Tan Tyls Phonology

Lots of phonological monkey business going on with Tan Tyls. I'll do my best to describe it all in as succinct a manner as possible, without sacrificing clarity. Stranger things have happened.


  Labial Alveolar Post-Alveolar Velar Post-Velar
Stops p, b t, d   k, g q, ɢ
Fricatives   s, z ʃ x h
Affricates   ts kx qx
Nasals m n   ŋ ɴ
Approximants   l j w ʕ


  Front Mid Back
High i ɨ, ʉ u
Mid e ə o
Low   a  

Consonantal Allophonic Variation

That table up there looks deceptively small. It could have looked even deceptively smaller, but I decided to let affricates be their own category. It also could have looked enormous, but I thought enormosity was not a plus when it comes to tables. Thus, what you have in the consonant table above is almost a phoneme inventory rather than a sound inventory. In this section, I'll explain exactly how it all works.

  1. Nasal Assimilation: The nasal phoneme /n/ assimilates in place to any non-labial consonant which follows it. This nasal will always be written as n, though.

  2. Voicing Assimilation: When a stop (all of which are inherently voiceless) follows a nasal and does not precede a word boundary, it voices to assimilate in voicing quality with the nasal. Additionally, the fricative /s/ can occasionally voice to [z] when it follows a nasal. These changes take place before any type of epenthesis.

  3. Pharyngealization: All consonants to the right of [ʕ] pharyngealize, save the back consonants /q/ and /h/. Additionally, these consonants block the spread of pharyngealization to any other consonants to their right.

  4. Palatalization: All non-approximant consonants palatalize preceding the phoneme /j/.

  5. Labialization: All non-approximant consonants labialize preceding the phoneme /w/.

  6. Aspiration: All non-approximant consonants become aspirated preceding the phoneme /h/, save the phoneme /h/. This includes fricatives and nasals (which devoice).

That's all there is to say about the consonants of Tan Tyls for now. There will be more to say about them soon. Now for their romanization.

Romanization of Consonants

The romanization of the consonants of Tan Tyls if fairly straightforward, thankfully. The vowels are a different story. For now, the consonants.

  • The following consonants are to be spelled exactly as their transcription suggests: p, b, t, d, k, g, q, s, x, h, m, n, l, and w. To the extent that the sound [z] is indicated at all on these pages, it will be spelled z. In most cases, s will suffice.

  • The following sounds are spelled with a letter that features a fancy diacritic: [ɢ] = ġ; and [ʃ] = š.

  • The following sounds are spelled with an entirely different letter: [j] = y; and [ʕ] = '.

  • A palatalized consonant will be spelled Cy; a labialized consonant Cw; an aspirated consonant Ch; and a pharyngealized consonant C'. [Note: A pharyngealization mark will be written the first time only, as everything to the right of it is automatically pharyngealized. Two or more marks in a word will only occur if the spread of pharyngealization is stopped blocked by a back consonant and then started by another pharyngeal approximant later on in the word. An example would be: 'áqhat'ó.]

  • Affricates will be romanized simply as a series of two sounds. So [tʃ], for example, will be romanized , instead of the ever-popular č.

  • The sounds [ŋ] and [ɴ] will each be written as n, as they only occur before homorganic stops.

So, aside from a few wrinkles (like a need for a dot above a "g"), the romanization system for consonants in Tan Tyls is pretty easy to get a handle on. Now for the vowels.

Vocalic Allophonic Variation

The table for the vowels also, in my opinion, looks deceptively small. There's some interesting stuff that goes on with the vowels of Tan Tyls. Interesting minute stuff. Below is a summary of that stuff.

  1. Vowel Lowering I: A high vowel lowers to a mid vowel when it follows a back consonant other than /h/ (i.e., /q/ or /ʕ/) or a pharyngealized consonant. This causes /i/ to become [e]; /u/ to become [o]; and both /ɨ/ and /ʉ/ to become [ə].

  2. Vowel Lowering II: The high vowels /i/ and /u/ lower to [e] and [o] respectively when they follow /j/ and /w/, respectively.

  3. Nasalization: All vowels nasalize to the right of a nasal in onset position, or to the left of a nasal in coda position. Nasalization spreads from left to right, but is blocked by a stop (voiced or voiceless).

  4. Reduced Vowel Harmony: Reduced vowels harmonize in height and rounding. A reduced vowel will be realized as high, the closest vowel to the left is a high vowel, or if there is a preceding coda glide [w] or [j]. If the high vowel or glide is unrounded, the reduced vowel will be an unrounded vowel, [ɨ]. If the high vowel or glide is rounded, the reduced vowel will be a rounded vowel [ʉ]. Additionally, a preceding coda [ʕ] will cause the reduced vowel to be realized as [ə]. This height and rounding spreads to all other reduced vowels from left to right.

There aren't many rules, but they can get complex when they start playing together. Now for the romanization sytem.

Romanization of Vowels

It was extremely difficult making a decision on how to romanize the vowels. One system would have necessitated the use of four orthographic vowels in a row. As a result, I had to make some tough choices when finalizing the romanization system for the vowels. This is the result of that tortuous process.

  • Reduced vowels will be represented with single, unaccented characters. The vowel [ə] will be represented by a; [ɨ] by i; and [ʉ] by u. The nasalized versions of these vowels will be (in order) ã, ĩ, and ũ.

  • Standard vowels will be represented with their normal characters with an accute accent placed above them: á, é, í, ó, and ú. The nasalized versions of these vowels will be represented with a circumflex accent: â, ê, î, ô, and û.

  • Long vowels will be represented by doubling the character in question. The standard vowels look like this: áá, éé, íí, óó, and úú. The nasalized versions of these vowels, then, look like this: ââ, êê, îî, ôô, and ûû.

So that's the romanization system for vowels. Admittedly, this makes it so that a lot of vowels will have accents. This is really the lesser of three evils, though, in my opinion. Anyway, as a result, the name "Tan Tyls" should probably really be spelled Tândyals (showing how it's related to its previous incarnation, Dangelis). Meh: wha'cha gonna do? Let the good times roll!

Word Building

Admittedly, up to this point, the phonology might seem a bit tame compared to other phonologies. Here is where the real stuff starts.

In Tan Tyls, there are really only two types of segments: Reduced vowels, and everything else. Consonants and vowels are treated alike in many ways. The set of consonants and vowels are separated into five main groups:

  1. Stops: p, t, k, q
  2. Fricatives: s, š, x, h
  3. Nasals: m, n
  4. Approximants: l, y, w, '
  5. Vowels: á, é, í, ó, ú

You can remember these five by the word they spell: sfnav. Or, if you prefer, you may also remember them with the acronym "Succulent Fudge Never Amasses Victuals". These are the two officially sanctioned mneumonic devices for remembering the five main groups of Tan Tyls phonological segments. All others must seek approval from the Board of Internally Tracked and Calculated Handballs, headed by the inimitable David Bowie.

So all words (or roots, rather) in Tan Tyls are composed of between two and five segments from the listing above. There are no rules about which ones you can have, in theory. In practice, there is no word with five m's, or four á's and a í. There is nothing to rules these out, though. Nevertheless, generalizations can be arrived at, concerning the number of consonants, which ones occur next to each other, how many you can have in a row, etc. I'll leave that one as a mystery to be solved by the curious, though.

The interesting stuff comes in when you actually take a root and try to pronounce. This is where the phonology happens. So let's take a random root like P-L-Y-K-M. This isn't actually pronounceable, because there's no vowel. So what you have to do is syllabify it and add some vowels. When you do this, you get the pronounceable word playkĩm. But how does one arrive at this word, and not something else? This will be explained below.

In building pronounceable words from a given set of roots, there are a set of rules that must be followed. The first rule is that, if possible, words of the following type should be built: CCVC(C)—repeating, of course. Thus, you always try to start with a consonant cluster. With the root P-L-Y-K-M, that works, because pl is an acceptable onset. With a root like W-T-K-S-Y, though, it doesn't work, and you end up with wataksay. But how does one know if it doesn't work? This is how.

Given a sequence of two segments (based on the five groups above), you can tell whether or not an epenthetic vowel will need to be inserted or not. To follow is a list of twenty five rules which explain how epenthesis works in Tan Tyls. In the lists below, S = stop; F = fricative; N = nasal; A = approximant; and V = vowel. Here are the lists:

  1. Stops
    1. _S: Epenthesize, unless the result is a non-word-final geminate, or unless for CxCy, Cy is [-back, -velar, αplace] and Cx is [-αplace], and the cluster is non-initial and non-final.
    2. _F: Epenthesize if #_, unless the two members are: t + š; t + s; k + x; q + x; or S + h.
    3. _N: Epenthesize unless medial.
    4. _A: Always allowed.
    5. _V: Always allowed.

  2. Fricatives
    1. _S: Epenthesize if #_, unless F is s or š.
    2. _F: Epenthesize unless the two members form a medial geminate, or the second member is h (and the first isn't).
    3. _N: Epenthesize if _#.
    4. _A: Always allowed.
    5. _V: Always allowed.

  3. Nasals
    1. _S: Epenthesize if N is [αlabial] and S is [-αlabial].
    2. _F: Epenthesize if non-medial, and also if m is followed by s, š, or x.
    3. _N: Epenthesize unless medial geminate.
    4. _A: Always allowed.
    5. _V: Always allowed.

  4. Approximants
    1. _S: Epenthesize if #_.
    2. _F: Epenthesize if #_.
    3. _N: Epenthesize if #_.
    4. _A: Epenthesize if non-medial.
    5. _V: Always allowed.

  5. Vowels
    1. _S: Always allowed.
    2. _F: Always allowed.
    3. _N: Always allowed.
    4. _A: Always allowed.
    5. _V: Always allowed.

As you can see, vowels and consonants aren't treated completely equally. But that would just be downright nutty. Or would it...?!

There does exist one exception, however, and it has to do with geminates. Take a form like X-S-S-Q. By the rules we've devised, the result should be xasasaq. It is not, however, because unless geminates occur word-finally or word-initially, they tend to stick together like glue. Thus, you keep the two s's together, and you get the form xassaq.

An interesting tidbit. Say you have the word N-P-M-S-I (the root for "man"). Remembering rule 2 from the consonantal allophony section above, you'll recall that the p becomes a b. Thus, at the initial breakdown, you get nbVmsí. Consonant clusters like nb (and ms), though, simply aren't tolerated, so a schwa must be inserted. The result is nãbamãsî, with a voiced b.

Another interesting note that has to do with the section below. Take a root like S-S-N-M-P, "dust". A number of rules are involved here. First, the word breaks down to ssVnmVp. Both consonant clusters are disallowed, though, so you get sVsVnVmVp. Thus, you'd predict sasanãmãp. However, there's a rule which deletes all unstressed schwas where permissible (i.e., where the result is a permissible consonant cluster). Thus, what you in fact get is sasanãmp. This is a rule which doesn't come into play that often, but which is consistent and very much a part of the language.

Anyway, a question arises. Given a root like Y-K-T-S-W, is the resulting word yaktasw or yakatsaw? That question can be answered via the notion of rule ordering, in a way. Essentially, the first step is to try to break the word up into CCVCCVCCVCC… form. So given the root Y-K-T-S-W, you get ykVtsVw. Now you evaluate that form. Is it acceptable? No. You need to break up that initial cluster. So you break it up to get yVkVtsVw. Now you asign the actual vowels to get the form yakatsaw. There's a missing step in this three step process, though. Between steps one and two stress is assigned. So the process is actually comprised of four steps. This next section is devoted to that second step.


Stress is uniformly assigned to the penultimate vowel in a form after its been broken up into CCVCC form, but before epenthesis occurs. This is important because of a curious phenomenon I call stress spreading (after all, what good is stress if you have to keep it to yourself? You know the song: Stress is something if you give it away! Give it away! Give it away!). I'll explain briefly.

First, stress is realized by a slight increase in volume, and also a higher intonation. So, in a word like N-T-S-K-T, you first get ntVskVt, you assign stress to the first vowel, and you get ndaskat, with stress on the first syllable. In a form like the one above (Y-K-T-S-W), though, if stress is assigned to at the second stage (ykVtsVw), then stress is put on the first syllable of a disyllabic form. Epenthesis rules, however, force the initial cluster to be broken up. As a result, the form becomes yVkVtsVw (and eventually yakatsaw), and both the second and the first vowel are stressed. What does this mean? This means that the first two syllables have a higher intonation than the last.

Compare the form just discussed to the form M-P-S-A-U. Here, the form is broken down as mpVsVV, and stress is assigned to the penultimate vowel, á, resulting in a trisyllabic form mbasáú, where the first syllable does not have a high pitch. Thus, you can have tonal minimal pairs (though no true minimal pairs). An additional tonal note is that if a word ends in a long vowel, this long vowel will be realized with a falling tone. The system is still a stress system, though, and otherwise acts as such.


What a long strange trip it's been... Strange, and entirely uneventful. What has happened between here and the beginning of this page? Actually happened? There's a question.

Back to Tan Tyls Main

This page was last modified on Thursday, January 13, 2011.
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