I started a conlang inspired by Marshallese and Navajo, and have taken several labialized, palatalized, and velarized consonants from them. However, I am unsure how to romanize these consonants.
4 Answers
You have three main options here:
- Use new letters for each one. Latin does this, with C
/k/
vs Q/kʷ/
. This will get infeasible if you have a lot of them. - Have diacritics for palatalization, labialization, and velarization. This is common in Eastern European languages that use the Latin alphabet, contrasting T D N vs Ť Ď Ň and so on. The downside to this method is that they can get difficult to stack, if you can have e.g. a palatalized labialized stop.
- Have separate letters that mean "palatalized", "labialized", and "velarized". Most modern forms of Cyrillic do this, with the letters Ь and Ъ indicating whether the preceding consonant is palatalized or not. IPA also does this: tʲ, tʷ, tˠ.
Which one you choose mostly comes down to your sense of aesthetics, and how you expect the language to be written/typed.
-
listing Cyrillic solely in the third category seems odd to me, seeing as East Slavic (which is what most people not specifically knowledgable about e.g. the languages of the Northern Caucaus, Central Asia, or historical languages tend to think of as the prototypical form of Cyrillic) is arguably more like the first option, but with the different letters being those for the following vowels (with the yer's only being used word-finally or before a palatalising vowel for /ʲj/ or /ˠj/ respectively)– TristanMar 23 at 16:32
-
1. I am not sure if infeasible is the right word because I have eight separate distinguishing pairs, but the other two seem more elegant. 2. Seems like something I can do, especially since I am handwriting everything in a notebook rather than typing it onto a spreadsheet. 3. Again seems attainable. In fact, I have in this short time been distinguishing palatalization with a j digraph because I do not have j in my romanization. Unsure what to do with labialization though. I am torn between options 3 and 2. Mar 24 at 4:04
In addition to the methods given by others, you could also try the Irish approach (which has similarities to the Cyrillic systems): divide your usual vowel letters into sets for each coarticulation, then only allow vowel letters that match the consonant's coarticulation to be written adjacent to it.
This works perfectly nicely when you have palatalising vowels before and after palatal consonants, but when you don't you end up needing to insert additional vowel letters that are not present as distinct vowel phonemes in order to ensure that no vowel is next to a consonant it doesn't match coarticulation with.
This is why the Irish form of the name Mary, Máire has an <i>, despite being pronounced with an /ɑ:/ or /a:/ in the first syllable (depending on dialect). The r is soft (palatalised) and so cannot appear next to the <á>, and so an <i> is inserted.
This process is also one major factor that can make Irish vowels difficult to predict from the spelling (even within a given dialect), as it's not always clear which of the vowel letters are phonemically present, and which are just there to mark coarticulation of the consonants.
Depending on your goals that could be either an advantage or a disadvantage.
Here's another option:
If basically every consonant has a palatalized (resp. labialized, velarized) form, consider transforming the following vowel letter instead of the consonant letter itself. This is what Cyrillic does, in addition to the "hard" and "soft" signs that Draconis mentioned. It has ten vowel letters, five "hard" (un-palatalizing) and five "soft" (palatalizing). It's a bit more complicated than that, since the Slavic languages have evolved considerably since the Cyrillic alphabet was invented to accommodate them. Case in point: the letter I only appear in digraphs now, to the point that those digraphs are in the alphabet, but I itself is not.
If you have a separate -ized form for each one, digraphs are probably your best bet. Depending on what other sounds you have and how you romanize them, you could use j, w, and x (i.e. tʲ tʷ tˠ are tj tw tx). But of course, it's up to you.
Another way you could do it (this is Mark Rosenfelder-style) would be to use the IPA characters themselves (i.e. tʲ tʷ tˠ are tʲ tʷ tˠ). This is another aesthetic, and works really well if that's the aesthetic you're going for. I find it gives exotic vibes. But it comes down to personal preference overall.