Tersamian

s̩qx n̲s̲̍ṛnxbx̣pm “Z’om Dä́ŕtmin’ej” /ʑə̃̌m ðǽr̥mîɲɪ̀j/

Tersamian

Tersamian is a southern Skomi language belonging to the Dartminic branch. It is native to the more than 300 million inhabitants of the Ältojowa plain but due to its cultural dominance in the Tersamian Empire, it has come to supplant the local languages of other parts of the Empire.

With at least 3000 years of history, the Tersamian language has undergone massive transformations that makes it unrecognisable from the original Proto-Skomi language. Along the way, the Tersamian language fractured into a rich tapestry of dialects, separated by geographical features like mountains and the vast area of the Ältojowa plain itself. These dialects range from the politically and culturally influential S’ḗnmäic (ṣp̤̍zxi̲o /ɕɛ̃́mæ̂i̯t͡s/) varieties of the plains north of the Jowa mountains, to the Kä́ldäŕt (hs̲̍ṇn̲s̲ṛn /kǽlðæ̂r̥/) vernacular spoken in the wealthy financial and maritime provinces along the Tersamian bay, to the Lóhom (ṇq̍rqx /lɔ́ɦɔ̃̂/) dialect in the heavily industrialised Lóhom river delta. With time, these dialects have developed their own distinctive characters to the point that they are almost entirely mutually unintelligible. In fact, it would be more accurate to call them separate languages under the Ältojowa branch. However, for the purpose of uniting the Empire, these languages are officially considered dialects, while the “language” distinction is reserved for other varieties not native to the Ältojowa plain.

Given the plethora of Tersamian varieties, it would be impossible to document all of them, so this article will only focus on the prestige variety spoken in the capital Ŕáncöem (especially in the neighbourhoods on the west bank of the Ŕáncöem river). This is the variety taught as “Modern Standard Tersamian” and preserves much of the ancient grammar of Middle Tersamian. However, its phonology is the most divergent compared to the other dialect groups. For example, it has lost most of the tenuis vs palatal distinction between consonants and length distinction between vowels. It has also added soft sounds for certain consonants that appeared mainly through proximity with long vowels and developed Japanese-type pitch-accent.

Phonology

Consonants

Modern Standard Tersamian has a large (according to WALS) consonant inventory size between 34 to 36 consonants. All consonants (including approximants) save for the palatal nasal /ɲ/ and velar nasal /ŋ/ have unvoiced vs voiced distinctions.

Consonants		Labial	Coronal	Palatal	Dorsal	Glottal
Nasal	Voiceless	m̥	n̥
Nasal	Voiced	m	n	ɲ	ŋ
Plosive	Voiceless	p	t θ¹	c	k	(ʔ)
Plosive	Voiced	b	ð²	ɟ ~ d͡ʑ³	g
Fricative	Voiceless	f	s	ɕ		h
Fricative	Voiced	v	z	ʑ	ɦ ~ ɣ⁴
Affricate	Voiceless		t͡s	t͡ɕ
Affricate	Voiced		d͡z	d͡ʑ ~ ɟ³
Approximant	Voiceless		l̥	ç	ʍ
Approximant	Voiced		l	j	w
Trill	Voiceless		r̥
Trill	Voiced		r

/θ/ is the soft version of /t/ and is thus traditionally analysed as a plosive.
/ð/ is a lenited form of /d/ and is thus traditionally analysed as a plosive.
Among younger speakers, /ɟ/ and /d͡ʑ/ are slowly beginning to merge.
/ɦ/ is typically pronounced as /ɣ/ in the western districts of Ŕáncöem.

Vowels

_{In this article, “vowels” will serve as an umbrella term for “monophthongs” and “diphthongs”.}

Tersamian has a very large vowel inventory with 20 vowel qualities and 8 diphthongs. Vowels can be divided into “Tenuis” and “Ablaut” categories. In addition, the back vowels can be divided into “short” and “long” vowels. These further distinctions will be described in greater detail further along the line.

In the 2 tables below, each full vowel is listed along with its reduced counterpart in this format: <full> → <reduced>. There is one notable duplicate in this table, /ɵ/, which serves as the associative vowel of /ɔ/ and /o/.

Monophthongs	Front		Back
Monophthongs	Tenuis	Ablaut	Tenuis Short	Tenuis Long	Ablaut Short	Ablaut Long
High	i → ɪ	y → ʏ	u → ʊ	ʉ → ʏ	ʌ → ɨ	ɯ → ɨ
Mid	ɛ → ɪ	e → ɪ	ɔ → ə	ɵ → ə	o → ə	ɵ → ə
Mid		æ → ɛ	a → ɐ	ɑ → ɐ

Mid-High Diphthongs	Front		Back
Mid-High Diphthongs	Tenuis	Ablaut	Tenuis	Ablaut
(High)	→ ɪʊ̯	→ ɪʏ̯	→ ʊɪ̯	→ ʊ
Mid	eo̯ → ɪ̯ə	øː → ɵ	oe̯ → ʊɪ̯	oː → ə
Low	ɑe̯ → ɐɪ̯	æi̯ → ɛɪ̯	ɑu̯ → ɐʊ̯	æɯ̯ → ɛɨ̯

Vowel Mutation

Middle Tersamian had a vowel system where each vowel could be divided into either “Tenuis” or “Ablaut” vowels. These vowels were/are distinguished using umlauts (e.g. “ö”) in Tersamian Romanisation and with a bar (e.g. “s̲”) in T’úrsoomajatta and C’ī́nhṻŕaet̄a. “Tenuis” and their corresponding “Ablaut” counterpart usually have opposite backness. For instance, back “u” has front “y” as its ablaut counterpart. The two exceptions to this rule are /ɑe̯/ vs /æi̯/ and /ɑu̯/ vs /æɯ̯/.

The table below shows a mapping between tenuis and ablaut vowels.

Tenuis	Ablaut
i	ʌ/ɯ
ɛ	o/ɵ
a/ɑ	æ
u/ʉ	y
ɔ/ɵ	e
eo̯	oː
oe̯	øː
ɑe̯	æi̯
ɑu̯	æɯ̯

Short and Long Vowels

Between Middle Tersamian and Modern Standard Tersamian, vowels underwent huge changes. One of these changes was the loss of length distinctions in all vowels. However, the distinction between some pairs of short and long vowels were preserved through a change in vowel quality. Although they are listed as “long” vowels, they are actually the same length as their “short” counterparts. For vowel harmony purposes, short and their corresponding long vowels are considered to have the same height, even though on an IPA vowel chart they are situated at different heights.

Middle Tersamian	Modern Short Counterpart	Modern Long Consonant
ɯ	ʌ	ɯ
o	ɔ	ɵ
ɤ	o	ɵ
ɑ	a	ɑ

Nasal Vowels

Tersamian also has nasal vowels. These are variants of normal vowels in a nasal environment. The process by which nasal vowels appear are discussed in the Nasalisation section.

Non-nasal Vowel	Nasal Vowel
a, ɑ, ɐ	ɑ̃
i, ɪ, e	ɪ̃
ɛ, æ	ɛ̃
u, ʊ, ʉ, o	õ
ɔ	ɔ̃
ə, ɵ, ʌ, ɯ, ɨ	ə̃
y, ʏ	ʏ̃
ɑe̯	ɑ̃ẽ̯
ɐɪ̯	ɐ̃ɪ̯̃
æi̯	æ̃ĩ̯
ɛɪ̯	ɛ̃ɪ̯̃
ɑu̯	ɑ̃ũ̯
ɐʊ̯	ɐ̃ʊ̯̃
æɯ̯	æ̃ɯ̯̃
ɛɨ̯	ɛ̃ɨ̯̃
eo̯	ẽõ̯
ɪ̯ə	ɪ̯̃ə̃
øː	ø̃ː
oe̯	õẽ̯
ʊɪ̯	ʊ̃ɪ̯̃
oː	õː
ɪʊ̯	ɪ̃ʊ̯̃
ɪʏ̯	ɪ̃ʏ̯̃

Vowel Harmony

Tersamian exhibits both height harmony and backness harmony, but never both in the same word. On top of that, vowel harmony can be progressive or regressive. The type of vowel harmony used by each word depends on its word class, more about it in grammar.

Due to great Tersamian vowel shift, some vowels have shifted height or backness, which would make explaining vowel harmony much harder if I were to use phonemic transcription. Instead, I’m using transliterated vowel letters to showcase vowel harmony, as the Tersamian writing systems are much better suited to display the transformations vowels undergo during vowel harmony. See the orthography section for mapping between the writing systems and the IPA representations.

Height Harmony

Tersamian vowels can be divided into 3 heights: Low, Mid and High. There are no neutral vowels. Vowels must travel through defined “channels” when changing heights.

Monophthongs

Height	Front Tenuis	Front Ablaut	Back Tenuis	Back Ablaut
High	i	ü	u	ï
Mid	e	ö	o	ë
Low	ä	ä¹	a	a²

ä can descend from but never ascend to ö or ü.
a can descend from but never ascend to ë or ï.

Diphthongs

Height	Diphthong 1	Diphthong 2	Diphthong 3	Diphthong 4
Mid	eo	öe	oe	ëo
Low	äu	äi	ae	au

Diphthong channels 1 and 3 allow the vowel to change backness as height changes.

Height harmony can cause a vowel to travel all the way to the height of the trigger vowel¹ or only travel by one category to approach the category of the trigger vowel². Whichever happens depends whether vowel harmony is triggered by an inflectional suffix or derivational suffix, as well as when the word entered general use in Tersamian.

Example: Trigger “i” causes target “a” to rise to “u”.
Example: Trigger “i” causes target “a” to rise to “o”.

Backness Harmony

In the following 2 tables, vowels can move from front to back or vice versa if triggered by backness harmony.

Monophthongs

Backness	Front	Back
High	i	ï
High	ü	u
Mid	e	ë
Mid	ö	o
Low	ä	a

Diphthongs

Backness	Front	Back
Mid	eo	ëo
Mid	öe	oe
Low	ae	au
Low	äe	äu

Progressive Harmony

Progressive vowel harmony is usually triggered by the root vowel, which in the majority of cases is the stressed vowel. The target vowels are all vowels after the root vowels until the end of the word.

Regressive Harmony

Regressive vowel harmony is triggered by the last vowel. All vowels preceding it, up to and including the root vowel in the word are target vowels. If there is no root vowel, all vowels become targets.

Pitch Accent

Tersamian has a system of pitch accent that developed from the stress system of Middle Tersamian. This system of pitch accent is closer to that of Japanese, where words are distinguished according to where a fall in pitch occurs (if there is one at all) compared to something like Serbo-Croatian, where one syllable has a certain tone. Tersamian words can have one or zero accents (places of stress), with most having one accent. In multi-syllabic words, this accent is typically realised as a high tone followed by a falling tone, with the remaining vowels carrying a low tone, but could also be realised as a rising tone. It’s complicated.

For Tersamian speakers, word boundaries play an important part in determining what tones stressed and unstressed syllables have. For example TODO and TODO are pronounced /TODO/ and /TODO/ respectively.

Furthermore, speakers also pay attention to what syllables are written or not. For Tersamian, syllables are phoneme sequences with one vowel in them. This is important in distinguishing the difference between orthographical syllables, non-orthographical syllables and non-syllables. Non-orthographical (or unwritten) syllables arose when complex consonant clusters from Middle Tersamian collapsed into simpler ones, with vowels interspersed between the consonants.

Orthographical syllables have one written vowel in them that is enunciated by speakers.
Non-orthographical syllables are syllables with no written vowel but are enunciated with a vowel.
Spoken syllables can either be orthographical or non-orthographical.
Non-syllables do not have a spoken vowel, and can either have a written vowel or not.

This all sounds very convoluted but I will explain more with an example.

Consider the word “l’áeva” (ṃi̍f̲s /l̥çɛ́vɐ̂/). This word has 2 syllables, “l’áe” (ṃi̍ /l̥çɛ́/) and “va” (f̲s /vɐ̂/). Both syllables are written and enunciated, therefore both syllables are “orthographical” and “spoken” syllables. Notice the áccent on “l’áe” (ṃi̍), this means that this syllable is stressed. Meanwhile, “va” (f̲s) does not have an accent and is unstressed.

Now consider the word “m’dz’ī́n’d’” (g̣t̲b̤̍x̣v̥ /mɪ̀d͡ʑĩ̌ɟ/). Like “l’áeva”, it has 2 spoken syllables: “m’” (g̣ /mɪ̀/) and “dz’ī́n’d’” (t̲b̤̍x̣v̥ /d͡ʑĩ̌ɟ/). Since “dz’ī́n’d’” (t̲b̤̍x̣v̥) has an accent, it is a stressed syllable. On the other hand, “m’” (g̣) does not have an accent and is unstressed. Unlike “va” (f̲s) in “l’áeva” (ṃi̍f̲s /l̥çɛ́vɐ̂/), “m’” (g̣ /mɪ̀/) does not have a written vowel but is pronounced with a vowel. Thus, “m’” is a non-orthographical unstressed syllable while “va” is an orthographical unstressed syllable.

Here is a table of what counts as what.

Syllable Type	Example (C’ī́nhṻŕaet̄a)	Romanised	IPA
Stressed Syllable	cl̍ṃ	séol’	/sēo̯l̥çɪ̂/
Orthographical Unstressed Syllable	rq̳̍ṛs	hṻ́ŕa	/çýr̥ɐ̂/
Non-orthographical Unstressed Syllable	ṛq̲̍c̲x	ŕōzm	/r̥ʉ̄zʏ̂m/
Non-syllable (with written vowel)	xp̍nb	méti	/měc/

With that out of the way, you now have the necessary background to understand how Tersamian pitch accent works.

In Tersamian, stressed vowels are marked with an upper vertical mark (s̍) in C’ī́nhṻŕaet̄a and as an accent (ó) in the romanisation. These vowels can take a high (ó), middle (ō) or rising (ǒ) tone. Meanwhile, unstressed vowels can take either a low (ò) or a falling (ô) tone.

Tones	IPA	Placement	Example (C’ī́nhṻŕaet̄a)	Romanised	IPA
˦	ó	Stressed vowel with one or more subsequent orthographical spoken syllables in the remainder of the word.	vb̤̍c̣s	pī́śa	/píɕɐ̂/
˦˨	ô	Spoken unstressed vowel immediately after the stressed syllable in the same word.	hq̲̍n̲d̲	kö́dü	/kýðŷ/
˨	ò	Unstressed vowels not immediately after a stressed vowel in the same word.	ṣp̤̍x̣pxs̲ṇ	s’ḗn’emäl	/ɕéɲêmɛ̀l/
˨˦	ǒ	Stressed vowel with no spoken syllables after it in the same word.	aq̍ṿ	ḱót’	/hɔ̌c/
˧	ō	Stressed vowel with no orthographical syllables after it in the same word but with at least one spoken unstressed syllable after it in the same word.	g̣d̳̍ṃ	m’ṻ́l’	/m̥ȳl̥çɪ̂/

An example of pitch accent in action:

- - :e
  - Oe
  - /òe̯
  - 1SG.NOM
- - ọq̍ṛe
  - ćóŕoe
  - t͡ɕɔ́r̥ôe̯
  - eat.G14-PRES.PFV
- - ad̲̍k̲g̣
  - ḱǘg’m’
  - çȳgjɪ̂m(jɪ̀)
  - meat.G24-ACC.PL
- - hsn̲sm
  - kadaj
  - kɛ̀ðɛ̀
  - all.GEN
- - s̳̍:
  - ā̈́.
  - ě/
  - PROX.G11.GEN.SG
I have eaten all the meat here.

Phonotactics

Modern Standard Tersamian has a maximum syllable structure of (X)(L)(G)V(L/G)(X), where:

X refers to all consonants minus liquids and glides.
L refers to all liquids: /l/, /l̥/, /r/, /r̥/
G refers to these glides: /j/, /ç/

Although /w/ and /ʍ/ are technically approximants, they are not considered glides. In fact, the phonosyntactical rules surrounding labiovelar approximants are quite restrictive. In the onset and coda, /w/ and /ʍ/ occupy the spots reserved for L and G, such that the onset can have a maximum of (X)(W) and for the coda, (W)(X), where W refers to the labiovelar approximants.

Orthography

v̩d̍ḷc̲in̤s n̲s̲̍ṛnxbx̣pm “D’úrzaet̄a Dä́ŕtmin’ej” /ɟúrʑɑ̂e̯tɐ̀ ðǽr̥mîɲɪ̀j/

The modern Tersamian language is written using “C’ī́nhṻŕaet̄a” (tb̤̍zrd̳̍ṛin̤s /t͡ɕín̥çŷr̥ʊ̀ɪ̯tə̀/), or the “Flying Lines System”, which is what I’m using as the “Native” writing system in this article. It is derived from the older “T’úrsoomajatta” writing system (ṿd̍ḷcqqxsmsnns /cúrʑʉ̂m̥ɐ̀ɪ̯tə̀/), or “Circular Writing System”, invented about 1000 years before present. The original T’úrsoomajatta script is a featural alphabet, with the letter designs determined by place of articulation, manner of articulation and palatalisation (but not voicing!) and is notable for its heavy use of circles and round swishes, hence the name “Circular” script. With time, cursive styles were developed for T’úrsoomajatta, and one of these styles evolved into the modern C’ī́nhṻŕaet̄a writing system. Unlike T’úrsoomajatta, C’ī́nhṻŕaet̄a is not featural, diacritics can be tacked onto letters to represent sounds completely different from the what the bare letter represents.

Although the two writing systems coexist side by side, the Tersamian writing system is NOT BICAMERAL (i.e. no uppercase and lowercase letters). Officially, the Tersamian Empire has sanctioned only C’ī́nhṻŕaet̄a for official purposes. However, the older T’úrsoomajatta script is still used for artistic and calligraphic reasons, and is notably used by some shopowners to impart a traditional feel to their establishments.

Since Modern Standard Tersamian is written using the C’ī́nhṻŕaet̄a writing system, you might want to skip to that section to understand how it works.

T’úrsoomajatta

ṿd̍ḷcqqxsmsnns “T’úrsoomajatta” /cúrʑʉ̂m̥ɐ̀ɪ̯tə̀/

The original “T’úrsoomajatta” (ṿd̍ḷcqqxsmsnns /cúrʑʉ̂m̥ɐ̀ɪ̯tə̀/) was invented 1000 years before present by merchants who were active along the Tersamian Bay. Before then, the different city states along the Tersamian coast usually adopted writing systems from other countries they were the most closely linked with, such that Middle Tersamian was written using a plethora of writing systems. Everything changed when some merchants came in contact with travellers and dignitaries who spoke the Rituppal language (documentation upcoming) and crucially had their own writing system. Inspired the flowing and circular nature of their letters, their style of writing was exported back to the Tersamian bay to create T’úrsoomajatta. Like the Rituppal writing system, T’úrsoomajatta is noted for the loops present in most letters and the lack of distinct unvoiced and voiced letters. Unlike the Rituppal abugida, T’úrsoomajatta is a featural alphabet. For example, N represents an alveolar plosive, V represents a labial plosive and W represents a palatalised labial plosive. Additional strokes are used to encode phonetic data.

Due to pragmatic concerns, T’úrsoomajatta possessed certain special characteristics, namely the featural nature of the script and the lack of separate glyphs for voiced and unvoiced consonants. These characteristics were carefully chosen for due to the phonology of the Kä́ldäŕt dialects, used by merchants on the Tersamian bay. Firstly, the Middle Tersamian language had a large number of phonemes (roughly 79 to 87). In contrast, the Rituppal language had a much smaller phoneme inventory. Instead of inventing many arbitrary shapes to represent the extra sounds, the merchants created symbols systematically, with similar-sounding phonemes having similar-looking glyphs. Secondly, the Kä́ldäŕt dialects had recently lost the distinction between unvoiced and voiced consonants from Middle Tersamian by the time T’úrsoomajatta was invented. As a result, these merchants did not create separate letters for voiced consonants.

On top of that, T’úrsoomajatta has letters for retroflex consonants. Retroflex consonants are not native to the Tersamian language. However, some loanwords from Rituppal containing retroflex consonants were borrowed over into the Kä́ldäŕt dialects via trade and contact. Instead of using the adjacent alveolar series of consonants in place of the retroflex ones, Kä́ldäŕt dialect speakers actually pronounced the retroflex consonants and they became part of the standard speech patterns, so T’úrsoomajatta has a retroflex series.

The following table details the letters of T’úrsoomajatta as used by the merchants 1000 years before present. Do note that the sounds correspond more to Middle Tersamian than to Modern Standard Tersamian. Hence, you will see tenuis and their corresponding palatal consonants. The current table is currently incomplete as I haven’t digitised all of the glyphs for T’úrsoomajatta yet.

Format: <T'úrsoomajatta> <Romanisation> /<IPA>/

Consonants	Labial		Alveolar		Retroflex		Velar		Glottal
Consonants	Tenuis	Palatal	Tenuis	Palatal	Tenuis	Palatal	Tenuis	Palatal	Tenuis	Palatal
Nasal	<m> /m/	<m'> /mʲ/	Z <n> /n/	<n'> /nʲ/	<ń> /ɳ/	<ń'> /ɳʲ/	<ŋ> /ŋ/	<ŋ'> /ɲ/
Plosive	V <p> /p/	W <p'> /pʲ/	N <t> /t/	<t'> /tʲ/	<t́> /ʈ/	<t́'> /ʈʲ/	H <k> /k/	K <k'> /kʲ/	<'> /ʔ/
Fricative	<f> /f/	<f'> /fʲ/	C <s> /s/	<s'> /sʲ/	<ś> /ʂ/	<ś'> /ɕ/	<ḱ> /x/	<ḱ'> /xʲ/	<h> /h/	<h'> /ç/
Affricate			O <c> /t͡s/	<c'> /t͡sʲ/	<ć> /ʈ͡ʂ/	<ć'> /t͡ɕ/
Approximant / Trill			<ŕ> /r/	<j> /j/			<w> /w/	<w'/y> /ɥ/
Lateral			<l> /l/	<l'> /lʲ/
Tap			<r> /ɾ/	<r'> /ɾʲ/

Monophthongs	Front	Back
High	<i> /i/	<u> /u/
Mid	<e> /e/	<o> /o/
Low		<a> /a/

Diphthongs	Front	Back
Mid	<eo> /eu̯/	<oe> /oi̯/
Low	<ae> /ɑi̯/¹	<au> /ɑu̯/

/ɑi̯/ is actually a back diphthong.

Long vowels and geminate consonants can be created by duplicating the appropriate letters:

<aa> /ɑː/
<tt> /tː/

C’ī́nhṻŕaet̄a

tb̤̍zrd̳ṛin̤s “C’ī́nhṻŕaet̄a” /t͡ɕín̥çŷr̥ʊ̀ɪ̯tɐ̀/

C’ī́nhṻŕaet̄a is an alphabet that arose from one of the many cursive styles of T’úrsoomajatta. It has 26 (what a coincidence!) basic letter forms and 8 diacritics to indicate voicing, length, alternative pronuniciations and stress. However, combining basic letter forms and diacritics do not form new letters, unlike many European languages and Arabic. For example, n <t> /t/ can be converted to ṇ <l> /l/, but ṇ is not counted as a separate letter. Instead it is just treated as n with a dot underneath. These diacritics are necessary as many of the letter forms from T’úrsoomajatta had merged in C’ī́nhṻŕaet̄a. As a result, C’ī́nhṻŕaet̄a has fewer letters than T’úrsoomajatta.

Ignoring the diacritics, the letters below are organised in the canonical C’ī́nhṻŕaet̄a alphabetical order. Each C’ī́nhṻŕaet̄a letter can represent the phonemes under the two IPA columns, but are not limited to those phonemes either. Bolded IPA sequences represent the main pronunciations indicated by the letter.

C’ī́nhṻŕaet̄a	T’úrsoomajatta	Romanisation	Middle Tersamian IPA	Modern Standard Tersamian IPA
n	N	t	/t/	/t/ /θ/
n̲	N	d	/d/	/ð/
ṇ		l	/l/	/l/ /l̥/
c	C	s	/s/	/s/ /ɕ/
c̲	C	z	/z/	/z/ /ʑ/
c̣		ś	/ʂ/	/ɕ/
c̩		ź	/ʐ/	/ʑ/
o	O	c	/t͡s/	/t͡s/
o̲	O	dz	/d͡z/	/d͡z/
ọ		ć	/ʈ͡ʂ/	/t͡ɕ/
o̩		dź	/ɖ͡ʐ/	/d͡ʑ/
z	Z	n	/n/	/n/ /n̥/ /∅/
ẓ		ŋ’	/ŋʲ/	/ɲ/ /ŋj/ /ŋ/ /∅/
r		h	/h/	/ɦ/ /h/ /ɣ/ /∅/
ṛ		ŕ	/r/	/r/ /r̥/
v	V	p	/p/	/p/ /f/
v̲	V	b	/b/	/b/ /v/
ṿ		t’	/tʲ/	/c/ /t/
v̩		d’	/dʲ/	/ɟ/ /d/ /d͡ʑ/
s		a	/ɑ/	/a/ /ɑ/ /æ/ /ɐ/ /ɛ/ /ɔ/ /ɵ/ /ə/
s̲		ä	/æ/	/æ/ /ɛ/ /a/ /ɑ/ /ɐ/ /o/ /ɵ/ /ə/
ṣ		s’	/sʲ/	/ɕ/
s̩		z’	/zʲ/	/ʑ/
t		c’	/t͡sʲ/	/t͡ɕ/
t̲		dz’	/d͡zʲ/	/d͡ʑ/
ṭ		ć’	/t͡ɕ/	/t͡ɕ/
t̩		dź’	/d͡ʑ/	/d͡ʑ/
x		m	/m/	/m/ /m̥/
x̣		n’	/nʲ/	/ɲ/
m		j	/j/	/j/ /ç/
ṃ		l’	/lʲ/	/l/ /l̥/ /lj/ /l̥ç/
f		f	/f/	/f/
f̲		v	/v/	/v/
f̣		f’	/fʲ/	/f/ /fç/
f̩		v’	/vʲ/	/v/ /vj/
w	W	p’	/pʲ/	/p/ /pç/
w̲	W	b’	/bʲ/	/b/ /bj/
ẉ		ḱ’	/ç/ (/xʲ/)	/ç/
w̩		ǵ’	/j/ /ɣʲ/	/j/ /ç/
g		ŋ	/ŋ/	/ŋ/
g̣		m’	/mʲ/	/m/ /m̥/ /mj/ /mç/
q		o	/o/	/ɔ/ /ɵ/ /ə/ /e/ /u/ /ʉ/ /ʊ/ /y/ /ʏ/ /a/ /ɑ/ /ɐ/ /æ/ /ɛ/
q̲		ö	/ø/	/e/ /ʏ/ /ɵ/ /ɔ/ /ə/ /y/ /u/ /ʉ/ /ʊ/ /æ/ /ɛ/ /a/ /ɑ/ /ɐ/
q̣		t́	/ʈ/ /t/	/t͡ɕ/ /t/
q̩		d́	/ɖ/ /d/	/d͡ʑ/ /d/ /r/
b		i	/i/	/i/ /ɪ/ /ʌ/ /ɯ/ /ɨ/ /ɛ/ /o/ /ɵ/ /ə/
b̲		ï	/ɯ/	/ʌ/ /ɯ/ /ɨ/ /i/ /ɪ/ /o/ /ɵ/ /ə/ /ɛ/
ḅ		ń	/ɳ/ /n/	/n/ /ɲd͡ʑ/ /r/
e		oe	/oi̯/	/oe̯/ /ʊɪ̯/ /øː/ /ɵ/ /ɪʏ̯/ /ɑe̯/ /ɐɪ̯/ /æi̯/ /ɛɪ̯/
e̲		öe	/øy̑/	/øː/ /ɵ/ /oe̯/ /ʊɪ̯/ /ɪʏ̯/ /æi̯/ /ɛɪ̯/ /ɑe̯/ /ɐɪ̯/
ẹ		t́’	/ʈʲ/ /tʲ/	/t͡ɕ/ /c/
e̩		d́’	/ɖʲ/ /dʲ/	/d͡ʑ/ /ɟ/
p		e	/e/	/ɛ/ /ɪ/ /o/ /ɵ/ /ə/ /i/ /ʌ/ /ɯ/ /ɨ/ /æ/ /a/ /ɑ/ /ɐ/
p̲		ë	/ɤ/	/o/ /ɵ/ /ə/ /ɛ/ /ɪ/ /ʌ/ /ɯ/ /ɨ/ /i/ /a/ /ɑ/ /ɐ/ /æ/
p̣		ś’	/ɕ/	/ɕ/
p̩		ź’	/ʑ/	/ʑ/
k		k’	/kʲ/	/k/ /c/ /kç/
k̲		g’	/gʲ/	/g/ /ɟ/ /gj/ /d͡ʑ/
ḳ		ń’	/ɳʲ/ /nʲ/	/ɲ/ /rj/
l		eo	/eu̯/	/eo̯/ /ɪ̯ə/ /ɪ̯ʊ/ /oː/ /ə/ /ʊ/ /æɯ̯/ /ɛɨ̯/ /ɑu̯/ /ɐʊ̯/
l̲		ëo	/ɤu̯/	/oː/ /ə/ /ʊ/ /eo̯/ /ɪ̯ə/ /ɪ̯ʊ/ /ɑu̯/ /ɐʊ̯/ /æɯ̯/ /ɛɨ̯/
ḷ		r	/ɾ/	/r/ /r̥/
d		u	/u/	/u/ /ʉ/ /ʊ/ /y/ /ʏ/ /ɔ/ /ɵ/ /ə/ /e/
d̲		ü	/y/	/y/ /ʏ/ /u/ /ʉ/ /ʊ/ /e/ /ɔ/ /ɵ/ /ə/
ḍ		r’	/ɾʲ/	/rj/ /r̥ç/
h		k	/k/	/k/
h̲		g	/g/	/g/
a		ḱ	/x/	/k/ /h/
a̲		ǵ	/ɣ/	/g/ /ɦ/
j		h’	/ç/	/ç/ /j/
j̣		w	/w/	/w/ /ʍ/
y		’	/ʔ/	/∅/
ỵ		w’/y	/ɥ/	/j/ /ç/ /w/ /ʍ/
i		ae	/ɑi̯/	/ɑe̯/ /ɐɪ̯/ /æi̯/ /ɛɪ̯/ /oe̯/ /ʊɪ̯/ /øː/ /ɵ/ /ɪʏ̯/
i̲		äe	/æi̯/	/æi̯/ /ɛɪ̯/ /ɑe̯/ /ɐɪ̯/ /eo̯/ /ɪ̯ə/ /ɪ̯ʊ/ /oː/ /ə/ /ʊ/
u		au	/ɑu̯/	/ɑu̯/ /ɐʊ̯/ /æɯ̯/ /ɛɨ̯/ /oe̯/ /ʊɪ̯/ /øː/ /ɵ/ /ɪʏ̯/
u̲		äu	/æu̯/	/æɯ̯/ /ɛɨ̯/ /ɑu̯/ /ɐʊ̯/ /eo̯/ /ɪ̯ə/ /ɪ̯ʊ/ /oː/ /ə/ /ʊ/

On top of the 26 basic letter forms, there are 8 diacritics. Using s as an example, these are:

Example (C’ī́nhṻŕaet̄a)	Romanisation	Description
s̲	ä	Adds voicing to a consonant or “umlaut” to a vowel.
ṣ	s’	Alternative pronunciation unrelated to the original.
s̩	z’	Adds voicing to the alternative pronunciation.
s̤	ā	“Doubles” the original consonant or vowel.
s̳	ā̈	“Doubles” the voiced consonant/ablaut vowel.
s̥	s̄’	“Doubles” the alternative pronunciation.
s͈	z̄’	“Doubles” the voiced alternative pronunciation.
s̍	á	Marks stress on a vowel.

The stress diacritic can be combined with a vowel character (or character+diacritic combo) to form a new sound. For example, s͈ (ā̈) and s̍ (á) can be combined to form s͈̍ (ā̈́).

Romanisation

The romanisation system I have employed for Tersamian maps each letter in C’ī́nhṻŕaet̄a to a letter/diacritic sequence in the Latin alphabet, and as a result, is more like a transliteration system. A mapping from basic C’ī́nhṻŕaet̄a to the romanisation system has been provided in the C’ī́nhṻŕaet̄a section. However, it does not explain the rules to transliterate geminate consonants/long vowels, palatalised consonants and diacritic order, which this section will do.

In C’ī́nhṻŕaet̄a, geminate consonants and long vowels are treated as “doubled” consonants and vowels respectively. “Doubled” letters can be written in one of two ways, whichever one is used depends on the etymology of the word in which they appear. The first (called the macron form) is to use any of the doubling diacritics like s̤ s̳ s̥ or s͈. In this case, the letters must transliterated using a combining macron U+304 (ā ā̈ s̄’ or z̄’). This way of doubling letters is seen in the spelling of C’ī́nhṻŕaet̄a, originally tb̤̍zrd̳ṛin̤s, where b̤̍ d̳ and n̤ are transliterated as ī́, ṻ and t̄ respectively. Another way doubled letters are written is by duplicating the letter (called duplicate form), so instead of s̤ s̳ s̥ s͈, the letters are written as ss s̲s̲ ṣṣ s̩s̩. When the original C’ī́nhṻŕaet̄a is written like this, the romanisation must reflect this by duplicating its letters as well (aa ää s’s’ z’z’). This is how the letters qq and nn in ṿd̍ḷcqqxsmsnns (T’úrsoomajatta) are transliterated.

Tersamian used to have a series of consonants with secondary palatal articulation. Such consonants are transliterated with an apostrophe (‘) after the main letter (e.g. k’ for k /kç/). The apostrophe is also used to transliterate the Middle Tersamian glottal stop (y) which may cause some ambiguity. In practice, this ambiguity rarely arises since y rarely appears immediately after another consonant. If a geminate palatal consonant must be written using the duplicate form, the apostrophe must appear for both consonants (e.g. t’t’ for ṿṿ).

Given the large number of phonemes in Tersamian, there are a lot of letters to be transliterated from C’ī́nhṻŕaet̄a, which by extension requires using multiple diacritics on some characters. For example, b̳̍ is transliterated as “ī̈́” U+69 U+304 U+308 U+301 (/ɯ́/). Although a human reader would not fuss over the sequence of the 3 diacritics on the letter “i”, a computer certainly would. Hence, there is a prescribed order for diacritics in the romanisation system, namely: macron (“ī” U+304), umlaut (“ī̈”U+308), acute (“ī̈́” U+301). An apostrophe is then written after the character if need be (such as for palatalised consonants). Such rules are necessary so that a computer can automatically and unabiguously transliterate C’ī́nhṻŕaet̄a writing to the romanised form and back.

The following code snippet contains the Phomo rules to convert romanised Tersamian to C’ī́nhṻŕaet̄a code points.

ae/ì
äe/i̲
áe/i̍
ä́e/i̲̍
au/ù
äu/u̲
áu/u̍
ä́u/u̲̍
eo/l̀
ëo/l̲
éo/l̍
ë́o/l̲̍
oe/è
öe/e̲
óe/e̍
ö́e/e̲̍
dz/o̲//_'
dź/o̩//_'
dz̄/o̳//_'
dz̄́/o͈//_'
dz'/t̲
dź'/t̩
dz̄'/t̳
dz̄́'/t͈
{q,ā,ä,ā̈,á,ā́,ä́,ā̈́,x}/{q,s̤,s̲,s̳,s̍,s̤̍,s̲̍,s̳̍,x}
b̄/v̳//_'
b'/w̲
b̄'/w̳
{q,ć,c̄,c̄́,x}/{q,ọ,o̤,o̥,x}//_'
c'/t̀
ć'/ṭ
c̄'/t̤
c̄́'/t̥
d̄/n̳//_'
d́/q̩//_'
d̄́/q͈//_'
d'/v̩
d́'/e̩
d̄'/v͈
d̄́'/e͈
{q,ē,ë,ē̈,é,ḗ,ë́,ē̈́,x}/{q,p̤,p̲,p̳,p̍,p̤̍,p̲̍,p̳̍,x}
f̄/f̤//_'
f'/f̣
f̄'/f̥
{q,ǵ,ḡ,ḡ́,x}/{q,a̲,h̳,a̳,x}//_'
g'/k̲
ǵ'/w̩
ḡ'/k̳
ḡ́'/w͈
h̄/r̤//_'
h'/j̀
h̄'/j̤
{q,ī,ï,ī̈,í,ī́,ḯ,ī̈́,x}/{q,b̤,b̲,b̳,b̍,b̤̍,b̲̍,b̳̍,x}
j̄/m̤
{q,ḱ,k̄,k̄́,x}/{q,à,h̤,a̤,x}//_'
k'/k̀
ḱ'/ẉ
k̄'/k̤
k̄́'/w̥
l̄/n̥//_'
l'/ṃ
l̄'/m̥
m̄/x̤//_'
m'/g̣
m̄'/g̥
{q,ń,n̄,n̄́,x}/{q,ḅ,z̤,b̥,x}//_'
n'/x̣
ń'/ḳ
n̄'/x̥
n̄́'/k̥
ŋ̄/g̤//_'
ŋ'/ẓ
ŋ̄'/z̥
{q,ō,ö,ō̈,ó,ṓ,ö́,ō̈́,x}/{q,q̤,q̲,q̳,q̍,q̤̍,q̲̍,q̳̍,x}
p̄/v̤//_'
p'/ẁ
p̄'/w̤
{q,ŕ,r̄,r̄́,x}/{q,ṛ,l̤,r̥,x}//_'
r'/ḍ
r̄'/d̥
{q,ś,s̄,s̄́,x}/{q,c̣,c̤,c̥,x}//_'
s'/ṣ
ś'/p̣
s̄'/s̥
s̄́'/p̥
{q,t́,t̄,t̄́,x}/{q,q̣,n̤,q̥,x}//_'
t'/ṿ
t́'/ẹ
t̄'/v̥
t̄́'/e̥
{q,ū,ü,ṻ,ú,ū́,ǘ,ṻ́,x}/{q,d̤,d̲,d̳,d̍,d̤̍,d̲̍,d̳̍,x}
v̄/f̳//_'
v'/f̩
v̄'/f͈
w̄/j̥//_'
w'/ỵ
w̄'/y̥
y/ỵ
ȳ/y̥
{q,ź,z̄,z̄́,x}/{q,c̩,c̳,c͈,x}//_'
z'/s̩
ź'/p̩
z̄'/s͈
z̄́'/p͈
'/ỳ
{q,a,b,c,d,e,f,g,h,i,j,k,l,m,n,ŋ,o,p,r,s,t,u,v,w,y,z,x}/{q,s,v̲,o,n̲,p,f,h̲,r,b,m,h,ṇ,x,z,g,q,v,ḷ,c,n,d,f̲,j̣,ỵ,c̲,x}
{q,à,b̀,c̀,d̀,è,f̀,g̀,h̀,ì,j̀,k̀,l̀,m̀,ǹ,ŋ̀,ò,p̀,r̀,s̀,t̀,ù,v̀,ẁ,ỳ,z̀,x}/{q,a,b,c,d,e,f,g,h,i,j,k,l,m,n,ŋ,o,p,r,s,t,u,v,w,y,z,x}

Spelling Rules

_{Pain and suffering}

_{In this section, diphthongs are treated like long vowels. You may assume diphthongs are included when I say long vowels, unless otherwise specified.}

Tersamian’s orthography was last standardised about 800 years ago when T’úrsoomajatta and C’ī́nhṻŕaet̄a were first introduced to the Ältojowan dialects through contact with Kä́ldäŕt literati. Since then, local vernacular speech has evolved to such a degree that the original Middle Tersamian spelling bears little resemblance to the sounds emitted by modern-day speakers. Despite the changes, there has been no concerted national effort to reform Tersamian spelling, particularly due to the many outspoken movements from the culturally and linguistically diverse regions of the Tersamian Empire protesting any such changes as reinforcing the Ŕáncöem dialect as the sole standard for the Modern Tersamian language. Instead, they prefer preserving the venerable Middle Tersamian spelling as a source of linguistic unity, which represents a stage when the Tersamian dialects hadn’t diverged as much yet.

As a result, the spelling of Modern Standard Tersamian differs greatly from its pronunciation. However, most of the significant sound changes between Middle Tersamian and MST were applied somewhat uniformly, so there is some method to the madness.

In this section, a letter may be phonetically transcribed for Middle Tersamian instead of Modern Standard Tersamian. If this is the case, then this transcription reflects the pronunciation of the letter in an earlier stage of the Middle Tersamian to MST transition. One critical difference between Middle Tersamian and MST is the presence of long vowels. In Middle Tersamian, long vowels were written with by adding a doubling the vowel letter or by adding/modifying the diacritic underneath the letter, such that q (o) /o/ becomes qq / q̤ (oo/ō) /oː/. Modern Standard Tersamian does not distinguish any long vowels except for /øː/ or /oː/, but long vowels are still written out as if they still exist. Furthermore, Middle Tersamian had a series of palatalised consonants that have since merged with the tenuis consonants. These consonants are typically romanised with an apostrophe ‘.

Labiopalatal Approximant Merger

The first change to the Ŕáncöem dialect is that the labiopalatal approximant w’ /ɥ/ merges with j /j/ or w /w/ depending on the environment. If w’ comes before a front vowel, such as e /e/ or eo /eu̯/, then w’ gets pronounced like w. If w’ comes before a back vowel such as a /ɑ/, then w’ gets pronounced like j. If w’ is in the coda, then the pronunciation it takes (w or j) is decided based on the vowel before it, so if the previous vowel is e, then w’ is pronounced like w. If none of the above environments are found, then w’ is pronounced as j.

- - ỵp̤
  - w’ē (pronounced like wē)
  - /ɥeː/ → /weː/ → /ʍɪ̀/
  - five
Five (5)

- - mqme̲ỵsh
  - jojöew’ak (like: jojöejak)
  - /jojøy̑ɥɑk/ → /jojøy̑jɑk/ → /jə̀çɪ̀ʏ̯çɛ̀(k)/
  - seventeen
Seventeen (15 in base 12)

Great Vowel Shift

_{Yes. I stole the name from English’s own Great Vowel Shift. Deal with it.}

The next major change is the raising of long vowels (including diphthongs). This applies to all long vowels unless the long vowel is flanked by unvoiced consonants or is adjacent to a h.

- - ṛs̤̍mp̳h̲
  - ŕā́jē̈g (like: ŕṓjī̈g)
  - /ˈrɑːjɤːg/ → /ˈroːjɯːg/ → /r̥ɵ́çɯ̂g/
  - monarch-supreme
Supreme King/Queen/Monarch

In this example, both ā́ /ɑː/ and ē̈ /ɤː/ are long vowels, and so their vowel qualities are raised to /oː/ and /ɯː/ respectively.

- - ṃi̍f̲s
  - l’áeva (like: l’óeva)
  - /ˈlʲɑi̯vɑ/ → /ˈlʲoi̯vɑ/ → /l̥çévɐ̂/
  - have
Have

Like long monophthongs, diphthongs are also raised in eligible positions.

- - cq̤̍
  - sṓ (like: sū́)
  - /ˈsoː/ → /ˈsuː/ → /sʉ̌/
  - inside
Inside*

In this example, only the onset s has an unvoiced consonant, so the sound change still applies to ṓ.

Of course, this sound change does not apply if the long vowel is flanked by unvoiced consonants. For the onset, the unvoiced consonant may be succeeded by the following sonorants: r (ḷ), r’ (ḍ), ŕ (ṛ), j (m), l (ṇ), l’ (ṃ), w (j̣), w’ (ỵ). The consonant after the long vowel must be unvoiced for the exception to apply. This consonant need not be part of the coda of the same syllable containing the long vowel for the exception to apply, instead it could be part of the onset of the next syllable.

- - hṛp̤̍n
  - kŕḗt
  - /ˈkreːt/ → /r̥ɛ̌θ/
  - type.of.fish
A type of fish.

For kŕḗt, the onset kŕ has the unvoiced consonant k followed by the sonorant ŕ, while the coda has the unvoiced consonant t, so the vowel ē is still pronounced as /eː/.

The Great Vowel Shift also gets overridden when the long vowel appears next to h even if the the long vowel is in contact with a voiced consonant. In this case, the vowel retains its original quality.

- - rs̳̍ḷs
  - hā̈́ra
  - /ˈhæːɾɑ/ → /hǽr̥ɐ̂/
  - push
Push

Lexical Vowel Harmony

Almost immediately after the Great Vowel Shift occurred, vowels within each word underwent vowel harmony according to the rules in the vowel harmony section. Previously, the vowels within each word may not align each other in terms of vowel harmony due to sound changes between Old Tersamian and Middle Tersamian, as well as borrowing of words from other languages that did not have the same vowel harmony system as Middle Tersamian, or simply did not have any vowel harmony at all. As a result, words like kédo /ˈkedo/ (which requires backness harmony due to it being a Tier 4 word), have vowels that are in conflict with each other according to vowel harmony. This step forces all words already in Middle Tersamian to undergo internal vowel harmony, it is not applied retroactively for words borrowed after that period. Such ambiguity usually serves as a source of confusion for learners.

All words in Tersamian are given a Word Class and this is typically written in the format Gxy where x is either 1 or 2 denoting its philosophical group and y is a number ranging from 1 to 5 denoting its exaltation tier (e.g. G23 means second philosophical group, third exaltation tier). More information on how word classes work can be found here. For this section, all you need to know that words have a certain word class and their exaltation tier determines which vowel harmony paradigm they undergo according to the following table.

Exaltation Tier	Vowel Harmony
Exalted (Gx1)	Progressive Height Harmony
Emergent (Gx2)	Progressive Backness Harmony
Anthropic (Gx3)	None
Fauna (Gx4)	Regressive Backness Harmony
Flora (Gx5)	Regressive Height Harmony

For progressive and regressive height harmony during lexical vowel harmony, the trigger vowel causes the target vowel to travel all the way to match the trigger vowel’s height. Thus, a trigger vowel “u” would cause target vowel “ä” to be raised to “i” instead of “e”.

When deciding how to evolve the pronunciation of a word at its stage, we must first consider its exaltation tier. The word kédo’s word class is G24, meaning it belongs to the “Fauna (Gx4)” exaltation tier and thus undergoes regressive backness harmony. This means the last vowel in the word “o”, causes all vowels before it and up until to the root vowel “é” to change their backness to match that of “o”. Hence …

- - hp̍n̲q
  - kédo (like: kë́do)
  - /ˈkedo/ → /ˈkɤdo/ → /kóðɔ̂/
  - city.G24
City

One exception to this vowel harmony rule is the word-final infinitive “-a”. This vowel appears in most verbs in their infinitive form and for the most part it is treated as transparent vowel, meaning that it does not trigger any regressive vowel harmony, and neither does it undergo any progressive vowel harmony.

- - tq̳̍n̲s
  - c’ō̈́da (like: c’ṻ́da)
  - /ˈt͡sʲøːdɑ/ → /t͡ɕýðɐ̂/
  - know.G15
Know

In c’ō̈́da, “-a” does not trigger regressive height vowel harmony.

- - xd̍ṛs
  - múŕa
  - /ˈmurɑ/ → /múrɐ̂/
  - go.G21
Go

In múŕa, “-a” is not affected by progressive height harmony due to “ú”.

However, if there is a vowel before infinitive “-a”, then the vowel can trigger regressive vowel harmony assuming the word is a Gx4 or Gx5 verb.

Consonant Softening

In the Great Vowel Shift, it was consonants that influenced the value of long vowels, but in this sound change, long vowels cause consonants to soften unless they exist as part of a consonant cluster. During softening, voiced nasals, liquids and glides become unvoiced and the unvoiced alveolar plosive /t/ lenitions to /θ/ respectively. Softened consonants are called “soft” consonants while unsoftened consonants are termed “hard” consonants.

- - ỵd̤̍ṇ
  - w’ū́l (like: h’ū́l)
  - /ˈɥuːl/ → /ˈjuːl/ → /ˈçuːl̥/ → /çʉ̌l̥/
  - dog
Dog

If a consonant that can be softened is part of a consonant cluster, then the consonant is not lenitioned. In the example below, since l’ /lʲ/ is part of the consonant cluster l’m’ /lʲmʲ/.

- - ṣb̤̍ṃg̣
  - s’ī́l’m’
  - /ˈsʲiːlʲmʲ/ → /ɕīlmjɪ̂/
  - farm
Farm

The following table shows the mapping between unsoftened consonants and softened consonants, if it is next to a long vowel and not part of a consonant cluster.

Romanised	C’ī́nhṻŕaet̄a	Middle Tersamian IPA	Softened	MST IPA
m	x	/m/	/m̥/	/m̥/
m’	g̣	/mʲ/	/m̥ʲ/	/m̥ç/
n	z	/n/	/n̥/	/n̥/
n’	x̣	/nʲ/	/n̥ʲ/	/ɲ/
ń	ḅ	/ɳ/ /n/	/ɳ/ /n̥/	/n̥/
ń’	ḳ	/ɳʲ/ /nʲ/	/ɳʲ/ /n̥ʲ/	/ɲ/
ŋ	g	/ŋ/	/ŋ/	/ŋ/
ŋ’	ẓ	/ŋʲ/	/ŋʲ/	/ɲ/
t	n	/t/	/θ/	/θ/
ḱ	a	/x/	/h/	/h/
ǵ	a̲	/ɣ/	/ɦ/	/ɦ/
ǵ’	w̩	/j/	/ç/	/ç/
h¹	r	/h/	/h/	/h/
l	ṇ	/l/	/l̥/	/l̥/
l’	ṃ	/lʲ/	/l̥ʲ/	/l̥ç/
ŕ	ṛ	/r/	/r̥/	/r̥/
j	m	/j/	/ç/	/ç/
r	ḷ	/ɾ/	/r̥/	/r̥/
r’	ḍ	/ɾʲ/	/r̥ʲ/	/r̥ç/
w	j̣	/w/	/ʍ/	/ʍ/
w’	ỵ	/ɥ/ → /w/ /j/	/ʍ/ /ç/	/ʍ/ /ç/

h is a special letter in that it remains unvoiced (/h/) if next to a long vowel. Otherwise, it becomes unvoiced /ɦ/. More will be explained in Voiced Glottal Fricative.

Voiced Glottal Fricative

Middle Tersamian only had one glottal fricative – the unvoiced glottal fricative /h/. However, in Modern Standard Tersamian this phoneme split into 2 distinct phonemes – /h/ and /ɦ/ – depending on the environment. If /h/ appears in a context where a consonant would be softened (i.e. next to a long vowel and not part of a consonant cluster), then it stays as /h/. Otherwise, it becomes unvoiced /ɦ/.

Consonant Mergers

Another critical difference between Middle Tersamian and MST is MST’s significantly simplified consonant inventory (albeit still a large one) due to a series of mergers between several consonants.

Romanised	C'ī́nhṻŕaet̄a	Middle Tersamian IPA	MST IPA
n ń	z ḅ	/n/ /ɳ/	/n/ (Hard)
n ń	z ḅ	/n/ /ɳ/	/n̥/ (Soft)
n' ń' ŋ'	x̣ ḳ ẓ	/nʲ/ /ɳʲ/ /ŋʲ/	/ɲ/ (Hard and Soft)
s' ś ś'	ṣ c̣ p̣	/sʲ/ /ʂ/ /ɕ/	/ɕ/
z' ź ź'	s̩ c̩ p̩	/zʲ/ /ʐ/ /ʑ/	/ʑ/
k ḱ (Hard)	h a	/k/ /x/	/k/
g ǵ (Hard)	h̲ a̲	/g/ /ɣ/	/g/
h (Soft) ḱ (Soft)	r a	/h/ /x/	/h/
h (Hard) ǵ (Soft)	r a̲	/h/ /ɣ/	/ɦ/
c' ć ć' t́ t́'	t ọ ṭ q̣ ẹ	/t͡sʲ/ /ʈ͡ʂ/ /t͡ɕ/ /ʈ/ /ʈʲ/	/t͡ɕ/
dz' dź dź' d́ d́'	t̲ o̩ t̩ q̩ e̩	/d͡zʲ/ /ɖ͡ʐ/ /d͡ʑ/ /ɖ/ /ɖʲ/	/d͡ʑ/
h' j (Soft) ḱ' ǵ' (Soft) w' (Soft, /j/ environment)	j m ẉ w̩ ỵ	/ç/ /j/ /ç/ /j/ /ɥ/	/ç/
j (Hard) ǵ' (Hard) w' (Hard, /j/ environment)	m w̩ ỵ	/j/ /j/ /ɥ/	/j/
w (Hard) w' (Hard, /w/ environment)	j̣ ỵ	/w/ /ɥ/	/w/
w (Soft) w' (Soft, /w/ environment)	j̣ ỵ	/w/ /ɥ/	/ʍ/
r (Hard) ŕ (Hard)	ḷ ṛ	/ɾ/ /r/	/r/
r (Soft) ŕ (Soft)	ḷ ṛ	/ɾ/ /r/	/r̥/

Loss of Long Vowels

The next sound change (I sound like a broken recorder) is the merger between long monophthongs and short monophthongs. Long monophthongs that have been raised due to the Great Vowel Shift are also subject to the changes based on their new vowel quality. For example, if ē /eː/ were to be raised to /iː/, then its new shortened value would be /i/ instead of /ɛ/.

Romanisation	C’ī́nhṻŕaet̄a	Middle Tersamian IPA	MST IPA
ī	b̤	/iː/	/i/
ī̈	b̳	/ɯː/	/ɯ/
ē	p̤	/eː/	/ɛ/
ē̈	p̳	/ɤː/	/ɵ/
ū	d̤	/uː/	/ʉ/
ṻ	d̳	/yː/	/y/
ō	q̤	/oː/	/ɵ/
ō̈	q̤	/øː/	/e/
ā	s̤	/ɑː/	/ɑ/
ā̈	s̳	/æː/	/æ/

In addition, Middle Tersamian a /ɑ/ becomes MST /a/ in most contexts unless it is adjacent to h, when it becomes MST /ɑ/ instead. This is due to a change preceding the step in consonant mergers.

Reduced Vowels

In unstressed positions, vowels turn into their reduced vowels according to the vowels section in the phonology. The vowel immediately after the stressed vowel in the same word does not get reduced, however.

- - ṣs̲̍x̣s̲xs̲ṇ
  - s’ä́n’ämäl
  - /ˈsʲænʲæmæl/ → /ˈɕæɲæmæl/ → /ˈɕæɲæmɛl/ → /ɕɛ́ɲɛ̂mɛ̀l/
  - heaven
Heaven, paradise

- - mlḷ
  - jeor
  - /jeu̯ɾ/ → /çeo̯r̥/ → /çɪ̯ər̥/ → /çɪ̯ə̀r̥/
  - seven
Seven (7)

Transition to Pitch Accent

Tersamian pitch accent was developed in 2 stages. The first stage involves converting the stress system into a tone system by giving stressed syllables a high tone (˦) than surrounding unstressed syllables (˨). Then, if there is unstressed syllable right after a stressed syllable in the same word, the unstressed syllable gets a falling tone (˦˨).

- - ṣs̲̍x̣s̲xs̲ṇ
  - s’ä́n’ämäl
  - /ˈsʲænʲæmæl/ → /ˈɕæɲæmɛl/ → /ɕǽɲæ̂mɛ̀l/ → /ɕɛ́ɲɛ̂mɛ̀l/
  - heaven
Heaven, paradise

Palatalised Vowels

If the consonant preceding them is palatalised, then certain vowels become palatalised. If the vowel is a front vowel, then it undergoes raising while back vowels undergo fronting.

- - md̍h̲s
  - júga (like: jū́ga but with hard “j”)
  - /ˈjugɑ/ → /júgɐ̂/ → /jʉ́gɐ̂/
  - sit
Sit

Unpalatalised IPA	Palatalised IPA
/a/	/æ/
/æ/	/ɛ/
/ɛ/	/e/ (unreduced), /ɪ/ (reduced)
/ɔ/	/ɵ/
/o/	/ɵ/
/u/	/ʉ/ (unreduced), /ʏ/ (reduced)

Simplified Syllable Structure

Middle Tersamian had a complex syllable structure of up to CCCVCCC. However, the syllable structure of Modern Standard Tersamian is heavily simplified, allowing up to one obstruent or nasal in the onset or coda. In some cases, the consonant clusters are merged, the syllable conforms to the allowed structure and all is fine, but this is by far the exception to the norm. In most cases, the consonant clusters are broken up using epenthetic vowels. The value of the epenthetic vowels are inherited from the nearest vowel before it (or after it if desperate).

- - ṛq̤̍c̲x
  - ŕṓzm (like “ŕū́zm” due to long vowel raising)
  - /ˈroːzm/ → /r̥ʉ́zm/ → /r̥ʉ̄zʏ̂m/
  - eye
Eye

When an epenthetic vowel is added, it becomes part of a non-orthographical unstressed syllable, with implications for pitch accent. This additional change in pitch accent is part of the second stage of Tersamian pitch accent development.

Depalatalisation

Then comes the loss of palatalised consonants. Most of these consonants (like /mʲ/) simply decompose into a tenuis consonant (/m/) + a palatal glide (/j/).

- - ṃq̍xs
  - l’óma
  - /ˈlʲomɑ/ → /lʲɵ́mɐ̂/ → /ljɵ́mɐ̂/
  - say
Say; Talk

If the consonant was unvoiced, then the palatal glide must also be unvoiced (/ç/).

- - g̣b̳̍c
  - m’ī̈́s
  - /ˈmʲɯːs/ → /m̥ʲɯ̌s/ → /m̥çɯ̌s/ → /m̥çɯ̌/
  - say
Say; Talk

Some consonants like /ɲ/, /ɕ/, /ʑ/, /c/, /ɟ/, /t͡ɕ/ and /d͡ʑ/ remain as they are.

- - v̩d̍ḷc̲s
  - d’úrza
  - /ˈdʲuɾzɑ/ → /ɟʉ́rʑɐ̂/
  - write
Write

Nasalisation

The final step in this convoluted step is nasalisation. If a vowel comes before a voiced nasal consonant in the coda, it becomes nasalised. If the consonant is /n/, then it gets deleted.

- - mp̍x
  - jém
  - /ˈjem/ → /jěm/ → /jɪ̃̌m/
  - inside
Inside

- - ṇp̍mz
  - léjn
  - /ˈlejun/ → /lěo̯n/ → /lẽ̌õ̯/
  - person.G13-ACC
Person

Onset nasals do not trigger nasalisation.

- - g̣p̤̍t
  - m’ḗc’
  - /ˈmʲeːt͡sʲ/ → /m̥çǐt͡ɕ/
  - white
White

Inherent Vowels

As part of MST grammar, inflected words can have inherent vowels. This vowels are not written but are assumed to be there by speakers. These vowels are an integral part of prefixes and suffixes and must be learned along with the affix. For example, the G1x accusative suffix is “-[u]n” where “[u]” is the inherent vowel. When inflecting a word, “n” gets written but the “[u]” does not. However, the “[u]” is still pronounced and is treated as part of a non-orthographical unstressed syllable. The inherent vowel also undergoes vowel harmony or triggers vowel harmony, depending on the word class of the stem it is being attached to.

If the suffix is inflectional, then during progressive or regressive vowel harmony, the target vowel only travels by one category towards the height of the trigger vowel. Hence, trigger “a” would cause target “u” to move up to “o” instead of “a”.

- - rs̤̍ṇz
  - hā́ln (like “hṓl[o]n” due to long vowel raising and progressive height harmony)
  - /ˈhɑːlon/ → /hɵ̄l̥õ̂/
  - god.G11
God

- - j̣p̍cz
  - wésn (like “wë́s[u]n” due to regressive backness harmony)
  - /ˈwɤsun/ → /wōsõ̂/
  - pole.G14
Cylinder; pole

In some cases, the inherent vowel may be omitted in some instances, such as the stem ending in vowel, rendering the inherent vowel unused. If so, the inherent vowel does not trigger any sort of regressive vowel harmony.

- - s̲̍z
  - ā́n (like “ṓn” due to long vowel raising)
  - /ˈɑːn/ → /ə̃̌/
  - PROX.G11
Here; Now

Grammar

Of the major languages/dialects that emerged from Middle Tersamian, Ŕáncöem Tersamian is the most grammatically conservative, preserving 5 out of 6 original cases from Middle Tersamian and maintaining its strict insistence on the Noun-Verb-Adjective (NVA) system. In addition, Modern Standard Tersamian has preserved most of the 10-gender word class system. Tersamian is a predominantly head-initial language with a SVO basic word order. The only exception to this are determiner phrases which are head-final. Topic-prominence features heavily in the language, which means all 6 possible basic word orders can appear in any text. Like most languages in the Skomi family, Tersamian has nominative-accusative alignment.

In Middle and Modern Standard Tersamian, all words except a few particles are either nouns, verbs or adjectives. There are no adpositions nor adverbs, and relative clauses are firmly attached to the modified noun or verb like adjectives. This system is called the NVA system and is the basis of Tersamian grammar.

Word Class

Middle Tersamian used to have 10 word classes. Unlike most nat- and conlangs, these word classes apply to nouns as well as verbs and adjectives. The word classes can be split into 2 philosophical groups, G1x and G2x. The G1x philosophical group are generally for words describing sentient/natural things, while G2x is for their “products”. Thus, G1x is usually called the “Necessary” group and G2x the “Sufficient” group. Additionally, the word classes can be divided into 5 exaltation tiers, from the highest Gx1 to lowest Gx5 tier.

Class Names	Necessary (G1x)	Sufficient (G2x)
Exalted (Gx1)	Sacred (G11) h-	Fundamental (G21) m-
Emergent (Gx2)	Royal (G12) ŕ-	Fundamental (G22) j-
Anthropic (Gx3)	Human (G13) l-	Man-made (G23) t-
Fauna (Gx4)	Animal (G14) v-	Organic (G24) k-
Flora (Gx5)	Plant (G15) s-	Static (G25) c-

The 2 philosophical groups primarily determine the inflectional affixes a word takes while the exaltation tiers determine the type of vowel harmony a word undergoes.

Exaltation Tier	Vowel Harmony
Exalted (Gx1)	Progressive Height Harmony¹
Emergent (Gx2)	Progressive Backness Harmony
Anthropic (Gx3)	None
Fauna (Gx4)	Regressive Backness Harmony
Flora (Gx5)	Regressive Height Harmony¹

For height harmony, inflectional suffixes typically only cause target vowels to shift up or down by one height category towards the height of the trigger vowel. See vowel harmony for more information.

Although Middle Tersamian has 10 word classes each with their own inflectional affixes and vowel harmony paradigms, there are really only 7 “functional” word classes left in Modern Standard Tersamian. This is due to the merger between several word classes, namely:

G14 and G15 merged so that G15 words behave like G14 words by undergoing regressive backness harmony.
G23, G24 and G25 merged so that such words behave like G23 words and do not undergo any vowel harmony.

For example,

- - ṣs̲̍n̲q
  - s’ä́d-o
  - /ɕǽðɔ̂/
  - walk.G15-PRES
I’m just walking.

although s’ä́do is a G15 word, it behaves like a G14 word and so the root vowel “ä́” undergoes regressive backness harmony and is pronounced like “á”. You may notice in the phonetic transcription that the “ä́” is still pronunced as /ǽ/. This is due to influence from palatalised “s’”.

This merger only applies to words coined or inflected in Modern Standard Tersamian, so a G25 (let’s say) Middle Tersamian word still undergoes regressive lexical height harmony as specified in lexical vowel harmony. However, if that word were to be inflected now, then it would undergo G23 vowel harmony (a.k.a. no vowel harmony).

TODO: Provide an example

Nouns

Nouns can decline for 5 cases — Nominative, Instrumental, Accusative, Dative, Genitive — and 2 numbers — Singular, Plural. In some of the declension paradigms below, there are vowels encased in square brackets like in -[u]k. These vowels are inherent vowels, which may be added to break up an illegal consonant cluster. Inherent vowels are not typically written, even if they are pronounced, meaning that if the inherent vowel does gets pronounced, it is treated as a non-orthographical syllable and affects the tones of neighbouring vowels according to Tersamian’s pitch accent rules.

G1x	Singular	Plural
Nominative	Base Form	-[e]j
Instrumental	-[ï]r	-[i]r'
Accusative	-[u]n	-[i]n'
Dative	-[u]k	-[i]k'
Genitive	Genitive Stem	Genitive Stem -[ï]j

The genitive stem is usually formed by adding an umlaut to the root vowel. In rare cases, a different genitive stem is used and sometimes it may be identical to the base stem.

G2x	Singular	Plural
Nominative	Base Form	Plural Stem
Instrumental	-[ï]r	Plural Stem -r'[i]
Accusative	-[o]m	Plural Stem -[ö]m'
Dative	-[o]k	Plural Stem -[ö]k'
Genitive	-[ï]l	Plural Stem -[i]l'

The plural stem is usually formed by adding an umlaut to the root vowel. In rare cases, a different plural stem is used and sometimes it may be identical to the base stem.

Noun Declension Examples

ṇp̍m léj (G13) "person"	Singular	Plural
Nominative	ṇp̍m léj /lě/	ṇp̍mm léjj /lǐ/
Instrumental	ṇp̍mḷ léjr /lējɨ̂r/	ṇp̍mḍ léjr' /lējɪ̂r/
Accusative	ṇp̍mz léjn /lẽ̌õ̯/	ṇp̍mx̣ léjn' /lǐɲ/
Dative	ṇp̍mh léjk /lěo̯k/	ṇp̍mk léjk' /lǐk/
Genitive	ṇp̲̍m lë́j /lǒe̯/	ṇp̲̍mm lë́jj /lōe̯jɪ̂/

ms̍ṇ jál (G22) "country"	Singular	Plural
Nominative	ms̍ṇ jál /jæ̌l/	ms̲̍ṇ jä́l /jɛ̌l/
Instrumental	ms̍ṇḷ jálr /jǣlɪ̂r/	ms̲̍ṇḍ jä́lr' /jɛ̄lɪ̂r/
Accusative	ms̍ṇx jálm /jǣlɪ̂m/	ms̲̍ṇg̣ jä́lm' /jɛ̄lɪ̂m/
Dative	ms̍ṇh jálk /jɛ̃̌k/	ms̲̍ṇk jä́lk' /jɛ̃̄cɪ̂/
Genitive	ms̍ṇṇ jáll /jǣlɨ̂l/	ms̲̍ṇṃ jä́ll' /jɛ̄ljɪ̂/

Verbs

Tersamian verbs inflect for mood, tense and aspect. The infinitive INF (base) form of a verb usually ends in -a, whose behaviour is between a that of a non-orthographical and an orthographical vowel. The verb stem can usually be found by removing the final -a.

G1x		Imperfective		Perfective
Imperative	-a	Non-progressive	Progressive	Non-progressive	Progressive
Realis	Present	-(öt)o	-(öt)oŕ	-(öt)oe	-(öt)oeŕ
Realis	Past	-(ës)e	-(ës)eŕ	-(ës)eo	-(ës)eoŕ
Irrealis	Present	Irrealis Stem -(öt)o	Irrealis Stem -(öt)oŕ	Irrealis Stem -(öt)oe	Irrealis Stem -(öt)oeŕ
Irrealis	Past	Irrealis Stem -(ës)e	Irrealis Stem -(ës)eŕ	Irrealis Stem -(ës)eo	Irrealis Stem -(ës)eoŕ

G2x		Imperfective		Perfective
Imperative	-a	Non-progressive	Progressive	Non-progressive	Progressive
Realis	Present	-	-(c)aŕ	-öe(cö)	-öe(cö)ŕ
Realis	Past	-'u	-'ṻ	-'uŕ	-'ṻŕ
Irrealis	Present	Irrealis Stem	Irrealis Stem -(c)aŕ	Irrealis Stem -öe(cö)	Irrealis Stem -öe(cö)ŕ
Irrealis	Past	Irrealis Stem -'u	Irrealis Stem -'ṻ	Irrealis Stem -'uŕ	Irrealis Stem -'ṻŕ

The syncretic inflection paradigm is an inbetween of the G1x and G2x inflection paradigms. Certain G1x and G2x verbs can undergo syncretic inflections, however, which verbs do so cannot be predicted and must be memorised.

Syncretic		Imperfective		Perfective
Imperative	-a	Non-progressive	Progressive	Non-progressive	Progressive
Realis	Present	-	-ar	-ō̈	-or'
Realis	Past	-in	-eo	-ed	-eor
Irrealis	Present	Irrealis Stem -	Irrealis Stem -ar	Irrealis Stem -ō̈	Irrealis Stem -or'
Irrealis	Past	Irrealis Stem -in	Irrealis Stem -eo	Irrealis Stem -ed	Irrealis Stem -eor

M-verbs may have an additional change in the verb stem (for both realis and irrealis) when inflecting for past tense. The steps to create the past stem are:

Step	Sound Changes	Example
Original		tl̍cs c’éosa
Add a nasal to the stem if the stem does not have a nasal immediately after the final vowel.	∅/N/V_((C-N)C*)?#/	tl̍zṣd c’éons’u
Remove non-nasal alveolar non-sibilant consonant after the nasal	C[+alveolar -sibilant]/∅/VN_C*#	tl̍zṣd c’éons’u

In the conjugation tables above, you will notice some letters encased in round brackets like in -(öt)o. These letters are optional additions for the conjugation and may be included to distinguish conjugations whose pronunciations are similar due to vowel harmony, especially due to progressive backness harmony characteristic of Gx2 class verbs.

- - :e
  - Oe
  - /òe̯
  - 1SG.NOM
- - md̍gh̲p:
  - júŋge.
  - jṍŋgô/
  - sit.G22-PAST
- - →
  - →
  - →
  - →
- - :e
  - Oe
  - /òe̯
  - 1SG.NOM
- - md̍gh̲p̲cp:
  - júŋgëse.
  - jṍŋgôsə̀/
  - sit.G22-PAST
I sat down. (TODO: Use a G12 verb. Why don’t I have one lmao)

Tense

Tersamian has 2 tenses, past and present (or non-past). Tense is used to mark the time reference in which the action or state is true (for realis verbs) or might be/could have been true (for irrealis verbs).

Aspect

Tersamian has 4 aspects, the Imperfective, the Progressive, the Perfective and the Perfect Progressive.

The imperfective is typically used for habitual actions, for actions whose current state of completion is unknown.

- - :ṇp
  - Le
  - /lɛ̀
  - G13.NOM.SG
- - rs̍hp
  - háko
  - ɦákâ
  - go.G11-PRES
- - os̤̍hx
  - cā́km
  - t͡sɑ̄kɔ̃̂m
  - shop.G25-ACC.SG
- - zbf̲
  - niv
  - ỹ̀
  - for
- - h̲q̲̍jg̣
  - gö́h’m’
  - gēçɪ̃̂m
  - food.G24-ACC.PL
- - ṃp̍ṛs
  - l’éŕa
  - ljérɐ̂
  - buy.G13-INF
- - q̲z:
  - ön.
  - ø̃̀ː/
  - RELX.G13-ACC
They/He/She went to the store to buy meat.

In this example, the speaker knows that the subject had definitely left for the store, but doesn’t know whether they have reached it or not, because of this, whether the person has actually bought anything is still up in the air and so the verb l’éŕa (to buy) can only be infinitive. Also note that “háko” is in the present tense; the speaker is talking about the subject’s current whereabouts, not when they had left for the store.

The progressive aspect is used to describe actions that are in progress or for ongoing actions with a definite end point.

- - :e
  - Oe
  - /òe̯
  - 1SG.NOM
- - ṃq̍xqṛ
  - l’ómoŕ
  - ljɵ́mɔ̂r
  - talk.G13-PRES.PROG
- - ms̍ṇex:
  - jáloem.
  - jǽl̥ø̂ːm̥/
  - country.G22-1SG.GEN-ACC.SG
I am talking about my country.

The example above implies that the speaker is talking about their country to give a brief introduction on it or to prove a point.

In contrast, the example below uses the non-progressive present tense on the same verb, which could imply that the speaker is rambling about their country to no apparent end, which is more in line with what a continuous aspect would describe. However, the non-progressive imperfective could also be used with a habitual sense to mean that the speaker regularly talks about their country.

- - :e
  - Oe
  - /òe̯
  - 1SG.NOM
- - ṃq̍xq
  - l’ómo
  - ljɵ́mɔ̂
  - talk.G13-PRES
- - ms̍ṇex:
  - jáloem.
  - jǽl̥ø̂ːm̥/
  - country.G22-1SG.GEN-ACC.SG
I am talking about my country. OR
I regularly talk about my country.

The perfective (PFV) aspect is used to denote a completed action. It can also denote past actions that have present relevance (thus fulfilling the function of the perfect PERF aspect).

- - :e
  - Oe
  - /òe̯
  - 1SG.NOM
- - w̲s̳̍zce
  - b’ā̈́nsoe
  - bjɛ̃́n̥sôe̯
  - write.G14-PRES.PFV
- - zpz
  - nen
  - nɛ̃̀
  - 3SG.ACC
- - xs
  - ma
  - mɐ̀
  - what
- - rd̤a:
  - hūḱ
  - hʉ̀
  - 2SG.DAT
I have written something for you.

The perfective progressive (PFPG) aspect is so called as its conjugation involves a combining the progressive suffix “-ŕ” with the perfective conjugation of a verb. In reality, it’s function is completely different from what it’s name implies. Originally, when used on a verb in Middle Tersamian, the perfective progressive describes an action that involves a series of steps that are in the process of being completed one-by-one.

In modern Tersamian, the role of the perfective progressive has shifted significantly, and depending on the modern Ältojowan dialect, the perfective progressive can stand for a varying number of aspects:

Dialect	Aspect	Gloss
Ŕáncöem	Iterative	ITER
???	Continuous	CONT
???	Durative	DUR
???	Defective	DFTV
???	Gnomic	GNO

In Ŕáncöem,

- - :nu̲̍ṇe
  - Tä́uloe
  - /θéo̯l̥ôe̯
  - car.G23-1SG.GEN-NOM.SG
- - d̤̍n̲eṛ:
  - ū́doeŕ.
  - ʉ́ðôe̯r̥/
  - stop.G11-PRES.ITER
My car keeps stopping.

Mood

On top of the default realis mood, Tersamian also has the irrealis and the imperative moods. The realis mood is used for states or actions that are known to be true (or will definitely true for future events), while the irrealis is mainly for hypothetical or conditional events. The realis and irrealis distinction arose from evidentiality prefixes back in Proto-Skomi, whose meaning eventually eroded.

The imperative form of a verb has the suffix “-a”, but although it is the same as the default infinitive ending “-a”, the imperative “-a” is pronounced with a full /ɑ/ instead of reduced /ɐ/. When the imperative is used in a sentence, an additional particle “a” (yes, another a) is appended to the front of the sentence and carries a high tone despite having no stress mark. Unlike English, the sentence retains V2 word order, with the direct object being elevated the first argument. If there is a subject, it is demoted to being the second argument, and is placed after the verb.

- - :s
  - A
  - /ɑ́
  - IMP
- - ws̲̍chx
  - p’ä́skm
  - pjǣskə̂m
  - door.G24N-ACC.SG
- - s̤̍ḷs:
  - ā́ra
  - ɵ́r̥ɑ̂/
  - open.G11-IMP
Open the door!

Adjectives

Tersamian adjectives serve a dual purpose of adjectives and adverbs; there is no distinction between the 2 concepts. Adjectives and adverbs are placed after their respective heads. In a noun phrase, adjectives must agree with the noun in terms of number, case and word class. In a verb phrase, adverbs can be used to describe the action, or to add information like modality such as wants (optative) or needs, which would require an additional noun in English (e.g. The child wants to play with this toy.)

As for nouns, the vowels in brackets signify unwritten inherent vowels that may be pronounced to break up complex consonant clusters.

Adjectives	G1x		G2x
Adjectives	Singular	Plural	Singular	Plural
Nominative	Base Form	-[e]j	Base Form	-[e]j
Instrumental	-[e]r'
Accusative	-[ü]l'	-[ü]n'	-[i]m'
Dative	-[i]k'
Genitive	-j[e]

Adjective Examples	s̍k ák' (G11) "big"		g̣ṣb̤̍v̩ m's'ī́d' (G21) "thin"
Adjective Examples	Singular	Plural	Singular	Plural
Nominative	s̍k ák' /ǎc/	s̍km ák'j /ākjɪ̂j/	g̣ṣb̤̍v̩ m's'ī́d' /mɪ̀ɕɪ̌ɟ/	g̣ṣb̤̍v̩m m's'ī́d'j /mɪ̀ɕɪ̄ɟɪ̂(j)/
Instrumental	s̍kḍ ák'r' /ākjɪ̂r/		g̣ṣb̤̍v̩ḍ m's'ī́d'r' /mɪ̀ɕɪ̄ɟɪ̂r/
Accusative	s̍kṃ ák'l' /ākjɪ̃̂l/¹	s̍kx̣ ák'n' /ākjɪ̃̂ɲ/¹	g̣ṣb̤̍v̩g̣ m's'ī́d'm' /mɪ̀ɕɪ̄ɟɪ̃̂m/
Dative	s̍kk ák'k' /ākjɪ̂c/		g̣ṣb̤̍v̩k m's'ī́d'k' /mɪ̀ɕɪ̄ɟɪ̂c/
Genitive	s̍km ák'j /ākjɪ̂/		g̣ṣb̤̍v̩m m's'ī́d'j /mɪ̀ɕɪ̄ɟɪ̂/

In case you’re wondering, the inherent vowel “ü” /ʏ/ is, due to progressive height harmony due to “á”, lowered to “ö” whose reduced vowel is /ɪ/.

Copula

There is no equivalent word for “to be” in Tersamian, instead adjectives and nouns can be verbalised using the suffix ”-[a]g’” (/ɟ/) with inherent vowel “a”. The new verb can then be conjugated like a normal verb.

- - :ỵd̤̍ṇ
  - W’ū́l
  - /çʉ̌l̥
  - dog.G14-NOM.SG
- - g̣p̤̍tk̲:
  - m’ḗc’g’
  - m̥çīt͡ɕɪ̂ɟ/
  - white.G21-be.PRES
The dog is white.

The following are the conjugation tables for the copula based on the word class of the noun or the adjective. Unlike normal verbs, copula verbs do not have an irrealis stem. Instead, the copula suffix changes completely from ”-[a]g’” to ”-[o]da”.

Copula Suffix		Imperfective		Perfective
Imperative	-[a]g'a	Non-progressive	Progressive	Non-progressive	Progressive
Realis	Present	-[a]g'	-[a]g'aŕ	-[a]g'ae	-[a]g'aeŕ
Realis	Past	-[a]g'i(mï)	-[a]g'ā̈ŕ	-[a]g'e(z'o)	-[a]g'e(z'o)ŕä
Irrealis	Present	-[o]dō	-[o]dūŕa	-[o]dövü	-[o]dūvaŕa
Irrealis	Past	-[o]dē(më)	-[o]dā̈ŕ	-[o]de(zo)	-[o]dezā̈ŕ

Just like normal verbs, copula verbs can conjugate for perfective and progressive aspects. The progressive aspect denotes a change of state, and functions like the phrase “is becoming”.

- - :ṣq̳̍ṇm
  - S’ō̈́lj
  - /ɕēlɪ̂j
  - grass.G15-NOM.PL
- - s̍kk̲sṛ:
  - ák’g’aŕ.
  - ácɟâr/
  - big.G11-be.PRES.PROG
The grass is growing.

The perfective aspect is for completed changes in state.

- - :ṣq̳̍ṇm
  - S’ō̈́lj
  - /ɕēlɪ̂j
  - grass.G15-NOM.PL
- - s̍kk̲i:
  - ák’g’ae.
  - ácɟâe̯/
  - big.G11-be.PRES.PFV
The grass has grown.

Unlike in normal verbs, the perfective progressive has retained its original meaning in copula verbs, that is, they denote a change in state that requires multiple smaller sub-steps to complete.

Prepositions

Tersamian prepositions aren’t technically prepositions. Instead, prepositions are nouns and are included in a genitive phrase to form adpositional phrases.

To create a locative prepositional phrase, the locative prefix (px “em”) is added to the prepositional noun, and the genitive of the noun phrase is placed after the preposition. The prepositional noun is then declined for cases, usually accusative, dative or instrumental.

- - :hs̲̍v
  - Kä́p
  - /kæ̌p
  - boat.G24-NOM.SG
- - k̲s̍z
  - g’án
  - gjɛ̃̌
  - float.G24-PRES
- - pxp̤̍hḷ
  - emjḗgr
  - ɪ̀m̥çīgɨ̂r
  - LOC-up.G21-INSTR.SG
- - md̳̍ṇṇ:
  - jṻ́ll.
  - çȳl̥ɪ̂l/
  - river.G22-GEN.SG
The boat is floating on the river.

In this example, “emjḗgr” or in English, “on”, is the preposition, made up of the locative prefix “em”, the fundamental G21 noun “jḗg” for “top”, and the instrumental suffix “r”, which is the required case the locative phrase must take for the verb “g’ana”. The word for river, “jṻ́l” has the genitive case, so the prepositional phrase roughly directly translates to “at on top of the river”.

For temporal phrases, the locative prefix “em” is replaced with the temporal prefix “om”.

- - qxs̤̍
  - omā́
  - /ə̀m̥ɵ̌/
  - TMP-here.G11
Now.

Prepositional Cases

As mentioned previously, prepositional nouns can decline for different cases. Usually this would be the accusative case, which is used to specify the time or place that an event takes place.

- - :zpz
  - Nen
  - /nɪ̃̀
  - 1SG.NOM
- - qxs̤̍z
  - omā́n
  - ə̀m̥ə̃̌
  - TMP-here.G11-ACC.SG
- - s̍xsḷ:
  - ámar.
  - ámâr/
  - do.G11-PRES.PROG
I’m doing it now!

In addition, the accusative case may be used for destinations such as for the verb “háka”.

- - :ṇp
  - Le
  - /lɛ̀
  - G13.NOM.SG
- - rs̍hp
  - háko
  - ɦákâ
  - go.G11-PRES
- - os̤̍hx:
  - cā́km.
  - t͡sɑ̄kɔ̃̂m/
  - shop.G25-ACC.SG
They/He/She went to the store.

Meanwhile, the instrumental case is generally used for transient time and places, so its use can be roughly translated as “through” or anything to that effect.

- - :ṇp
  - Le
  - /lɪ̀
  - G13.NOM.SG
- - ṣs̲̍n̲p
  - s’ä́de
  - ɕɛ́ðɛ̂
  - walk.G14-PAST
- - pxṛlxḍ
  - emŕeomr
  - ɪ̀mrēo̯m̥ɪ̂r
  - LOC-around.G12-INSTR.SG
- - hp̍n̲qṇ:
  - kédol.
  - kóðɔ̂l/
  - city.G24-GEN.SG
They walked around the city.

Syntax

Modern Standard Tersamian is a generally head-initial language. This is the case for noun phrases, prepositional phrases and verb phrases. The only exception here are determiner phrases, which are head-final. MST also has SVO basic word order, however, word order is usually more flexible than this, so it is more accurate to call MST a V2 language. On top of that, MST has optional topic-prominence, with the topic placed before the first argument of the sentence.

The following sentence illustrates Tersamian’s syntax:

TODO: I need more words in my dictionary.

Relative Clauses

Relative clauses allow nouns and verbs to modified with a dependent clause. The main strategy to form relative clauses is by starting with a normal clause, leaving a blank for the shared noun. The relative noun “i” is added to the front of the relative clause, inflected for the appropriate case in the relative clause. Following that, the positions of the remaining arguments are readjusted so that V2 word order is maintained in the relative clause. If necessary, a relative clause terminator “ö” /øː/ is added to the end of the clause. With this method, the following statement…

TODO: use an example with a non-nominative relative pronoun

- - :e
  - Oe
  - /òe̯
  - 1SG.NOM
- - x̣p̍vl
  - n’épeo
  - ɲépêo̯
  - catch.G14-PAST.PFV
- - f̩p̍ḷz:
  - v’érn.
  - vjə̃̌r |
  - bird.G14-ACC.SG |
- - :f̩p̍ḷ
  - V’ér
  - vjěr
  - bird.G14-NOM.SG
- - rq̳̍ṛpṛ
  - hō̈́ŕeŕ
  - çýr̥îr
  - fly.G11-PAST.PROG
- - op̍gf̲ph:
  - céŋvek.
  - t͡sɛ́ŋvɛ̂k/
  - nest.G25-G14.GEN-DAT.SG
I caught a bird. The bird was flying back to its nest.

can be converted to this:

- - :e
  - Oe
  - /òe̯
  - 1SG.NOM
- - x̣p̍vl
  - n’épeo
  - ɲépêo̯
  - catch.G14-PAST.PFV
- - f̩p̍ḷz
  - v’érn
  - vjə̃̌r
  - bird.G14-ACC.SG
- - b
  - i
  - jɪ̀
  - REL.G14-NOM.SG
- - rq̳̍ṛpṛ
  - hō̈́ŕeŕ
  - çýr̥îr
  - fly.G11-PAST.PROG
- - op̍gf̲ph:
  - céŋvek.
  - t͡sɛ́ŋvɛ̂k/
  - nest.G25-G14.GEN-DAT.SG
I caught a bird which was flying back to its nest.

Using this construction, only relative pronouns that are nominative, instrumental, accusative or dative are accessible. To access a genitive relative pronoun, the relevant pronoun is added after the noun in the relative clause that the genitive shared noun possesses.