Adarain · June 5, 2018 20:55
diff --git a/sascha grammar and lexicon.txt b/sascha grammar and lexicon.txt
 ##LEXICON##
 b           n count         demonstrative pronoun (this, that)
 bañ         v intr cont     to speak, talk
 bañdaut     v tr mom        to say
 gag         n count         1st person pronoun
 ges         v tr mom        separate, split in twain, give birth 
 gind        v cop stat      to be (used with permanent states, nouns)
 gida        v cop cont      to be (used with temporary states, locations)
 gidat       v cop mom       to become
 h           adj             no, none of
 ham         adj             sick, ill
 ísi         n count         3rd person pronoun
 káb         n count         man
 kir         prep dat        down to
 kirbu       adv             after
 kirbuñg     adv             downstream, later, in the future
 krin        adj             old
 kun         adv             only
 mbámbuñg    adv             upstream, earlier, in the past
 mbáñ        n count         head
 mbekat      v tr mom        to agree with
 mbúre       subj            for, to
 mesittoh    n count         person
 ndás        n mass flat     ground, earth
 ndásapper   v tr cont       to bury
 pen         n mass liquid   grass
 piɀti       conj            then, thus, so, therefore
 pod         v tr mom        to send
 ruk         v tr stat       to obey
 rvs         n count         woman
 sopit       v intr cont     to be called (name takes essive case)
 sá          conj            and
 sá          prep inc        with (comitative)
 sítam       adv             more
 ȿapithe     v tr stat       to inhabit
 ȿápper      v intr mom      to stop doing sth
 ȿoɀ         v               to do so (copies aspect and valence of verb it refers to)
 ȿvpper      v intr mom      to die
 tan         v tr mom        to catch, capture, kill (in the context of hunting)
 tar         n count         father
 ten         adj             healthy
 tet         n count         2nd person pronoun
 tirda       v intr mom      to see, spot, notice
 tsok        v tr cont       to feel, smell, taste
 tȿún        v tr cont       to bind
 tȿúnda      v tr mom        to entangle, tie
 tȿúnheat    v tr mom        (of a woman) to marry
 zundhe      adj             weak
 ɀámkegȿap   n count         dream

 PROPER NOUNS
 Kar
 Soúrsvr
 Tohoúir
 Úoi
 Úortvr

 #MORPHEME LISTS#
 #Case Suffixes#
 +-----+-------+------+
 |     | Count | Mass |
 +-----+-------+------+
 | ABS | -s    | -i   |
 +-----+-------+------+
 | ESS | -t    | -e   |
 +-----+-------+------+
 | DAT | -ñ    | -uñg |
 +-----+-------+------+
 | INC | -k    | -ak  |
 +-----+-------+------+

 #Number Suffixes# 
 +----+----+-----+-----+
 |    | ∅  | SG  | PL  | 
 +----+----+-----+-----+
 | ∅  |    | -o  | -v  |
 +----+----+-----+-----+
 | SG | -n | -no | -nv |
 +----+----+-----+-----+
 | PL | -k | -ko | -kv |
 +----+----+-----+-----+
 Columns (vowels) represent the “internal person”, i.e. Absolutive on verbs/the noun itself on nouns; rows (consonants) represent the “external person” i.e. ergative/possessor
 This table could be just as well represented as two distinct morphemes. It just gets a bit weird for glossing, I prefer being able to gloss it like SG>PL for -nv

 Personal agreement suffixes
 +------+---------+-----------+------------+
 |      | Stative | Momentane | Continuous |
 +------+---------+-----------+------------+
 | 1NOM |         ga-         |     a-     |
 +------+---------------------+------------+
 | 2ERG |         te-         |     e-     |
 +------+---------------------+------------+
 | 3ERG |                i-                |
 +------+----------------------------------+
 |      | Stative | Momentane | Continuous |
 +------+---------+-----------+------------+
 | 1ACC |    -k   |           -a           |
 +------+---------+------------------------+
 | 2ABS |    -t   |           -e           |
 +------+---------+------------------------+
 | 3ABS |          -s         |     -i     |
 +------+---------------------+------------+
 Possessive prefixes take the stative/momentane column ones.

 Slot 2 TAM suffixes:
 Non-Past        -pv
 Non-Future      -∅
 Negative        -he
 Infinitive      -ko
 Imperative      -∅

 ##GRAMMAR##
 #INTRODUCTION#
 Mesak is a strictly agglutinative naturalistic conlang set in an Alt-Earth Iron Age setting. Its primary distinguishing features are an extremely prevailing progressive ±ATR harmony, deep ergativity and head- or sometimes double-markedness. 

 #PHONOLOGY#
 Mesak’s phonology as far as this relay is concerned has about four important features:

 1) Vowel harmony:
    Mesak has a [+ATR] feature which affects the pronunciation of vowels and spreads rightwards. Thus if any vowel in a word is [+ATR], all subsequent vowels receive this feature too. 
    •In the romanization, marked and unmarked vowels are spelled identically, except that the [+ATR] vowel in a word receives an acute. Thus ⟨sítam⟩ /si.təm/ “more” (both vowels [+ATR]) and ⟨mesittohvs⟩ /mɛ.set.tɔ.ʁəs/ “humans, people” (all vowels [-ATR]). The romanization is thus phonemic in this regard.
 2) Voiced plosive allophony:
    There are two rows of voiced plosives: implosives /ɓ ɗ ʄ/ and prenasalized stops /ᵐb ⁿd ᵑg/. These are contrastive in the coda, but in complementary distribution in the onset: implosives only appear before [-ATR] vowels and prenasalized stops only before [+ATR] ones. These rows additionally contrast with voiceless plosives /p t k/ in all environments. Thus the verbal formation gind-o-s “he/she is” takes on the form /ɠeɗɔs/, but if a root with a dominant vowel is incorporated, not only the vowels but also the plosives change: húh-gind-o-s “it is an owl” /ʁuʁᵑgiⁿdɔ̝s/
    •In the romanization, prenasalized stops are spelled ⟨mb nd ñg⟩, implosives ⟨b d g⟩ and voiceless stops ⟨p t k⟩. The above words are spelled ⟨gidos⟩ and ⟨húhñgindos⟩.
 3) Plosive clusters reduce to a single geminate. The behaviour for this is slightly complicated:
    -Plosives always take on the place of articulation of the latest in the cluster: t-k → kk, b-d → dd
    -Plosives also typically take on the voicing of the latest consonant in the cluster: t-d → dd, b-p → pp
    -However, if any of the plosives involved is a prenasalized stop, then the resulting geminate will remain voiced.
        -The resulting sound may be either prenasalized or implosive depending on the environment (default is prenasalized)
 4) HOMORGANIC Nasal-Plosive clusters turn into prenasalized stops. There is no nasal assimilation.


 #VERBAL MORPHOLOGY#
 Aight from now on no evil surprises I think.

 Verbs are templatic in structure, employing a somewhat straightforward polypersonal agreement structure. Additionally, they inflect for aspect, voice, tense and mood (in decreasing order of importance). Here’s the template:

      -1   |   0  |   1   |       2      |    3   |    4   |    5
    person | STEM | voice | tense & mood | person | number | person
    
 (affixes themselves can be found listed in MORPHEME LISTS earlier. Here only usage will be elaborated)
    
 -1) Person: In this slot, person prefixes cross-referencing 1.NOM, 2.ERG and 3.ERG are placed. These prefixes vary based on aspect, see below.

 0) Stem: The only truly mandatory slot can be further broken down into three subslots:
    a. Incorporated nouns (explained later)
    b. Root 
    c. Derivational suffixes. Most derivations are nontransparent and have been added to the dictionary. The exceptions to this are listed in the morpheme list
 
 1) Voice: The stem is followed by one or sometimes multiple suffix for voice. In this text, antipassive, passive and causative occur, at times on the same verb.

 2) Tense and Mood: This slot is filled one or multiple tense/mood/polarity markers. There is an internal order to these, and some may force changes elsewhere in the verbal template:
    a. Tense: Only two tenses are distinguished: non-past and non-future. 
    b. Negative: negates the whole predicate
    c. Infinitive: If the infinitive is used, then no person or number agreement takes place. The use of the infinitive will be dealt with later on.
    d. Imperative: If the imperative is used, then no person or number agreement takes place.
    
 3) Person: In this slot, person suffixes cross-referencing 1.ACC and 2.ABS are placed. These suffixes vary based on aspect, see below.
 
 4) Number: In this slot, number for both internal and external person are crossreferenced. Technically this can be broken down into two morphemes, but I choose not to for the sake of notation.
 
 5) Person: In this slot, person suffixes cross-referencing 3.ABS are placed. These suffixes vary based on aspect, see below.
 
 #Aspects#
 Verb stems inherently carry one of three aspects: stative, momentane or continuous. These more or less correspond to actions of general truth, single moments and longer timespans, but often there may be a disconnect between syntax and semantics, as aspects are primarily determined by the derivational affixes (which themselves can often only attach to certain aspects). Aspects are important in verbal morphology, as the personal affixes reflect them. Apart from that you don’t have to care too much about it.

 #Tense#
 Mesak storytelling always happens in the present, the two tenses are used to highlight whether an action is more closely related to the future or past. Often, a given paragraph, verse &c will be mostly written in the same tense. The non-future is unmarked, but cannot be assumed to be more neutral in meaning; a non-past is not necessarily a future tense.

 #NOMINAL MORPHOLOGY#
 Nouns in many ways mirror verbs in their morphology. Nouns also form a (much simpler) complex, with this structure:

    Possessor | STEM | Number | Case
    
 The possessor prefixes are identical to the personal prefixes found on stative verbs. The number suffixes are exactly identical to those found on verbs. The case suffixes at times show similarity with some other verbal affixes. 

 #Number#
 Nouns are split into count and mass nouns. The latter do not take a number suffix (except in possessive constructions), number agreement with them is done with the “logical” number, defaulting to singular. The two classes also take different case suffixes.

 #Case#
 Five cases are distinguished: Ergative, Absolutive, Dative, Instrumental-Causal (INC) and Essive. Of these, the INC only appears in a prepositional phrase here and will not be discusses further.
 •The ABS only has a few important uses, encoding the S and P arguments of verbs, the subject of copulas and the comparee in comparatives
 •The ERG is only used to encode the A argument. Its construction is directly derived from the absolutive and is explained a bit later.
 •The DAT is used with some prepositions, and as a third argument for some verbs. If the DAT is present then it is preferred over the ABS for verbal agreement
 •The ESS is Mesak’s catch-all case. Its prototypical use is to form a sort of adverbial, often temporal in meaning (e.g. this year-ESS = “in this year”). It is also used for the standard in comparatives, the case to which P gets demoted in an antipassive and the case into which possessors in a possessive construction are put.

 #Possession#
 If a noun is possessed then a marker indexing the person of the possessor is prefixed to it. The number marker on the possessed noun also indexes both number of the possessor and possessee (possessor corresponding to the ergative number index on verbs, i.e. the consonantal part of the affix).
 For example, the noun ñúsos ⟨NU·SO·S⟩ means “friend”, while gañúskos ⟨KA·NU·S·KO·S⟩ means “our friend” (breaking down to ga-ñús-ko-s → 1-friend-PL>SG-ABS).
 If the possessor is another noun, then it may be put immediately preceding the possessed noun, in the essive case: mesittohot iñúsnvs (human-SG-ESS 3-friend-SG>PL-ABS) “the human’s friends”.

 #Ergative#
 To construct the ergative, take the absolutive and treat it as if it was possessed by the particle ñ (which precedes it immediately, is encoded as singular and does not inflect). Thus ñúsos “friend” → ñ-iñúsnos.

 #Adjectives#
 Adjectives agree in person, countability and number with their nouns, but do not reflect possession.

 If an adjective already carries plural number (e.g. “many” or “five”) then the noun itself will generally be marked singular. Plural marking in this case would be interpreted as counted instances of groups:

    ñúsos dɀános “many friends” (singular marker -o-)
    ñúsvs dɀánvs “many groups of friends” (plural marker -v-)
    
 #Comparatives#
 Adjectives have a comparative degree, but this is not featured in the text. A comparative sentence will usually incorporate the adjective into the copula and precede it with sítam “more”. The entire construction would thus be structured like this: 

    X-abs Y-ess sítam ADJ-be-{inflections} 
    meaning “X is ADJer than Y”

 #INTRA-CLAUSAL SYNTAX#
 #Constituent Order#
 Mesak’s inter-clausal syntax is pretty straightforward. It can be more or less arranged by the cases involved:

    [A] [S/P] [DAT] [PP, INC, Adv] [ESS] [V]

 Where A/S/P stand for (underlying!) syntactic core relations, PP stands for prepositional phrases and Adv for adverbials, and stuff within square brackets is interchangable. Since A/S/P pronominals are (typically) realized on the verb and the other cases are somewhat rarer, it’s not uncommon for a sentence to just consist of an adverbial or essive followed by a verb.

 #NPs and PPs#
 NPs have a rigid internal order that goes as follows:

    [Possessor NP] [Adj] [N] [Rel Clause]

 PPs are of the form [Prep] [NP]

 Each NP is associated with a case and (if the head noun is countable) a number, which are referenced on the noun, adjectives and the relativizer clitic (see later). Prepositions have an inherent case (DAT or INC), which is used with definite NPs; indefinite NPs in a prepositional clause take the ESS. Possessor NPs always take ESS.


 #Voice operations#
 Mesak has several valency-altering operations: Antipassive, Passive, Causative, Reflexive and Reciprocal are the ones I’ve worked out so far and I think it could still do with some applicatives. But you don’t have to care about that. Note that multiple operations can be applied to the same verb, in which case they happen in the order they are marked for. E.g. if you see -úk- on a verb that means it’s first turned antipassive, and then causativized, whereas -kú- is the opposite.

 The antipassive is marked on the verb with the suffix -ú in slot 1 and turns a transitive verb intransitive, mapping the underlying A to S and the P to non-core (marked with Essive):

    Underlying      A       P           V-tr
    Derived         S       non-core    V-intr
    Marked with     ABS     ESS         -ú-

 The passive is marked on the verb with the suffix -gi in slot 1 and turns a transitive verb intransitive, mapping the underlying P to S and the A to non-core (marked with Instrumental-Causal):

    Underlying      A        P           V-tr
    Derived         non-core S           V-intr
    Marked with     INC      ABS         -gi-
    
 The causative is marked on the verb with the suffix -k- in slot 1 and increases the valence of the verb by one. The underlying S/P is mapped to P, the causer becomes the A and if the original verb was transitive then the A becomes non-core (marked with dative)

    Underlying      Causer  A        S/P     V-tr/intr
    Derived         A       non-core P       V-tr
    Realization     ERG     DAT      ABS     -k-
    
 There is an absolutive pivot, which more or less entails that when chaining clauses, the same NP has to feature in the pivot, i.e. be the S or P of both clauses. This is a syntactic requirement for subordinate clauses, but only a pragmatic trend for sentences within a text. 


 #Incorporation#
 In this text there’s only one use of incorporation employed: incoproration of a noun or adjective into the copula (so that X-be = to be (an) X).


 #INTER-CLAUSAL SYNTAX#
 There are three kinds of clause connections that need to be taken into account here. All of them interact with the absolutive pivot in one way or another.

 #Coordinated Clauses#
 When two clauses are coordinated, e.g. [X] and [Y], then the second clause is typically introduced with a conjunction. Both clauses are just normal main clauses behaving normally. However, they require the absolutive pivot to be identical, otherwise coordination is simply not possible. The second clause will not include an overt S or P NP (but still have verbal affixes for it).

 #Relative Clauses#
 A relative clause requires its head to be either its S or P. Relative clauses are marked by adding a clitic =neh- to the verb, which then inflects as if the relative clause was an adjective.

 #Complement Clauses with Identical Subject#
 If there is a complement clause (usually introduced by mbúre “to, for”) which has the same subject as the main clause, then the pivoting constraint must be fullfilled as well. This has the consequence that the main clause must be intransitive in this case, usually enforcing an antipassive. The subordinate verb in this instance does not inflect but is put in the infinitive instead.

 #Discourse Pragmatics#
 It is common for the same things to be put into the absolutive pivot throughout longer pieces of text, even if there is no real advantage to this.
	##LEXICON##
	b n count demonstrative pronoun (this, that)
	bañ v intr cont to speak, talk
	bañdaut v tr mom to say
	gag n count 1st person pronoun
	ges v tr mom separate, split in twain, give birth
	gind v cop stat to be (used with permanent states, nouns)
	gida v cop cont to be (used with temporary states, locations)
	gidat v cop mom to become
	h adj no, none of
	ham adj sick, ill
	ísi n count 3rd person pronoun
	káb n count man
	kir prep dat down to
	kirbu adv after
	kirbuñg adv downstream, later, in the future
	krin adj old
	kun adv only
	mbámbuñg adv upstream, earlier, in the past
	mbáñ n count head
	mbekat v tr mom to agree with
	mbúre subj for, to
	mesittoh n count person
	ndás n mass flat ground, earth
	ndásapper v tr cont to bury
	pen n mass liquid grass
	piɀti conj then, thus, so, therefore
	pod v tr mom to send
	ruk v tr stat to obey
	rvs n count woman
	sopit v intr cont to be called (name takes essive case)
	sá conj and
	sá prep inc with (comitative)
	sítam adv more
	ȿapithe v tr stat to inhabit
	ȿápper v intr mom to stop doing sth
	ȿoɀ v to do so (copies aspect and valence of verb it refers to)
	ȿvpper v intr mom to die
	tan v tr mom to catch, capture, kill (in the context of hunting)
	tar n count father
	ten adj healthy
	tet n count 2nd person pronoun
	tirda v intr mom to see, spot, notice
	tsok v tr cont to feel, smell, taste
	tȿún v tr cont to bind
	tȿúnda v tr mom to entangle, tie
	tȿúnheat v tr mom (of a woman) to marry
	zundhe adj weak
	ɀámkegȿap n count dream

	PROPER NOUNS
	Kar
	Soúrsvr
	Tohoúir
	Úoi
	Úortvr

	#MORPHEME LISTS#
	#Case Suffixes#
	+-----+-------+------+
	\| \| Count \| Mass \|
	+-----+-------+------+
	\| ABS \| -s \| -i \|
	+-----+-------+------+
	\| ESS \| -t \| -e \|
	+-----+-------+------+
	\| DAT \| -ñ \| -uñg \|
	+-----+-------+------+
	\| INC \| -k \| -ak \|
	+-----+-------+------+

	#Number Suffixes#
	+----+----+-----+-----+
	\| \| ∅ \| SG \| PL \|
	+----+----+-----+-----+
	\| ∅ \| \| -o \| -v \|
	+----+----+-----+-----+
	\| SG \| -n \| -no \| -nv \|
	+----+----+-----+-----+
	\| PL \| -k \| -ko \| -kv \|
	+----+----+-----+-----+
	Columns (vowels) represent the “internal person”, i.e. Absolutive on verbs/the noun itself on nouns; rows (consonants) represent the “external person” i.e. ergative/possessor
	This table could be just as well represented as two distinct morphemes. It just gets a bit weird for glossing, I prefer being able to gloss it like SG>PL for -nv

	Personal agreement suffixes
	+------+---------+-----------+------------+
	\| \| Stative \| Momentane \| Continuous \|
	+------+---------+-----------+------------+
	\| 1NOM \| ga- \| a- \|
	+------+---------------------+------------+
	\| 2ERG \| te- \| e- \|
	+------+---------------------+------------+
	\| 3ERG \| i- \|
	+------+----------------------------------+
	\| \| Stative \| Momentane \| Continuous \|
	+------+---------+-----------+------------+
	\| 1ACC \| -k \| -a \|
	+------+---------+------------------------+
	\| 2ABS \| -t \| -e \|
	+------+---------+------------------------+
	\| 3ABS \| -s \| -i \|
	+------+---------------------+------------+
	Possessive prefixes take the stative/momentane column ones.

	Slot 2 TAM suffixes:
	Non-Past -pv
	Non-Future -∅
	Negative -he
	Infinitive -ko
	Imperative -∅

	##GRAMMAR##
	#INTRODUCTION#
	Mesak is a strictly agglutinative naturalistic conlang set in an Alt-Earth Iron Age setting. Its primary distinguishing features are an extremely prevailing progressive ±ATR harmony, deep ergativity and head- or sometimes double-markedness.

	#PHONOLOGY#
	Mesak’s phonology as far as this relay is concerned has about four important features:

	1) Vowel harmony:
	Mesak has a [+ATR] feature which affects the pronunciation of vowels and spreads rightwards. Thus if any vowel in a word is [+ATR], all subsequent vowels receive this feature too.
	•In the romanization, marked and unmarked vowels are spelled identically, except that the [+ATR] vowel in a word receives an acute. Thus ⟨sítam⟩ /si.təm/ “more” (both vowels [+ATR]) and ⟨mesittohvs⟩ /mɛ.set.tɔ.ʁəs/ “humans, people” (all vowels [-ATR]). The romanization is thus phonemic in this regard.
	2) Voiced plosive allophony:
	There are two rows of voiced plosives: implosives /ɓ ɗ ʄ/ and prenasalized stops /ᵐb ⁿd ᵑg/. These are contrastive in the coda, but in complementary distribution in the onset: implosives only appear before [-ATR] vowels and prenasalized stops only before [+ATR] ones. These rows additionally contrast with voiceless plosives /p t k/ in all environments. Thus the verbal formation gind-o-s “he/she is” takes on the form /ɠeɗɔs/, but if a root with a dominant vowel is incorporated, not only the vowels but also the plosives change: húh-gind-o-s “it is an owl” /ʁuʁᵑgiⁿdɔ̝s/
	•In the romanization, prenasalized stops are spelled ⟨mb nd ñg⟩, implosives ⟨b d g⟩ and voiceless stops ⟨p t k⟩. The above words are spelled ⟨gidos⟩ and ⟨húhñgindos⟩.
	3) Plosive clusters reduce to a single geminate. The behaviour for this is slightly complicated:
	-Plosives always take on the place of articulation of the latest in the cluster: t-k → kk, b-d → dd
	-Plosives also typically take on the voicing of the latest consonant in the cluster: t-d → dd, b-p → pp
	-However, if any of the plosives involved is a prenasalized stop, then the resulting geminate will remain voiced.
	-The resulting sound may be either prenasalized or implosive depending on the environment (default is prenasalized)
	4) HOMORGANIC Nasal-Plosive clusters turn into prenasalized stops. There is no nasal assimilation.


	#VERBAL MORPHOLOGY#
	Aight from now on no evil surprises I think.

	Verbs are templatic in structure, employing a somewhat straightforward polypersonal agreement structure. Additionally, they inflect for aspect, voice, tense and mood (in decreasing order of importance). Here’s the template:

	-1 \| 0 \| 1 \| 2 \| 3 \| 4 \| 5
	person \| STEM \| voice \| tense & mood \| person \| number \| person

	(affixes themselves can be found listed in MORPHEME LISTS earlier. Here only usage will be elaborated)

	-1) Person: In this slot, person prefixes cross-referencing 1.NOM, 2.ERG and 3.ERG are placed. These prefixes vary based on aspect, see below.

	0) Stem: The only truly mandatory slot can be further broken down into three subslots:
	a. Incorporated nouns (explained later)
	b. Root
	c. Derivational suffixes. Most derivations are nontransparent and have been added to the dictionary. The exceptions to this are listed in the morpheme list

	1) Voice: The stem is followed by one or sometimes multiple suffix for voice. In this text, antipassive, passive and causative occur, at times on the same verb.

	2) Tense and Mood: This slot is filled one or multiple tense/mood/polarity markers. There is an internal order to these, and some may force changes elsewhere in the verbal template:
	a. Tense: Only two tenses are distinguished: non-past and non-future.
	b. Negative: negates the whole predicate
	c. Infinitive: If the infinitive is used, then no person or number agreement takes place. The use of the infinitive will be dealt with later on.
	d. Imperative: If the imperative is used, then no person or number agreement takes place.

	3) Person: In this slot, person suffixes cross-referencing 1.ACC and 2.ABS are placed. These suffixes vary based on aspect, see below.

	4) Number: In this slot, number for both internal and external person are crossreferenced. Technically this can be broken down into two morphemes, but I choose not to for the sake of notation.

	5) Person: In this slot, person suffixes cross-referencing 3.ABS are placed. These suffixes vary based on aspect, see below.

	#Aspects#
	Verb stems inherently carry one of three aspects: stative, momentane or continuous. These more or less correspond to actions of general truth, single moments and longer timespans, but often there may be a disconnect between syntax and semantics, as aspects are primarily determined by the derivational affixes (which themselves can often only attach to certain aspects). Aspects are important in verbal morphology, as the personal affixes reflect them. Apart from that you don’t have to care too much about it.

	#Tense#
	Mesak storytelling always happens in the present, the two tenses are used to highlight whether an action is more closely related to the future or past. Often, a given paragraph, verse &c will be mostly written in the same tense. The non-future is unmarked, but cannot be assumed to be more neutral in meaning; a non-past is not necessarily a future tense.

	#NOMINAL MORPHOLOGY#
	Nouns in many ways mirror verbs in their morphology. Nouns also form a (much simpler) complex, with this structure:

	Possessor \| STEM \| Number \| Case

	The possessor prefixes are identical to the personal prefixes found on stative verbs. The number suffixes are exactly identical to those found on verbs. The case suffixes at times show similarity with some other verbal affixes.

	#Number#
	Nouns are split into count and mass nouns. The latter do not take a number suffix (except in possessive constructions), number agreement with them is done with the “logical” number, defaulting to singular. The two classes also take different case suffixes.

	#Case#
	Five cases are distinguished: Ergative, Absolutive, Dative, Instrumental-Causal (INC) and Essive. Of these, the INC only appears in a prepositional phrase here and will not be discusses further.
	•The ABS only has a few important uses, encoding the S and P arguments of verbs, the subject of copulas and the comparee in comparatives
	•The ERG is only used to encode the A argument. Its construction is directly derived from the absolutive and is explained a bit later.
	•The DAT is used with some prepositions, and as a third argument for some verbs. If the DAT is present then it is preferred over the ABS for verbal agreement
	•The ESS is Mesak’s catch-all case. Its prototypical use is to form a sort of adverbial, often temporal in meaning (e.g. this year-ESS = “in this year”). It is also used for the standard in comparatives, the case to which P gets demoted in an antipassive and the case into which possessors in a possessive construction are put.

	#Possession#
	If a noun is possessed then a marker indexing the person of the possessor is prefixed to it. The number marker on the possessed noun also indexes both number of the possessor and possessee (possessor corresponding to the ergative number index on verbs, i.e. the consonantal part of the affix).
	For example, the noun ñúsos ⟨NU·SO·S⟩ means “friend”, while gañúskos ⟨KA·NU·S·KO·S⟩ means “our friend” (breaking down to ga-ñús-ko-s → 1-friend-PL>SG-ABS).
	If the possessor is another noun, then it may be put immediately preceding the possessed noun, in the essive case: mesittohot iñúsnvs (human-SG-ESS 3-friend-SG>PL-ABS) “the human’s friends”.

	#Ergative#
	To construct the ergative, take the absolutive and treat it as if it was possessed by the particle ñ (which precedes it immediately, is encoded as singular and does not inflect). Thus ñúsos “friend” → ñ-iñúsnos.

	#Adjectives#
	Adjectives agree in person, countability and number with their nouns, but do not reflect possession.

	If an adjective already carries plural number (e.g. “many” or “five”) then the noun itself will generally be marked singular. Plural marking in this case would be interpreted as counted instances of groups:

	ñúsos dɀános “many friends” (singular marker -o-)
	ñúsvs dɀánvs “many groups of friends” (plural marker -v-)

	#Comparatives#
	Adjectives have a comparative degree, but this is not featured in the text. A comparative sentence will usually incorporate the adjective into the copula and precede it with sítam “more”. The entire construction would thus be structured like this:

	X-abs Y-ess sítam ADJ-be-{inflections}
	meaning “X is ADJer than Y”

	#INTRA-CLAUSAL SYNTAX#
	#Constituent Order#
	Mesak’s inter-clausal syntax is pretty straightforward. It can be more or less arranged by the cases involved:

	[A] [S/P] [DAT] [PP, INC, Adv] [ESS] [V]

	Where A/S/P stand for (underlying!) syntactic core relations, PP stands for prepositional phrases and Adv for adverbials, and stuff within square brackets is interchangable. Since A/S/P pronominals are (typically) realized on the verb and the other cases are somewhat rarer, it’s not uncommon for a sentence to just consist of an adverbial or essive followed by a verb.

	#NPs and PPs#
	NPs have a rigid internal order that goes as follows:

	[Possessor NP] [Adj] [N] [Rel Clause]

	PPs are of the form [Prep] [NP]

	Each NP is associated with a case and (if the head noun is countable) a number, which are referenced on the noun, adjectives and the relativizer clitic (see later). Prepositions have an inherent case (DAT or INC), which is used with definite NPs; indefinite NPs in a prepositional clause take the ESS. Possessor NPs always take ESS.


	#Voice operations#
	Mesak has several valency-altering operations: Antipassive, Passive, Causative, Reflexive and Reciprocal are the ones I’ve worked out so far and I think it could still do with some applicatives. But you don’t have to care about that. Note that multiple operations can be applied to the same verb, in which case they happen in the order they are marked for. E.g. if you see -úk- on a verb that means it’s first turned antipassive, and then causativized, whereas -kú- is the opposite.

	The antipassive is marked on the verb with the suffix -ú in slot 1 and turns a transitive verb intransitive, mapping the underlying A to S and the P to non-core (marked with Essive):

	Underlying A P V-tr
	Derived S non-core V-intr
	Marked with ABS ESS -ú-

	The passive is marked on the verb with the suffix -gi in slot 1 and turns a transitive verb intransitive, mapping the underlying P to S and the A to non-core (marked with Instrumental-Causal):

	Underlying A P V-tr
	Derived non-core S V-intr
	Marked with INC ABS -gi-

	The causative is marked on the verb with the suffix -k- in slot 1 and increases the valence of the verb by one. The underlying S/P is mapped to P, the causer becomes the A and if the original verb was transitive then the A becomes non-core (marked with dative)

	Underlying Causer A S/P V-tr/intr
	Derived A non-core P V-tr
	Realization ERG DAT ABS -k-

	There is an absolutive pivot, which more or less entails that when chaining clauses, the same NP has to feature in the pivot, i.e. be the S or P of both clauses. This is a syntactic requirement for subordinate clauses, but only a pragmatic trend for sentences within a text.


	#Incorporation#
	In this text there’s only one use of incorporation employed: incoproration of a noun or adjective into the copula (so that X-be = to be (an) X).


	#INTER-CLAUSAL SYNTAX#
	There are three kinds of clause connections that need to be taken into account here. All of them interact with the absolutive pivot in one way or another.

	#Coordinated Clauses#
	When two clauses are coordinated, e.g. [X] and [Y], then the second clause is typically introduced with a conjunction. Both clauses are just normal main clauses behaving normally. However, they require the absolutive pivot to be identical, otherwise coordination is simply not possible. The second clause will not include an overt S or P NP (but still have verbal affixes for it).

	#Relative Clauses#
	A relative clause requires its head to be either its S or P. Relative clauses are marked by adding a clitic =neh- to the verb, which then inflects as if the relative clause was an adjective.

	#Complement Clauses with Identical Subject#
	If there is a complement clause (usually introduced by mbúre “to, for”) which has the same subject as the main clause, then the pivoting constraint must be fullfilled as well. This has the consequence that the main clause must be intransitive in this case, usually enforcing an antipassive. The subordinate verb in this instance does not inflect but is put in the infinitive instead.

	#Discourse Pragmatics#
	It is common for the same things to be put into the absolutive pivot throughout longer pieces of text, even if there is no real advantage to this.