Proto-Bakarh

Setting and Information
Proto-Bakarh is reconstructed a fully polysynthetic language in the narrowest sense: it featured both noun incorporation and polypersonality; it supposedly was a primarily fusional, secondarily agglutinating language. Spoken some 6C00 years before the Golden Age, it is the ancestor of all Bakaric languages henceforth spoken.

Unless explicitly noted, all material here is to be taken as dubious and uncertain.

Proto-Bakarh was indeed a language that was written (as noted in Psarh archive records dating to the era) but nothing has survived. The cause of this might be systematic destruction of documents at the beginning of the Golden Age, initiated and done by Ḱægx́a and his brood.

Phonotactics
The phonotactics of PB have been reconstructed to approximately this format:

Here, the one-letter tokens represent the following:
 * N - any nasal or approximant or /l/
 * S - any obstruent
 * C - any consonant
 * V - any vowel
 * F - any fricative or /ɦ/

No obstruent can follow a geminate in the onset, while no such rule exists for the coda. Aspirated plosives cannot be word-final, while no such restriction is imposed upon fricatives. Geminate plosives cannot be initial, while no such restriction is imposed upon fricatives.

Assimilation
Proto-Bakarh had an assimilatory retroflex rule - any /ST/ and /TS/ cluster, where "S ∈ {s ʂ}" and "T ∈ {t ʈ}", must agree in place of articulation - if any one element is retroflex, the other assimilates to it in point of articulation.

Basics
Proto-Bakarh was a verb-initial language in which modifiers followed the modified and heads often preceed dependents.

Terminology
Several lexical categories are found in Proto-Bakarh. They are:
 * Nouns - any lexical item that can take a case and become an argument.
 * Verbs - any lexical item that conveys an action and can (but isn't obligated to) take an argument.
 * Modifiers - any lexical item that modifies another lexical item.
 * Particles - any lexical item that serves a role in the regulation of sentences and expresses relations between phrases.

There are several items which don't fit into any of the categories above, and as such are sorted as Uncategorised. Most of the uncategorised items can fit in into some or all of the categories according to their function at that moment, but a few have unique functions.

Morphology
In essence, Proto-Bakarh morphology is divided into two large categories: synthetic morphology and analytical morphology. Synthetic morphology refers to direct changes to the words themselves, while analytical morphology to the various associated functional morphemes that are not phonologically bound to content words.

Often these two morphologies mix and the results and the processes themselves are reflected differently in various descendants.

Analytic Morphology
Proto-Bakarh analytic morphology is preserved in all daughter languages as it was favoured more over the synthetic kind. Analytic morphology of Proto-Bakarh is divided into adjective morphology and adverb morphology.

Synthetic Morphology
In essence, synthetic morphology had already begun simplifying by the Proto-Bakarh period. Almost all of it is preserved, but certain aspects of it, not reflected in any daughter language, show up only in foreign loans of the period. The language's synthetic morphology is divided according to the two major word classes: verb morphology and noun morphology.

Suffix
In Proto-Bakarh, suffixes are attached directly to the root. They're usually derivative in nature. A root with a suffix attached forms a complete core. A root may have multiple suffixes. Generally speaking, there are a few morphophonological suffixation rules in PB: Proto-Bakarh suffixes:
 * Short vowels merge with /ɦ/ in varying ways, depending on their position in the mouth.
 * If the vowel is back, the combination /ɦV/ becomes /ɰu/.
 * If the vowel is front, the combination /ɦV/ becomes /ŋi/.
 * Two short vowels will collapse into one long vowel whose height is determined by the first and frontness by the second vowel.
 * Two long vowels become seperated by a consonant - the exact consonant is determined by the first vowel.
 * If the first vowel is high, the seperator consonant is /ŋ/.
 * If the first vowel is low, the seperator consonant is /x/.

Noun Morphology
The Proto-Bakarh noun is made up of the following components:

The abbreviations above mean the following:
 * - a prefix attached to the core of the noun
 * - the basic form of the noun, not simplifiable further
 * - a suffix (often derivational in nature) attached to to form the core
 * - the declension of the noun
 * - an optional clitic added to the noun to form a noun complex.

Root
The root is the smallest coherent syntactial and morphological unit we find in Proto-Bakarh. While it is not a word, it is more independent than pure morphemes. Roots in Proto-Bakarh are almost exclusively of a single syllable, although disyllabic roots can also be found but in smaller amounts. There is only one reconstructed trisyllabic root in the language:  (Ōktarh species) - as it is a loanword from Proto-Dnaric, it isn't considered a proper example of the structure of Proto-Bakarh roots.

Monosyllabic roots in Proto-Bakarh never contained both a lateral and a rhotic at the same time, although disyllables sometimes did.

Disyllabic roots in Proto-Bakarh contained either only front or only back vowels in their base form,  but this could change when modified.

The peculiar property of Proto-Bakarh roots was that they changed their vowels during certain processes. This is called ablaut, and there were three kinds of ablaut:
 * 1) Front-back ablaut (shīŋH ~ shūŋH)
 * -This form of ablaut substitutes the vowel in question with its opposite on the horizontal axis.
 * 2) High-low ablaut (nłosH ~ nłusH)
 * -This form of ablaut substitutes the vowel in question with its opposite on the vertical axis.
 * 3) Short-long ablaut (ṣisxam ~ ṣisxām)
 * -This form of ablaut substitutes the vowel in question with its opposite in regards to length.

All three forms of ablaut are morphologically conditioned and occur only in specific enviroments. The vowel undergoing ablaut is always the last vowel in the root. Each ablaut can be applied only once (no stacking of ablauts is possible), and any further ablauts of the kind previously used shall invariably be ignored.

A few special roots to keep in mind:


 * 1) <ŋasq> - its basic usage is as an anchor in top-level phrases (phrases that do not stand below anything except the sentence unit).
 * This root doesn't change the length of its vowel if any other ablaut applies.
 * 2) <ŋēt> - its basic usage is an anchor in non-top-level phrases (phrases that are under other phrases).
 * This root doesn't change its vowel quality.
 * 3)  - it is used as a general negative.
 * This root doesn't change the length of its vowel.
 * 4) - it is used as a general positive.
 * This root doesn't change the length of its vowel.

Ending
Endings of nouns show its number, case and definiteness. Case and number are correlated but not truly conflated. The table below demonstrates the relationship between case and number:

The consonants in brackets surface only if the word they're attached to ends in a vowel. Marking for definiteness simply requires an additional ending, <-qqū> - indefinite is the implied form.

Verb Morphology
Proto-Bakarh verbs are formed with a verb template.

The abbreviations above mean the following:
 * - the inchoative slot
 * - a prefix attached to the core of the verb
 * - the basic form of the verb, not simplifiable further
 * - an incorporate
 * - a suffix (often derivational in nature) attached to to form the core
 * - the "subject" argument agreement slot
 * - the "object" argument agreement slot
 * - the evidential slot
 * - a suffix showing the age of the information
 * - the dynamicity slot
 * - the benefactive argument agreement slot
 * - the malefactive argument agreement slot
 * - the instrument argument agreement slot
 * - the aspect slot
 * - the tense slot
 * - the intentive slot
 * - the polarity slot
 * - the deixis slot
 * - deictic polarity
 * - the deontic slot
 * - the realis/deliberative slot
 * - the subordinate slot

Inchoative Slot (-2)
The inchoative slot is the first affix in the Proto-Bakarh verb template, being denoted by its position, "-2". The slot can be filled by either no or one prefix.

There are two possible prefixes that can fill the slot: the inchoative proper and the prospective. The inchoative proper refers to the beginning of the state or action denoted by the verb. In verbs marked with the inchoative, the action denoted by the verb is considered to have begun at the time of reference. The prospective differs from the inchoative in that it refers to the time just prior to the beginning of the state or action denoted by the verb.

The two prefixes are:

Incorporation (0)
Proto-Bakarh had a rudimentary system of basic, bleached incorporated roots that were used to show further semantic dimensions of the noun. These incorporates weren't related to incorporated arguments.

A few common incorporates:

Core Agreement (+1)
The macrotheme's and brachytheme's consonants are added only if they are preceded by a vowel or .

Explanation
Proto-Bakarh core agreement was a rather interesting feature of the language. Despite the use of terms "subject" and "object" agreement (an artefact of Ktarh linguistic terminology), Proto-Bakarh had a much different system.

In reality, the "subject" slot usually agrees with the narrative argument while the "object" simmilarly usually agrees with the receptive argument. Deviation from this is due to argument promotion either in the direction "narrative → receptive" or, less frequently, "receptive → narrative".

The narrative was, essentially, a high-activity marker; arguments most changed by the action get marked with it. It was also sometimes used to mark sole arguments of intransitive verbs. The narrative had a secondary marking of topicality - arguments marked with it tended to be marked in prominence as well. The receptive worked in reverse most of the time; arguments marked with the receptive are the ones that were least changed by the action. It was also infrequently used to mark sole arguments of intransitive verbs.

Intransitive arguments marked with the narrative are usually also considered to be volitive participants, while the ones marked with the receptive aren't.

To mark a receptive argument for prominence, one would shift its marking to the narrative; to mark a narrative argument for prominence, one would shift the receptive argument's marking to the genitive and the narrative's marking to a receptive. Even when the case suffixes change, the verb still encodes for the arguments as if they remained the same (so, if a receptive argument changed its marking to narrative, it would still be receptive to the verb). This is optional when the narrative changes to the receptive.

Evidentiality (+2)
Two morphophonological rules apply with evidentiality and information age:
 * 1) If a verb has both core agreement and an evidentiality suffix beginning in /r/, the combination /sr/ assimilates to /ʂ/ (  = <ṣ> ).
 * 2) Any sequence of /uu/ shall assimilate to /u:/, left-to-right. Note that /ɰ/ surfaces only when there would theoretically be a /uuu/ sequence (that is, a theoretical /uuu/ sequence would be realised as /u:ɰu/).

Explanation
Evidentiality in Proto-Bakarh primarily deals with the source of information and secondarily deals with the reliability of the information. It is analysed in tandem with information age which shows the datedness of the information.

Any verb marked with visual evidentiality is stated to be information directly observed by the speaker. Its source may be both visual, as per its name, and tactile - Proto-Bakarh doesn't distinguish between the two sources.

Any verb marked with olfactory evidentiality is stated to be information gleaned by the speaker through his nasal olfactory organs. Information marked with olfactory evidentiality is considered more reliable than visual although its scope is more narrow.

Verbs marked with gustatory evidentiality are stated to be information gathered by the speaker through oral organs. As Oktarh oral apparatus are diverse, gustatory evidentiality can also include information gleaned through scents by oral organs.

The reported evidentiality concerns itself with information heard, either by the speaker or by someone else and thus reported to the speaker - Proto-Bakarh doesn't distinguish between these types of auditory evidentialities (instead relying on context) although there is evidence that it used to at an earlier stage. It is considered less reliable than other kinds of evidentiality.

Assumed evidentiality is used to mark information deduced, assumed or hypothetically set up by the speaker for further consideration. It isn't marked for reliability, but is considered sideline, exemplative or alternative information.

Information age is usually analysed together with evidentiality as both have similar mechanisms of action and abide by the same morphophonological rules.

Information not previously known before being introduced is marked with the suffix <-(w)uł>. It is information assumed not to have been brought into discourse or to attention beforehand.

Information that was recently brought up is marked with <-ŋaṣ> - it was brought up before but not that far before and is therefore considered relevant information.

Information marked with <-Hoŋ> is, on the contrary, considered old and/or common knowledge and is less likely to be relevant to any current discussion. Information that is old but relevant to the current discussion is instead marked with <-(w)uł>, possibly implying it's been overlooked.

Dynamicity (+3)
The dynamicity slot is one of the two "host slots", as they are called, in Proto-Bakarh. The other such slot referred to as a host is the intentive (+5).

The reason the dynamicity slot is called a host slot is that the secondary agreement complex can only be after either it or the intentive slot.

The dynamic is a marker that denotes that the action had brought upon a noticeable, significant and relevant change of state, while the static instead denotes a more invariant situation.

The dynamicity slot is always filled if the second agreement complex is present and isn't distanced from the root.