decided against it because it wasn't clear what I would replace <hr hm hn
hny hng> with in the sister language Cidilib. More recently when writing
*in Swedish* about the Sohldar universe for a non-conlanger audience I've
considered following AngloRomance conventions and use <ch sh j zh kh gh>
rather than <c ç j j x q> -- being mostly concerned about the Cidilib
placename Jdrig/Zhdrig The problem is that most Swedes would even pronounce
<j> as [j] when speaking English! OTOH the 'SuedoCyrillic' <tj sj dj zj ch
gh> might look silly to the audience (AngloRomance rather than Swedish
expectations in a Swedish audience!) and moreover <sj ch> might be
misleading. I got so despondent that I considered going for a straight
transliteration of the native spellings using <ty sy zy zy x q> in Cidilib
and <tx sx dz zz x q> in CS! Should I even go so far as to use <ă> or <ȧ>
for /ʁ/ and CS <bb dd gg> for /p t k/ and Cidilib <hb hd dy hg> for /p t tɕ
k/? Then why not <hh> for /s/ as is actually the case in the native script?
 I think that would introduce bogus alienness were none actually exists and
make the two conlangs seem more different than they are or would be to
illiterate native speakers.


Den fredagen den 18:e januari 2013 skrev Herman Miller:

> 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.
> Su tiski marvi žihl jĕŕastajan vë łivi žeġ jĕlak.
> The quick brown fox jumps over the lazy dog.
> Rĕkezanuj my zjaniki tanigira da, rĕlinajžataj vë rĕsarga.
> If you seek a pleasant peninsula, look around you.
> 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.
> I do have accents on some vowels (e.g. "Sujinán"), and dieresis/umlaut
> marks over vowels for additional vowel sounds, but I'm thinking that it
> would be a lot easier if I just started using "ts" for the /ts/ sound in
> Tirelat instead of "ċ", and "dz" for /dz/ instead of "ż". Using "gh" for
> the voiced velar fricative /ɣ/ instead of "ġ" would be more convenient for
> typing. And why not use "hr" for a voiceless r? That would be ambiguous if
> I continue using -h for long vowels (is "lahra" pronounced /la:ra/ or
> /lar̥a/?), but I can write long vowels as double (so /la:ra/ would be
> spelled "laara").
> For the world map, I've also used a more English-like convention where "j"
> represents /dʒ/ and "y" is /j/. With these conventions, Tirelat spelling
> might look something like this:
> Su tiski marvi zhiil yëhrastayan vë hlivi zhegh yëlak.
> The quick brown fox jumps over the lazy dog.
> Rëkezanuy mï zyaniki tanigira da, rëlinayzhatay vë rësarga.
> If you seek a pleasant peninsula, look around you.
> (The difference between "ĕ" and "ë" in Tirelat romanization is only a
> spelling convention, following the way Tirelat is spelled in the Kjaginiċ
> alphabet.)