Sathir Phonology

Sathir phonology ain't no picnic with Jasper Johns, I'm sorry to say. It's more like a hurried luncheon with the father you never met. It does the job, though, I will say that. I'll give you the table of consonant sounds below, then the vowel sounds, and then explain how they work. Or...maybe I'll list the consonants, explain how they work, then list the vowels, and then explain how they work. I haven't decided yet. I'll have to consult my statistician. (I'm sooooooo glad that the government decided to assign every American a personal statistician after the whole Iran hostage thing.) In either case, the consonants will come first. Here they be:


  Bilabial Dental Alveolar Palatal Velar Lab.-Vel. Glottal
Normal Stops p, b t, d     k, g    
Aspirated Stops        
Lab. Stops        
Pal. Stops pʲ, bʲ     tʃ, dʒ kʲ, gʲ    
Fricatives ɸ, β θ, ðʷ s, z ʃ, ʒ x ʍ h
Nasals m, mʷ, mʲ n n ɲ ŋ, ŋʷ    
Approximants   l ɾ, ɾʷ, ɾʲ j   w  

* I've elected to proceed directly to the vowels and leave the explanations for later, for the sole reason that explaining the allophonic variations illustrated above would require far more footnote room than I'm willing to allow.


  Front Central Back
High i, iː   u, uː
Mid ɛ, ɛː ə ɔ, ɔː
Low   a, aː  

Consonantal Allophonic Variation

Now the tough part begins. If you look at the orthography of Sathir, it all makes sense. Divorcing from that, though, is tough on the children.

Below, I'll try to give as straightforward an account as I can about the allophonic variation of the consonants. After that, I'll give an account of vowel alternations. Then I'll detail the romanization system I'll be using. So, without further ado, here are the rules, so to speak:

  1. Voiceless obstruents voice intervocalically. This means that /p/, /t/, /k/ and /s/ become [b], [d], [g] and [z] between vowels (respectively).

  2. Historically, sequences of CuV and CiV became CwV and CjV. These sequences then turned into labialized and palatalized consonants, respectively. These correspond to the series of labialized and palatalized stops in the table above. It didn't only happen to stops: It happened to all consonants. The only orthographically marked phonemes, though, were the labialized stops. These were deemed significant enough to become orthographic phonemes. It's important to note that intervocalic voicing (rule 1) happened before this happened. This means that there were voiced labialized stops intervocalically at one point in time. Notice also that palatalization produced the phonemes /ʃ/ and /tʃ/ (and, since they occurred before vowels when intervocalic voicing happened, they become [ʒ] and [dʒ] intervocalically, respectively).

  3. A change that co-occurred with the labialization and palatalization listed in rule 2 was that sequences of Ch (where C was a stop) became aspirated stops, responsible for the row of aspirated stops in the table above. Notice that there is no aspirated palatal stop. This is because, historically, palatal stops came from Cj sequences, and aspirated stops came fromo Ch sequences. Neither a cluster of *Cjh nor *Chj was ever allowed in the language (and the same is true if you replace /j/ with /w/).

  4. A result of palatalization not yet discussed is what I call "normal palatalization". In Sathir, the consonants /t/ and /s/ became [tʃ] and [ʃ] before the vowel [i] (even when it was nuclear). A similar change happend to /h/. It became [ʃ] before both [j] and [i]. Some other processes of note relate only to CjV sequences, not to Ci sequences. Both /nj/ and /ŋj/ sequences became [ɲ]. All /lj/ sequences ultimately became [jj] sequences, which, in turn, became [j], via rule 6. Turning our attention to labialization, sequences of /hw/ became [ʍ], and sequences of /lw/ became [ww], which, in turn, became [w], via rule 6. Finally, sequences of /nw/ became [ŋʷ].

  5. Certain high vowels were lost word-finally in particular situations. Specifically, word-final /u/ was lost after /r/ and /l/, and word-final /i/ was lost after /s/ (which, via palatalization, became [ʃ]). When suffixes are attached to words ending in /r/, /l/ or /ʃ/, these vowels reappear (and they always appear, for neither /r/, /l/ nor /ʃ/ were allowable codas before this vowel loss occurred).

  6. All geminates degeminated. The allowable geminates were nasals, stops and the phoneme /s/ (though, as I said above, the derived geminates [jj] and [ww] also simplified). The result of this was an intervocalic contrast between the voiced and voiceless obstruent pairs [p] and [b]; [t] and [d]; [tʃ] and [dʒ]; [k] and [g]; [s] and [z]; and [ʃ] and [ʒ]. [Note: The palatal geminates arose from the following sequences: /...ssi(V).../ and /...tti(V).../.]

  7. Oh, almost forgot. The one rule that always seems to occur in a language with nasal codas is nasal assimilation (what's more interesting is examining languages where this doesn't occur. It's been said that the /n/ in the English prefix /un-/ doesn't assimilate. So the nasal in a word like "unpopular" will always surface as an [n], and not an [m], like it does in "improper"). In Sathir, there's only one allowable nasal coda, and that coda is /n/. This /n/, however, assimilates in place to whatever is following it, causing it to surface as [m], [n] (dental), [n] (alveolar), [ɲ], and [ŋ].

  8. Finally, spirantization occurred intervocalically to two series of stops. It occurred to aspirated stops, causing /pʰ/, /tʰ/ and /kʰ/ to become [ɸ], [θ] and [x], respectively. Note that this occurred after the intervocalic voicing rule, so these fricatives remain voiceless (and the stops remain voiceless because they didn't occur intervocalically when rule 1 occurred). Voiced labialized stops also spirantized intervocalically (not voiceless labialized stops, which were produced via the same rule that produced a voicing contrast between [tʃ] and [dʒ] intervocalically). The spirantization happened a little irregularly. The voiced /b/ and the labial coarticulation kind of merged to become [β]. A similar type of merger occurred at the velum, turning /gʷ/ into [w]. The intervocalic stop /dʷ/, though, was a little funny. The /d/ became [ð], rather predictably, but it didn't merge with the labialization, so the resulting phoneme is a labialized voiced interdental fricative: [ðʷ]. Given that this is a weak fricative, it's probably on it's way to becoming [w], as with /gʷ/. That change hasn't occurred yet, though, so Sathir is left with the irregular phone [ðʷ].

I think that's it... I may be forgetting something, though. If it turns out that I am, I'll write it in later.

Vocalic Allophonic Variation

These aren't nearly as exciting as the consonants. Nevertheless, there's something to them. A little something, but something, nonetheless.

  1. Something that could've been put in the section on consonants is what happened to coda /h/. Coda /h/ was lost in all environments (both word-finally and word-internally), and the loss caused compensatory lengthening of the previous vowel. This produced long vowel phonemes (and there are minimal pairs). (Note: This change occurred after rule 1, so a voiceless stop that followed a coda /h/ remained voiceless. It did, however, occur before spirantization, so aspirated stops do spirantize following a long vowel.) An upshot of this change (and this discussion should really be in the section on consonants, but wha'cha gonna do?) was what happened with suffixes. The phoneme /h/ was never at any time allowed as a medial onset. In other words, the sequence /VhV/ was strictly prohibited. It never occurs in stems, but when a word ends in an /h/ and a suffix is added, problems ensue (the same goes for prefixes which are added to /h/-initial words, and this discussion applies to them, as well). First of all, if a suffix of the form -CV is added to an /h/-final word, there is no problem. The reason is that /h/ is perfectly allowable as a coda. In this instance, the /h/ is allowed to stay (this produces forms ending in /-hl/, /-hr/ and /-hʃ/, as well [see rule 5 above]). If a suffix of the form -C is added, though, there is a problem. A suffix of the basic form /-C/ usually has two allomorphs: [-C] and [-VC]. Neither would work. First of all, complex codas are disallowed (remember: the complex codas discussed above were derived), so you can't add [-C]. You can't add [-VC] either, though, because /h/ can't serve as a medial coda. What to do? Well, the result is that the form [-C] will replace the final /h/, so a word that looked like /CVCVh/ will now be [CVCVC], where the final [C] is the suffix. All vocalic suffixes have allomorphs of [-V] and [-CV], so the [-CV] form is used after /h/. Now, with prefixes, all the same rules apply, but backwards. There's an added catch, though. If the prefix is a voiceless stop, it will combine with the /h/ to produce an aspirated stop, since onset /Ch/ sequences are allowed, when the C is a voiceless stop. That does it for /h/.

  2. The vowels /ɔ/ and /a/ change in quality in certain environments. Specifically, when not stressed, or when not directly preceding the stressed vowel, both /ɔ/ and /a/ reduce to [ə].

  3. All long vowels shorten word-finally. The resulting vowels are not subject to rule 2 above, because the shortening happened after reduction. The result is a word-final contrast between [ɔ], [a] and [ə].

  4. In the earliest form of Sathir, the plural was formed by reduplicating the last syllable. This caused stress to switch, since stress was always penultimate. After a time, though, all reduplicated syllables were lost. This means that the singular and plural forms became indistinguishable, save by stress, for all plurals are stressed word-finally.

  5. When stressed as a result of rule 4, the mid vowels /ɛ/ and /ɔ/ raise to [i] and [u], respectively. This rule only applies to vowels that are phonologically short (i.e., it doesn't affect the vowels mentioned in rule 3).

That does it for allophony in Sathir (for the time being).

The Sathir Romanization System

I had to make a lot of tough choices in devising a romanization system to use on this site. But isn't that what designing a web page is all about? After all, it's the web designer, not the law makers or politicians, that make the choices that affect our everyday lives. Choices such as, "Will the picture be centered or not?", and, "Should I italicize this headline?" This is where the action is. That's why I say that America's greatest heroes are the ones who design the websites we surf to (or "at" or "on" or "in") each and every day. Who's with me!

  • The following letters are equivalent in both the phonetic transcription and in the romanization: a, i, u, b, p, t, d, k, g, s, z, x, h, m, n, l, and w.

  • Here are some single-segment changes: [tʃ] will be written č; [dʒ] will be written j; [ɾ] will be written r; [j] will be written y; [ŋ] will be written ng (note: the sequence [ŋg] is impossible in Sathir, so there should be no confusion); [ɸ] will be written f; [β] will be written v; [θ] will be written þ; [ʃ] will be written š; [ʒ] will be written ž; [ɛ] will be written e; and [ɔ] will be written o. The sound [ə] will be written as either a or o, depending on what the underlying vowel is (the pronunciation is predictable, so I won't make a point of representing it in the romanization).

  • Initially, labialized, palatalized and aspirated segments will be written with an orthographic w, y or h. This corresponds to the following: [pʷ] will be pw; [tʷ] will be tw; [kʷ] will be kw; [ðʷ] will be ðw; [sʷ] will be sw; [zʷ] will be zw; [ʍ] will be hw; [mʷ] will be mw; [ŋʷ] will be ngw; [ɾʷ] will be rw; [pʲ] will be py; [bʲ] will be by; [kʲ] will be ky; [gʲ] will be gy; [mʲ] will be my; [ɲ] will be ny; [ɾʲ] will be ry; [pʰ] will be written ph; [tʰ] will be written th; and [kʰ] will be written kh.

  • All geminates will be written as singletons, since voicing will give them away as former geminates. Nasals, however, will be written as geminates. They will look like (in order from labial to velar): mm, nn, nny, and nng.

  • Long vowels will be written as double vowels: aa, ee, ii, oo, uu (this includes word-finally).

  • Words with the sequences /lwV/ and /ljV/ will be written out as lwV and lyV, rather than as wV and yV. They will still, of course, be pronounced [wV] and [jV], respectively.

  • Finally, stress will be marked with an acute accent on words where the stress is not penultimate and is not predictable. [Thus, most plurals will be marked for stress.] The vowels will look like this: á, áá, é, éé, í, íí, ó, óó, ú, and úú.

All right, that's it. Now, I know I've said it before, and I'm sure I'm fated to say it again, but the way I write "Sathir" isn't the way it should be written in the romanization. In fact, it should be written Saþír. As I've said before, though, I prefer to keep my irregular spelling, and will only use the romanization if I use the name of the language in an example using the language itself. Now I entreat you to explore the orthography of Sathir, or, if you want to see the romanization in action, take a look at the Swadesh list. With that, I bid you a happy day.

Back to Sathir Main

This page was last modified on Friday, August 3, 2018.
This website was last modified on .
This page can be viewed normally, as a milk or dark chocolate bar, in sleek black and white, or in many other ways!
All languages, fonts, pictures, and other materials copyright © 2003- David J. Peterson.

free counters