Zompist Bboard Again

Working on a romanization for Old Japanese. I'd like to only use single characters for each phoneme (as understood currently), and preferably only letters that would be easy for English speakers to guess the pronunciation, but I'm having trouble mostly with the vowels.

/*p *t *s *k/ <p t s k>
/*ᵐb *ⁿd *ⁿz *ᵑɡ/ <b d z g>
/*m *n/ <m n>
/*r/ <r>
/*w *j/ <w j>
/i1 i2 e1 e2 o1 o2/ <i y e ? o ?>

I know that making /j/ <j> isn't the best option for ease of pronunciation by English Speakers, but if I used <y> for that I'd have to find something for i2, which often came from a combo of /i and /u/ or vice versa in earlier stages, and so I figured would be close to /y/ and some English speakers would be familiar with languages with <y> for /y/ or similar. I might use the schwa symbol for o2, although it was actually more common than o1 which might have been /o/ or /wo/. I'm not sure what to do with the two e's. Vovin writes them as <e ɛ> while Frellesvig and others have <ye e>. I might go with <e ɛ> since e1 is the more common of the two and it still gets the idea that they're both e like across.

So I'm tentatively at <i y e ɛ o ə> for /i y (j)e ɛ (w)o ə/ but I'm not 100% satisfied with it? Anyone have any thoughts?

linguistcat wrote: ↑Mon May 08, 2023 3:53 pm So I'm tentatively at <i y e ɛ o ə> for /i y (j)e ɛ (w)o ə/ but I'm not 100% satisfied with it? Anyone have any thoughts?

In many materials I've encountered, */i₂/ is "wi", */o₁/ is "wo", and /e₁/ is "ye", so I would go with:

/*p *t *s *k/ <p t s k>
/*ᵐb *ⁿd *ⁿz *ᵑɡ/ <b d z g>
/*m *n/ <m n>
/*r/ <r>
/*w *j/ <w y>
/i₁ i₂ e₁ e₂ o₁ o₂/ <i wi ye e wo o>

This has the advantage of keeping things close to the Hepburn we're all used to using for Romanising Japanese, and avoids using IPA characters in narrative text (which I admittedly tend to find both jarring and horribly ugly) while still giving a good idea of how a word is meant to be pronounced.

If you want people to pronounce this correctly *mb, *nd, *nz and *ng are probably essential.

foxcatdog wrote: ↑Mon May 08, 2023 5:49 pm If you want people to pronounce this correctly *mb, *nd, *nz and *ng are probably essential.

Yeah fair point. I guess I can be looser with the vowels than with the consonants.

This is for a novel, right? So you're approaching it backwards. This isn't a linguistic text where you can look up the exact phonetic information. This is something that would exist entirely within English phonotactics, as understood by the reader. You are working inside English, trying to create a certain aesthetic feel. Otherwise, what comes out will just be gibberish. Give someone a word like {jy} and they're not going to get /jy/ from it; they'll get /dZai/.

So what do you want Old Japanese to sound like? If you want speakers to get that this is a Japonic language with schwas and pre-nasalization, you have to write that: 母 becomes {muh}, 具 becomes {ngoo}. You can write these in a way that makes more sense in a linguistic textbook, but then you have to accept that you are telling your readers to mispronounce these words, and loose the vibe you are going for.

You have to always ask "what is this orthography for? Who is reading it and how do I want them to read it? What is the effect I intend?" It's not always appropriate to start with IPA and work back from there.

Yeah... I don't know what I was thinking of with <j> for /j/ since most people reading it won't get that. Not sure what to do with i2 then, but then the three main issues would be the vowels, with <y> as /j/ and <mb nd nz ng> for /ᵐb ⁿd ⁿz ᵑɡ/.

So I'm currently at:

/*p *t *s *k/ <p t s k>
/*ᵐb *ⁿd *ⁿz *ᵑɡ/ <mb nd nz ng>
/*m *n/ <m n>
/*r/ <r>
/*w *j/ <w y>
/i1 i2 e1 e2 o1 o2/ <i ? e ? o ?>

linguistcat wrote: ↑Mon May 08, 2023 9:23 pm Yeah... I don't know what I was thinking of with <j> for /j/ since most people reading it won't get that. Not sure what to do with i2 then, but then the three main issues would be the vowels, with <y> as /j/ and <mb nd nz ng> for /ᵐb ⁿd ⁿz ᵑɡ/.

So I'm currently at:

/*p *t *s *k/ <p t s k>
/*ᵐb *ⁿd *ⁿz *ᵑɡ/ <mb nd nz ng>
/*m *n/ <m n>
/*r/ <r>
/*w *j/ <w y>

I would go with this and the suggested /i₁ i₂ e₁ e₂ o₁ o₂/ <i wi ye e wo o>.

I know my ignorance of Old Japanese, but looking on Wikipedia, maybe /i1 i2 e1 e2 o1 o2/ ⟨ee ea e (u)i uo o⟩? Admittedly it makes the phonetician in me want to scream, but has the advantage that the usual English readings correspond approximately to what’s been reconstructed (at least given what I can see on Wikipedia). Taking some random example words, this would give e.g. /waki2ratuko1/ wakearatukuo, /kazo1pe2/ kazuopi, /tirike1mu/ teereekemu, /nage2ki1/ naguikee, /yamato2 ko2to2ba/ Yamato kotoba.

EDIT: Or, of course, you could use the analysis of Pre-Japanese as having /i u ə a/, which fairly straightforwardly gives you a transcription ⟨i u e a⟩, or if you prefer ⟨ee u i a⟩ with the same approach as above.

I can't really add anything more insightful than what has already been said, but you haven't said how important the inclusion of Old Japanese is to the novel. Is it just for names? Will there be dialogue in it? Will it be used for very Japanese things for which we have no satisfactory English word? Depending on how important it is, you could include an appendix which explains how Old Japanese should be pronounced and explain the difficulties an English speaker would encounter (pre-nasalisation, phonemic schwa, etc). This gives your readers the opportunity to learn, if they are so inclined, how these Old Japanese words should really be pronounced vs. how their anglobrain has been doing it throughout the novel.

Do something which gives a fairly good idea to the non-expert reader what those words sound like, even if it misses some details. Like how we usually spell the names of those two great Hindu deities Vishnu and Shiva, even though they contain different 'sh'-sounds.