zompist bboard

To put it in as few words as possible, this thread is all about conlang sketches I'm working on at the moment.

Tantšak Notes

1. Tumetıęk, the Parent Language
2. Sound Changes
3. Phonology

Areyatta Notes

1. Proto-Areyattan Basics – Phoneme Inventory and Verbs

Sound Changes
Last week I have been working on and off on a descendant of my language Tumetıęk. So far I've decided to name the said descendant Tantšak, and sound changes have been slowly taking shape over last few days. Today I finally managed to get them in a more-or-less presentable state, so I decided to open a thread here to present them. The time depth for these sound changes is intended to be 1200 years or so.

Some allophonic changes described in the grammar are ignored here because I found it easier to describe the sound changes if I omit them, and the outcome is the same.
If not stated otherwise, sound changes affecting oral vowels affect nasal vowels as well.

Let us start with describing the notation used in the sound changes. It is for most part quite straightforward, but there are nonetheless several oddities, such as sonority marking.

• C – consonant
  P – plosive
  N – nasal
  V – vowel
  
  Ø – absence of a phoneme
  ~ – free variation
  # – word boundary
  
  $ – syllable
  . – syllable boundary
  
  [±V] – velarity
  [±L] – labiality
  [±N] – nasalisation
  [±R] – rounding
  [±S] – stress
  [±SN] – sonority
  [±V] – voicing
  
  [+_] – presence of a feature
  [-_] – absence of a feature
  
  >[+SN] – the following consonant is less sonorous
  <[+SN] – the following consonant is more sonorous

If a feature becomes absent, it is simply left out: for example, vowel denasalisation is written V[+N] → V, instead of V[+N] → V[-N].

Allophony
Voiceless plosives aspirate in anlaut when followed by a vowel, and certain vowels undergo a quality change. The palatal /j/ labialises before a rounded vowel, in this case /u/ and its nasal counterpart /ũ/.

1. P[-V] → Pʰ / #_V
2. /e ẽ a ã/ → [ɛ ɛ̃ ɑ ɑ̃]
3. /j/ → [ɥ] / #_V[+R]

Nasalisation
The first sound change which we can securely ascertain was fairly minor: nasals in a complex syllable coda nasalised the prevoius vowel and then dropped.

1. V → V[+N] / _NC.
2. N → Ø / _C.

Lenition
Voiced stops lenite to their corresponding fricatives, while the sibilant [z] lenites to [r]. The resulting fricatives change further, becoming more sonorous.

1. [d] [ɡ] [z] → [β] [ð] [ɣ] [r]
2. [β] → [ʋ]
3. [ð] → [r]
4. [ɣ] → [j]

Retroflexion
All coronal and dorsal consonants, except the already palatal /j/, palatalise before front vowels. If a rising diphthong beginning with any of /i ɛ/ followed a palatalisable consonant, it fused with that consonant into a palatalised one.
The palatal consonants subsequently depalatalised to corresponding retroflexes.

1. [n{i ɛ}] [t{i ɛ}] [k{i ɛ}] [s{i ɛ}] [r{i ɛ}] → [n̠ʲ] [t̠ʲ] [c] [s̠ʲ] [r̠ʲ] / _V
2. [n] [t] [k] [s] [r] → [n̠ʲ] [t̠ʲ] [c] [s̠ʲ] [r̠ʲ] / _{ [ɛ]}
3. [c] → [t̠ʲ]
4. [t̠ʲ] → [ts̠ʲ]
5. Cʲ → C
6. [n̠] [ts̠] [s̠] [r̠] → [ɳ] [tʂ] [ʂ] [ɽ]

The sound changes onward are mainly concerned with vowel development and subsequent cluster simplification.

Vowel Deletion
Unstressed pretonic vowels, including diphthongs, simply drop.

1. V[-S] → Ø / _$[+S]

Vowel Coalescence
Onglides are lost altogether, but offglides coalesce with the prevoius vowel.

1. V̯ → Ø / _V
2. {[ai̯] [ei̯]} {[au̯] [eu]} → [e]

Vowel Shift
The back vowels /ɑ/ /u/ shift to /o/ /ɑ/, respectively. Resulting instances of [ɥ] before an unrounded vowel and [j] before a rounded vowel are leveled out.

1. [ɑ] → [ɒ ~ ɔ]
2. { [ʊ]} → [ɨ ~ ə]
3. [ɨ ~ ə] → [a]
4. [ɒ ~ ɔ] → [o]
5. [a] → [ɑ]
6. [ɥ] → [j]
7. [j] → [ɥ] / _V[+R]

Vowel Denasalisation
Nasal vowels lose their nasalisation without any change in quality, except for [ɛ̃], which changes into [ɑ̃] and then into [ɑ].

1. [ɛ̃] → [ɑ̃]
2. V[+N] → V

Vowel Creation
The recent vowel deletion resulted in some cases with large consonant clusters which did not obey the sonority hierarchy.
Word-initially, clusters with three consonants, and those which did not obey the sonority hierarchy, acquired a copy vowel before them.
Consonants became syllabic if less sonorous consonants surrounded them. These syllabic consonants later vocalised to /ɑ/.
Finally, an epenthetic glide was inserted before a word-initial vowel. The choice of the glide depended on vowel quality: unrounded vowels acquired an epenthetic [j], while the only rounded vowel [o] acquired an epenthetic [ʋ]. The glides were then reinterpreted as phonemic.

1. Ø → V₁ / #_CCCV₁
2. Ø → V₁ / #_C>[+SN]V₁
3. C → C̩ / C<[+SN]C>[+SN]C
4. [j̩] [ʋ̩] → [o]
5. C̩ → [ɑ]
6. [ɑ] [ɛ] [o] → [jɑ] [jɛ] [ji] [ʋo] / #_

Cluster Simplification
Certain disallowed clusters undergo simplification or assimilation: double consonants resolve into single ones, nasals assimilate in place of articulation to the following consonants, etc.

1. [m] → [n] / _C[-L] (Doesn't affect [mn] [mr].)
2. {[n] [ɳ]} → [m] / _C[+L]
3. C[+K] → Ø / _C[+K] / _C[+L]
4. C[+L] → Ø / _C[+K] / _C[+L]
5. C₁C₁ → C₁
6. C → Ø / _CC (Except in [{s ʂ}{p t k}{ʋ r j}], and intervocalically in clusters consisting of a nasal followed by a fricative+stop, stop+fricative or obstruent+one of [ʋ] [r] [j].)

Comments, suggestions, and questions are all welcome.

Would be nice if you could cite some examples just to make things all the more clear. Moving forward, overall very common changes from what I can tell, but they seem to come together well.

Cael wrote:Would be nice if you could cite some examples just to make things all the more clear. Moving forward, overall very common changes from what I can tell, but they seem to come together well.

Thanks!

I've been quite busy with schoolwork this week, so I couldn't post some examples. Going to post them quite soon, though.

bump, will try to make stuff presentable

It took much longer than I expected because school, but I've managed to get the phonology of Tantšak in a presentable state right on my fifteenth birthday. Well-timed, I guess.

Phonology
Obviously enough, this section describes the phonology of Tantšak.

Phoneme Inventory
The phoneme inventory of Tantšak is quite small, although not excessively so, with a total of seventeen phonemes. Thirteen of these are consonants and the other four are vowels, so the consonant-vowel ratio is near-average at 4.25.

Consonants
Tantšak has a quite small consonant inventory with a total of thirteen phonemic consonants. They are listed below.

Obstruents: /p t s (tʂ) ʂ k/
Sonorants: /m n ɳ r ɽ ʋ j/

• – The retroflexes are written with a háček (caron): ň tš š ř.
  – The glides are written v y.
  – The transcription otherwise complies with IPA.
  – The affricate /tʂ/ is marginally phonemic – it is listed as an independent phoneme only because it can appear in syllable codae where consonant clusters are disallowed.

Onward to the allophony.

The retroflex consonants are usually pronounced as unpalatalised postalveolar [n̠], [ts̠], [s̠] and [r̠], and /j/ is realised as palatalisation on the previous consonant. Retroflexes palatalise to [n̠ʲ], [ts̠ʲ], [s̠ʲ] and [r̠ʲ], and the sole velar palatalises to [k̟].
Stops aspirate to [pʰ], [tʰ], [ts̠ʰ] and [kʰ] in anlaut, i.e. word-initially, before a vowel.
Obstruents voice after sonorants and intervocalically in unstressed syllables.
Nasals assimilate to the place of articulation of the following consonant.
The glide /j/ labialises to [ɥ] before the sole rounded vowel /o/. If it follows a consonant, it is realised as labiopalatalisation on that consonant.

Vowels
The four vowels of Tantšak are shown below.

Vowels: /i e a o/

• – The vowel /i/ is written without the dot – ı.

Now the allophony; the "mid" vowels /e/ and /o/ aren't equal in closeness. The former is realised as a low [ɛ] and the latter is realised as a true close-mid [o]. These two vowels are reduced to [ɜ] and [ʊ] ~ [ɵ] in unstressed syllables. For /o/, the latter unstressed pronunciation is preferred over the former.
The open vowel /a/ is pushed towards [ɑ] because of proximity of /e/, and it reduces to [ʌ] in unstressed syllables.
The vowel /i/ reduces to [ɪ] in unstressed syllables.
Unstressed vowels nasalise before nasals without any additional quality change.

Stress
Tantšak has a dynamic-accentual stress system, i.e. stressed syllables are louder and pronounced with a higher pitch than unstressed ones.
Stress is quite variable, but not phonemic at all. Minimal pairs are very rare and usually come from borrowings. It is indicated with an acute accent on the stressed vowel: á é í ó, except in monosyllabic words.

Phonotactics
The syllable structure of Tantšak can be summarised as (C)(C)CV(C). This is very oversimplified, though, and it should be noted that there are many constraints which prohibit syllables such as /ssop/, /pka/ and /ttʂi/.

Tantšak has a moderate amount of onset clusters, which fit certain patterns, listed below.

A sibilant followed by a stop: /sp/, /st/, /sk/, /ʂp/, /ʂt/ and /ʂk/
A stop followed by a sibilant: /ps/, /ts/, /ks/, /pʂ/, /tʂ/ and /kʂ/
An obstruent followed by a nasal: /pn/, /tm/, /tn/, /sm/, /sn/, /ʂm/, /ʂn/ and /kn/
An obstruent followed by a rhotic: /pr/, /tr/, /sr/, /kr/, /pɽ/, /tɽ/, /sɽ/ and /kɽ/
Any of the preceding (except obstruent-nasal clusters) or a single consonant followed by one of /ʋ/ and /j/.

Syllable nuclei are limited to vowels. The coda is optional, but every consonant can be there.

Bump. A post on nouns is in the making.

Time to stop neglecting this thread. I haven't posted here in quite a while, but I guess that it is nonetheless worth saying that I am still working on Tantšak. Perhaps most obviously, the name changed from Tantšák because stress is not phonemic, despite being unpredictable, so I decided there's no need to mark it everywhere.

Tumetıęk, the parent language
This is precisely what I've been working on over last month and half. As said in the first post, Tumetıęk is the ancestor language of Tantšak, developed for 2012/2013 Reconstruction Game, and it was for most part developed this August.

However, some two months ago I felt I needed a more solid base to derive Tantšak from, so I went on a endeavour to improve the grammar with better descriptions and tighten up some loose ends. This resulted in me reworking the grammar (PDF, 28 pages) of the parent language to finally have a solid base from which to work on, complemented with a 400-odd word lexicon (PDF, 15 pages). I'm quite proud of my clickable table of contents.

Let's head for a short description of Tumetıęk with snippets from the grammar.

• Tumetıęk has a small vowel inventory which numbers only four distinct vowel qualities. These four phonemic vowels are organised into an imperfect square-shaped system [...]

  Tumetıęk can be said to have a small consonant inventory which consists of only twelve constituent phonemes in total. [...]

This is pretty much the entire pholosophy 'bout the phoneme inventory. There are sixteen distinct basic phonemes which I list below.

• /m n/ m n
  /p b t d k ɡ/ p b t d k g
  /s z/ s z
  /r j/ r y
  
  /i e a u/ ı e a u

These four monophthongs can coälesce together to form a total of twelve diphthongs: ae, aı, au, ea, eı, eu, ıa, ıe, ıu, ua, ue, uı. Some of them are very infrequent and appear in only a handful of words.
Vowels can appear nasalised, in which case they are marked with an ogonek (tail). This extends to diphthongs, of course, but nasalisation is not marked on the glide: ąe, ıų.

There is some fancy allophony, which makes words such as kıayu and nunıegıand be realised as [ˈcʰæ̞ɥʊ] and [nʊnɪ̯ɛˈʝæ̞nd]. It is explained in detail in the grammar under 1.1.1. Vowels and 1.1.2. Consonants.

The canonical syllable structure is (C)(C)N(C)(C), but there is only a moderate amount of legal consonant clusters, which are exceedingly rare in syllable coda. Strict rules govern which clusters are legal, and illegal clusters are resolved through complex rules described in 1.5. Moprohophonology.

Stress is weight-sensitive and governed by a set of three rules, described in 1.3. Stress.

Being a heavily agglutinative/fusional language, Tumetıęk uses rich inflection to convey various grammatical functions and add shades of meaning to words. Perhaps my favourite feature of the system is how verbs inflected with a secondary voice acquire an impersonal meaning when they take on the passive voice suffix -un.

• Taneun zı̨rıu.
  3-INE-live-PASS house
  One lives in a house. (A house is lived in.)

Nouns inflect for pronominal possession, case and deixis, the latter two of which are marked with a set of fusional suffixes that encode both. There are three cases: nominative, genitive and accusative and three degrees of distance: proximal, medial and distal.
Combined case-deixis suffixes can render the root noun a bit hard to recognise: taking psųdıebra “bridge” bridge as example and adding the proximal suffix -·`tı we get psųdıeprı “this bridge” instead of a hypothetical *psųdıetı (← *psųdıebrtı).
The entire system of noun inflection is covered in 2.1. Nouns.

On a first glance, Tumetıęk has a very minimal pronoun system, with only three pronouns: the first person pronoun, the second person one plus the interrogative one. This is somewhat complicated by all of them having a pervasive number distinction, something not seen on nouns. Moreover, personal pronouns have distinct polite forms: nı̨ru, tęru, nąru and kıęru, as opposed to casual tı, sı, nı and grı.
Unlike nouns, pronouns can't take any possessive or deictic marking at all.

Verbs are very complex, inflecting for a total of six grammatical categories, displayed in the following template. They are described in 3. Verbs.

• -2 – subject, person and tense
  -1 – secondary voice
  0 – STEM
  +1 – mood
  +2 – primary voice

The voice system is best described with the following snippet.

• The voice system of Tumetıęk is very large and as such it spans two slots in the verb template. Voice inflection is carried out on two tiers: the primary tier, which hosts two primary voices and the secondary tier, which hosts nine secondary voices.
  Tumetıęk has a total of eleven voices including both primary ones and secondary ones, so the voice system is very large when compared with an average one.

There are only two voices on the primary tier. The active is unmarked while the passive is parked with -un.

Also dubbed applicatives, the secondary voices appear in the second slot of the verb template. They promote an indirect or otherwise oblique argument to direct object. They can only be used with intransitive verbs, except for the dative -sa-. All secondary voice prefixes are listed under 3.3.2. Secondary Tier.

Let's close off the talk about verbs with a short description of the moods, again conveniently snipped from the full grammar. A list of modal suffixes is available under 3.2. Mood.

• The first suffixal slot in the Tumetıęk verb template is occupied by modal suffixes. There is a total of four moods: realis, irrealis, imperative and optative.
  The realis mood is unmarked, and as such it is usually left out from the interlinear glosses.
  The irrealis mood is heavily used, and does not map exactly to English usage of irrealis modals such as “would”; an average speaker would use the irrealis to add even a slightest element of doubt to their utterance in all but most formal speech.
  Verbs conjugated into the imperative mood accept only second person subjects.

There are several derivational suffixes, but they are in my opinion nor worthy of describing in this short summary, so let's carry on to 5. Syntax.

Starting with describing noun phrase syntax as I've done in the grammar, I should give its structure which, by the way, displays quite equally distributed left- and right-branching: (REL) (DET) (APP) NOUN (PREP) (REL).

Honestly, there's not much to say about this besides the two following snippets about 5.1.1. Relative Clauses and 5.1.2. Determiners.

• Tumetıęk employs the gap strategy to form relative clauses, and this is described in more detail under 5.3. Clause Linking.
  Relative clauses can either precede or follow the head noun, depending on their weight. Relative clauses consisting of a single verb are counted as light, and all other relative clauses are counted as heavy.
  Heavy relative clauses mandatorily follow the head noun; light relative clauses can either precede the head noun or follow it, with former being preferred.

  Determiners constitute a small closed word class distinct from both verbs and nouns. This class includes quantifiers, numerals, and specific words describing size or age such as betre “old” and pı̨dre “big; large”.

Possession is a bit complex, and it comes in two flavours: alienable possession and inalienable possession.

• Alienable possession is handled by placing the possessor in a prepositional phrase introduced by the preposition kı “of” and inflecting the possessor into the genitive.

  On the other hand, inalienable possession is handled in a much simpler manner – the possessor is placed in apposition with the possessee.

  Pronominal possession, regardless of alienability, is handled with possessive prefixes.

Moving onwards to word order, described in 5.2.1. Word Order.

The basic and most common word order in a Tumetıęk clause is S Aux V O, and the language handles indirect objects through a double object construction. There is a modification to that, snipped from the grammar.

• The most common variation on the word order is Aux V P A, ubiquitous in passive clauses because it gives the patient a greater prominence than to the agent.

Adverbial phrases immediately follow the last verb in the verb phrase, noting that

• Tumetıęk productively forms adverbials out of verbs by nominalising the verb with the abstract nominaliser -rıa. The resulting noun is placed in a prepositional phrase introduced by the appropriate preposition, most commonly dru “in”.

The auxiliary verbs are as complex as they can be, so I've elected to solely use snippets to describe them.

• Tumetıęk makes frequent use of auxiliary verbs to convey information which verbal morphology is not able to indicate, and they are often stacked next to one another to convey richer and more complex meanings.
  They make the main verb non-finite, so it has to take the infinitive suffix •ta.
  Auxiliary verbs do not constitute a distinct verb class – it is more that certain verbs, in fact eight of them, can be used both indepedently and as auxiliaries.
  Auxiliaries are conventionally divided into four groups: modal, aspectual, tensal and polar auxiliaries, the distinction between which is based on the distance between the main verb and the auxiliary, as described below.
  Tensals are closest to the main verb, and modals are furthest away. Aspectuals come in between these, so a general ordering of these three is modal-aspectual-tensal.
  The sole polar auxiliary drık can come before any other auxiliary, or in the absence thereof, immediately before the main verb. It negates all verbs, both full and auxiliary, which follow it.

Tumetıęk employs several question-forming strategies which fall under one of two types of questions: polar (yes-or-no) and content questions. The strategies, one by one:

• Tumetıęk usually forms polar questions using the particle gıa.
  The particle is placed clause-initially to question the clause as whole, but it can also be placed before a specific constituent to inquire whether the latter is/was involved in the event described, in which case it is cliticised to that constituent.

  There are two more interrogative particles, which can be used to add finer shades to meaning of the question. These can appear in same positions as gıa.
  

  • dra –The rhetorical question particle; indicates that the speaker is not really asking for information; may be associated with both doubt or support of the expected answer. Glossed RHET.
    uk –The possibility question particle; does not ask for a statement about reality, but about the listener's assessment of likelihood of the described situation. Glossed POS.

  Tumetıęk forms content questions by replacing the questioned constituent with the interrogative particle tın, which takes appropriate morphological marking if needed. However, the interrogative pronoun kıe and its inflected forms may also be used. There is no wh-movement whatsoever.

And now onward to 5.3. Clause Linking.

Tumetıęk strongly prefers coordination over subordination to the point that the only types of subordinate clauses are complement and relative clauses, both of which are formed using the gap strategy. Complement clauses take the place of an object of the matrix clause; placement of relative clauses is described in detail in 5.1. Noun Phrase Syntax section.
Other kinds of relation between clauses are expressed by simply juxtaposing them.

Bump, with some hopefully cool new content! [B)]

Proto-Areyattan Basics – Phoneme Inventory and Verbs
The phoneme inventory of Proto-Areyattan is quite small – ten consonants and four vowels.

• p t k s r v y ɣ m n
  ı u ə a
  
  • – v, y, ɣ and ı are /w/, /j/, /ɰ/ and /i/. The transcription otherwise complies with IPA.

Stress is phonemic and this marked with an acute accent on the stressed vowel: í, ú, ə́ and á.
Syllabic allophones of /m/, /n/ and /r/ are marked with an underdot: ṃ, ṇ and ṛ.

Several vowels and sequences contract in syllables made unstressed by morphophonological processes, only if adjacent to the stressed syllable:

• ə{m|n|r} → ṃ|ṇ|ṛ | if tautosyllabic¹
  {m|n|r}ə → ṃ|ṇ|ṛ | if tautosyllabic²
  
  {ı|u|a}{m|n|r} → ə{m|n|r}
  {m|n|r}{ı|u|a} → {m|n|r}ə
  
  {ı|u|ə|a}y {ı|u|ə|a}ɣ {ı|u|ə|a}v → ı a u | except prevocalically³
  y{ı|u|ə|a} ɣ{ı|u|ə|a} v{ı|u|ə|a} → ı a u | except in anlaut⁴ and postvocalically⁵
  
  {ı|u|a} → ə

Further phonological processes:

• C̣V → C̣ɣ̆V
  VC̣ → Vɣ̆C̣
  
  {əa|aə} → a
  
  {ı|u} → {y|v} | #_V
  
  C₁C₁ → C₁
  V₁V₁ → V₁
  V₁V₁ → V₁

Verbs inflect for number, person and mood.

The singular number is unmarked, while the plural is formed by reduplicating the first syllable of the root. The reduplicant⁶ is always stressed in past tense, but stress does not relocate in the present.
In case the first syllable begins with a cluster, only the most sonorous consonant is reduplicated. If all consonants in a cluster are equally sonorous, the only the last consonant in the cluster is reduplicated.
There are two tenses: present and past. The present is unmarked, while the suffix -ná marks the past tense. The a is suppressed before an another vowel, but stress is still carried onto that vowel.
Mood suffixes follow the tense suffix if there is one. There are three moods: realis, irrealis and imperative. The realis is unmarked, the irrealis is marked with the suffix -ak, and the imperative is marked with -ı, which changes to -yı postvocalically⁷.

Also, an example of conjugation for sər 'stand'.

• Realis: sər, sṛsə́r, sṛná, sərsṛná
  Irrealis: sə́rak, sṛsə́rak, sṛɣ̆ák, səsṛɣ̆ák
  Imperative: sə́rı, sṛsə́rı, sṛɣ̆í, sərsṛɣ̆í

• ɣ̆ is an epenthetic consonant of a rather unclear quality which appears between a syllabic consonant and a vowel.

• ¹ ² In same syllable.
  ³ Before a vowel.
  ⁴ Word-initially.
  ⁵ ⁷ After a vowel.
  ⁶ The reduplicated part.

What do you all think? Is everything plausible and sensible?