Mestian

General information
Mestian (natively Mástā) is the language of some of the dragon-herding peoples in the central lowlands and on the Dragonforge. It is part of the lowlands Sprachbund. It is part of the Adaric language family and is a distant relative of Sarhan.

Speaking from a mixed diachronic and synchronic point of view, Mestian is a heavily fusional language descended from a heavily agglutinating stage. Many of the previously agglutinative components of the now-fusional morphemes are still vaguely evident and partially identifiable, although many have long since merged beyond transparency. Its tonal system descends from a mixture of stress accent and the disappearance of /*h/: most dipping long vowels come from a vowel that was lengthened by the loss of that /*h/.

Vowels
Mestian features distinctive stress and a pitch accent system: short vowels can either be stressed, marked with an acute accent such as in <ú>, or unstressed (unmarked), while long vowels can have either a peaking or dipping intonation when stressed. Dipping is marked with a tilde diacritic, such as <ũ>, and peaking is marked with a circumflex, such as <û>. The dipping and peaking diacritics replace the macron of long vowels.

Nominals
The category of nominals in Mestian includes nouns, pronouns and adjectives.

Nouns
Mestian has a moderately complex fusional noun system that distinguishes ten cases and three numbers. Mestian nouns are grouped based on their declension patterns -- the suffixes they take -- and their accentual paradigms -- how their accent shifts when they inflect.

They have five genders: ignic, aquatic, animate, inanimate and neuter. The division into these gender classes is largely arbitrary.

The ten cases of Mestian are:
 * Nominative
 * Accusative
 * Partitive
 * Possessive¹ (inalienable)
 * Possessive² (alienable)
 * Dative
 * Vocative
 * Commitative
 * Lative
 * Locative

Accentuation Patterns
Nouns generally fit into one of four large accentual patterns:


 * Short
 * Mobile -- most frequent
 * Static -- very rare
 * Long
 * Mobile -- common
 * Static -- slightly uncommon

While each of the patterns has its irregularities, the most common type of motion is relocation of stress in the direction of extension when the stem gets extended. Mobile long nouns generally either leave behind a long vowel upon relocation (either tone) or geminate the following consonant (only dipping). Static-stressed short nouns are overwhelmingly those that have a stressed short vowel followed by an unstressed long vowel: stress in these cases doesn't shift to the long vowel.

Many long nouns occasionally change their tone between paradigms.

(I) A-Stems
Mestian a-stems make up the first declension class in the language. By definition they are nouns with an <-a> in the nominative, optionally followed by a single consonant. A-stems can be in any of the five genders. The declension class has two subtypes: The two subtypes differ minimally in the suffixes they take.
 * Hard a-stems -- ending in a hard consonant followed by <-a>
 * Soft a-stems -- ending in any one of {š ž tš dž j} followed by <-a>

The three main suffix classes are:
 * a-class
 * as-class
 * an-class

The as-class nouns merge the nominative and accusative singulars and plurals, while an-class nouns merge the nominative and commitative cases in all numbers. The most common accentuation patterns of the a-stems are the short mobile-stressed nouns that make up the bulk of its lexical mass, as well as both mobile-stressed and static-stressed long nouns.

Neuter and inanimate nouns merge the two possessive cases.

An example aquatic a-class noun: pírka, pirkássa (fish, aqu):

An example inanimate a-class noun: rhýka, rhykássa (pebble, ina):

An example animate an-class noun: itáran, itarássa (swan, ani):

An example ignic as-class noun: apákas, apakássa (breath, ign):

(II) UR-Stems
Mestian ur-stems make up the second declension class in the language. They are nouns that end in <-ur> in the nominative. They are most often ignic, inanimate or neuter. They do not have subclasses.

Neuter and inanimate nouns merge the two possessive cases.

An example ignic noun: ĩkur, ikkússa (snake, ign):

An example neuter noun: âghur, āghússa (egg/ovum, neu):

(III) Ā-Stems
Mestian ā-stems make up the third declension class in the language. They are nouns that end in <-ā> in the nominative, and are almost exclusively neuter. All nouns of the class merge the two possessive cases, regardless of gender.

The ā-stems come in two subclasses:
 * regular-class
 * džā-class

The džā-class category specifically includes abstract nouns derived from adjectives using the derivational <-džā> suffix. These nouns have their own unique declension that doesn't quite line up with the regular declension of this declension class, although similarities are too great to be ignored.

An example regular neuter noun: qásā, qasášša (sound, neu):

An example regular neuter noun (different stress paradigm): elétā, elétašša (table knife, neu):

An example džā-class neuter noun: tándžā, tandžâšša [tán(es)-džā] (tension, neu):

(IV) Ę/Ą-Stems
The fourth declension superclass in Mestian includes nouns that end in the vowels <-ę> and <-ą>. The super-class has two paradigms, corresponding to the endings, that have parallels in declension and are close enough to be grouped into one category.

An example ę-stem aquatic noun: ýšpę, yšpímmę (tadpole, aqu):

An example ą-stem animate noun: óllą, ollímmą (sky, ani):

(V) Ū/Ī-Stems
The fifth Mestian declension class includes nouns that end in the vowels <-ū> and <-ī>. The two endings do not actually cause a difference in inflection: the only difference is that nouns respectively take <-u-> and <-i-> as thematic vowels.

An example ū-stem aquatic noun: dásū, dasúššą (blackfish, aqu):

An example ī-stem inanimate noun: akãkī, akakkíššą (revelation, ina):

(VI) Nasal Stems
Mestian nasal stems are nouns that end in any of {m n ŋ}, preceded by either a vowel or a single consonant. The category systematically includes nouns ending in <-an> as standard members; many of these overlap with a-stems that end in <-an>. A good number of these nouns are slightly irregular in unpredictable and uncategorisable ways.

An example ignic noun: trąkún, trąkúnnu (small dragon, ign):

An example animate noun: awãraŋ, awarhúnnu (muscle fibre, ani):

(VII) L-Stems
Mestian l-stems are nouns that end in <-l> in the nominative; many of them are irregular and have complex stem alternations.

An example l-stem animate noun: tâlhęl, tąllêlle (blink, ani):

(VIII) Alveopalatal Stems
Mendian alveopalatal stems are nouns that end in an alveopalatal consonant in the nominative, namely one of {š ž tš dž}.

An example alveopalatal-stem inanimate noun: béč, béčella (boulder, ina):

(IX) Consonantal Stems
Mestian consonantal stems are generically nouns that end in consonants other than {m n ŋ l š ž tš dž}, while also excluding those that end in <-ur> (but not nouns that generically end in <-r> preceded by another vowel).

Example consonant-stem animate noun: dârab, dârabašša (fowl, ani):

Pronouns
The pronominal system of Mestian only has dedicated pronouns for first and second persons; third person pronouns as such don't quite exist as their role is taken up by demonstratives that all have full declensions.

First and Second Persons
Both the first and second person pronouns in Mestian have emphatic and unstressed forms for all case-number combinations. The first person pronoun lacks a vocative (due to semantics), and both persons lack the possessive cases, compensating with dedicated possessive pronouns.

The first person pronoun is:

The second person pronoun is:

Third Person
The third person pronominal system is more open: all the pronoun types follow one declension pattern.

The proximate pronoun dínaj (this one) serves as a good example:

All other third person pronouns decline the same. Some common ones:
 * dívaj -- that one (that you can see)
 * kévaj -- that one (that you cannot see)
 * kénaj -- this one (that you cannot see)
 * âssaj -- the previously mentioned one
 * ŷlaj -- the same one
 * ûttaj -- relative pronoun

The third person pronouns all take a gender determiner that depends on the gender of that which they refer to. These determiners aren't added when people are referred to. The determiners are:

These determiners cause stress shift and are attached directly onto the inflected pronoun.

Numbers
Numbers in Mestian are an intermediate category: they act as nominal modifiers, akin to adjectives, but always decline like fifth declension nouns, without regard to gender. Nominalised and pronominalised numbers can take any gender, as appropriate to what they refer to.

Mestian uses a duodecimal (base 12) counting system. It constructs multiples of higher positions than one in a compound fashion: they are shaped like compounds of a radix number in the lative singular, optionally without its final number, and a declinable position multiplir that follows it. Numbers built as sums of different radix-multiplier multiples are written sequentially in an ascending order (ones, then twelves, and so on) either with a linking tā, or naked.

Adjectives
Mestian nominal adjectives follow a similar paradigm pattern as do nouns: they are divided into declensions and inflect for number and case. Adjectives have two more dimensions: they also inflect to indicate the gender of nouns they stand with and have degrees of comparison. Neuter nominative and lative singular adjectives can also be used as adverbs and thus modify both verbs and other adjectives.

A full adjective citation is made up from anywhere from two to ten principal parts, indicating gender forms and accent alternations. Some adjectives merge some of their genders and thus do not make a full distinction. Gender in adjectives is most commonly indicated by the adjective taking a different, gender-appropriate declension: some declensions are linked to one gender while others may hold adjectives of several different genders or multiple genders at once.

Comparison of adjectives is handled with a derivational process: comparatives and superlatives are handled by accentually weak suffixes that are attached to the stem of the adjective. Adjectives distinguish positive, comparative, superlative and excessive degrees:

Adjectives generally follow stem patterns: many adjectives can be grouped into principal part families or classes based on similar patterns. The most common classes are:

(Ia) AK-Stems
Mestian ak-stem adjectives generally make up the bulk of inanimate and neuter adjectives. They closely follow the nominal a-stem declension, distinguishing the possessive cases even when the nouns they stand with do not. An example ak-stem adjective: sânnak, sānnássa (cruel, ina/neu):

(Ib) AS-Stems
Mestian as-stem adjectives make up the plurality of Mestian ignic adjectives. They closely follow the nominal as-class declension. An example as-stem adjective: ũlas, ullássa (beige/cream, ign):

(IIa) U-Stems
Mestian u-stem adjectives generally contain ignic and animate adjectives. They share some suffix shapes with ū-stems and ur-stems. Example u-stem adjective: tánu, tanúšša (tense, ign):

(IIb) US-Stems
Us-stem adjectives are exclusively ignic. They are often somewhat irregular as they often derive from syncopated or contracted adjectives and nouns of other classes. Their declension follows several patterns, and as such aren't specifically tied to one single declension class. Most adjectives of this class exist only in it or exist in only it and within u-stems.

An example us-stem adjective: ĩmus, īmbíllu (bubbly, ign):

(IIIa) EJA-Stems
Mestian eja-stem adjectives include aquatic, ignic and neuter adjectives. They are for the most part derived adjectives, but also contain the majority of aquatic adjectives. They follow a relatively simple declension pattern. An example eja-stem adjective: gháreja, ghárajšša (black, aqu/ign):

(IIIb) ESA-Stems
Esa-stem adjectives are primarily ingnic, although there exists a significant inanimate and neuter minority. They follow a relatively simple declension pattern. A sample esa-stem adjective: zýlesa, zýlišša (blind, ina/neu):

(IIIc) EŠIS-Stems
Mestian ešis-stems are a special class of gender-agnostic adjectives. They are primarily derived and participal. An example ešis-stem adjective: galéšis, galemmášša (incoherent, ign/aqu/ani/ina/neu):

Verbs
As the concept of clear-cut morphemes is a bit more loose in Mestian verbs, this section is divided into a theoretical and practical section. This is due to the complex morphophonological alternations that exist between morphemes, that's evolved out of a highly agglutinative and isolating system.

Theory
Mestian verbs are quite complex: they inflect for person of multiple arguments, aspect, telicity, tense and mood. They can take on multiple sequential clitics that further narrow their meaning and extract diverse additional semantic information.

Verb Shapes
All Mestian verbs follow a basic shape pattern that determines the order of their morphemes. A full Mestian verb consists of the following components:

These are, in order:
 *  -- preverb; usually prefix or postposition
 *  -- Slot A, one of the two slots for personal agreement
 *  -- the root of the verb
 *  -- the verb's morphosemantic characteristic suffix; can contain more than one such suffix
 *  -- Slot B, the other of the two slots for personal agreement
 *  -- the inflection suffix slot

Person Agreement
The verbs can take four different kinds of personal agreement affixes, grouped into two groups:
 * Primary
 * Nominative agreement affixes
 * Secondary
 * Accusative agreement affixes
 * Partitive agreement affixes
 * Oblique agreement affixes

The positioning of the affixes is complex: there is a strong tendency for the primary affix to be in Slot B, while the secondary affixes are more closely bound to Slot A. Their position is mostly conditioned by grammatical voice.

Both the primary and secondary affixes have four persons and three numbers. These are:
 * Persons
 * First person (speaker)
 * Second person (addressee)
 * Third person proximate (prominent)
 * Third person obviate (sidelined)
 * Numbers
 * Singluar
 * Dual
 * Plural

Mood and Modality
Mestian modality is primarily dependent on modal affixes and mood. It makes a distinction between three classes of moods:
 * 1) Realis moods
 * 2) *Indicative
 * 3) *Gnomic
 * 4) Irrealis moods
 * 5) *Subjunctive
 * 6) *Volitive
 * 7) *Potential
 * 8) *Interrogative
 * 9) Deontic moods
 * 10) *Imperative
 * 11) *Hortative
 * 12) *Permissive

Mood, polarity and tense are inflected together with the same affix. Some moods have different behaviours in the negative: these are primarily the prohibitive (from the permissive) and imprecative (from the hortative); they have irregular semantics but regular morphology. Not all mood-tense combinations exist.

Aspect and Telicity
Aspect and telicity are closely connected in Mestian. The combinations of telicity and aspect are:
 * 1) Telic
 * 2) *Inchoative
 * 3) *Momentane
 * 4) *Delimitative
 * 5) *Terminative
 * 6) *Iterative
 * 7) Atelic
 * 8) *Defective
 * 9) *Momentane
 * 10) *Progressive
 * 11) *Frequentative

The aspect combinations line up fairly evenly:

Aspects are either an inherent property of the root or are marked in the morphosemantic suffix slot.

Voice
Mestian verbs can have one of four (two simple and two 'augmented') voices: The applicative is extremely limited and generally no longer productive in the language; it exists in semantically shifted applicatives and similar fosilised constructions. The active and passive are generally indicated by order of person affixes, while the mediopassive requires a dummy pronoun (remnant of older reflexive construction); applicatives used to be formed by infixation and reordering of person affixes, and these processes have by large been thoroughly irregularised.
 * Simple
 * Active
 * Passive
 * Augmented
 * Applicative (†)
 * Mediopassive

Tense
The Mestian tense system is deceptively simple: there are only three tense-like forms extant in the language, alongside a few fossilised remnants:
 * Present
 * Nonpresent
 * Timeless
 * Telic
 * Present Telic (†)
 * Future Telic (†)
 * Future (†)

The telic tenses remain in a few verbs and are generally unproductive. The future is mostly preserved in hortatives.

The present and nonpresent are the primary temporal distinctions in the language; the nonpresent is the result of the merger of the future and an unattested past tense. The timeless is a tensified aorist that doesn't encode temporal information as such.

The distribution of tenses against aspects is somewhat skewed -- the atelic aspects occur more frequently in the present, whereas telic ones are more frequent timeless or nonpresent -- and some tenses are found only in some aspects -- telic tenses can only have telic aspects, and the future can have only atelic aspects.

Tenses are usually marked either by way of a morphosemantic suffix, or in the inflection slot.

The distribution of tenses is closely linked to moods:

Praxis
The practical section of Mestian conjugation deals with the morphology of verbal inflectional morphemes.

Tense and Aspect
The tense-aspect-mood (TAM) system of Mestian is partially derivational: there exist various fusions -- as well as various degrees of fusion -- of the three categories marked by one or more affixes that can sometimes carry semantic baggage.

The tenses are marked using a huge variety of suffixes and alternations: the timeless has four classes, the nonpresent has six and the present has eighteen different unique classes. The present classes are:


 * Type 1: Morphosemantic presents (seven classes)
 * Type 1a: -n- nasal root-infixed presents
 * Type 1bp: Primary -ešte- augmentations
 * Type 1cp: Primary -ŋ- stems
 * Type 1dp: Primary -ŋ- augmentations
 * Type 1bs: Secondary -ešte- augmentations (ablauting)
 * Type 1cs: Secondary -ŋ- stems (ablauting)
 * Type 1ds: Secondary -ŋ- augmentations (ablauting)
 * Type 1e: -lh- extensions
 * Type 1f: -r/rh- extensions
 * Type 1gp: -w- ~ -j- alternating extensions
 * Type 1gs: -u- ~ -i- ablauting
 * Type 2: Doubly-kinetic presents (three classes)
 * Type 2an: -n- nasal root-infixed, nasal-prefixed presents
 * Type 2ak (†): -n- nasal root-infixed, velar-prefixed presents
 * Type 2as (†): -n- nasal root-infixed, s- prefixed fientives
 * Type 2bs: -(a)gha- augmented, s- prefixed denominal presents
 * Type 2bw: -(a)gha- augmented, w- prefixed detelic presents
 * Type 3: Doubly-static presents (five classes)
 * Type 3ap: Primary -s- root-infixed, -aw- ~ -aj- augmentations
 * Type 3as (†): Secondary -s- root-infixed, -aw- ~ -aj- stems (ablauting, irregular)
 * Type 3b (†): -esse-, -gh- suffixed augmentations (ablauting, irregular)
 * Type 3c (†): -(a)šše-, -gh- suffixed augmentations (ablauting, irregular; suppletive)
 * Type 3d: -ŋ- augmented, -gh- suffixed causatives (ablauting; semantically shifted)
 * Type 3e: -zd- augmented, -d- suffixed perfectives
 * Type 4: circumfixed (two classes)
 * Type 4a (†): -s- prefixed, -gh- suffixed
 * Type 4b: -s- prefixed, -d- suffixed
 * Type 5: irregular
 * - ablauting raw stems
 * - primary raw stems
 * - suppletive
 * - irregular raw
 * - irregular pseudo-augmented

The nonpresent classes are:
 * Type 6: Morphosemantic nonpresents (five classes)
 * Type 6a: Primary -axi- ~ -axu- stems
 * Type 6b: Primary -rh- stems
 * Type 6cp: Primary -gha- stems
 * Type 6cs: Secondary -gha- stems (ablauting)
 * Type 6d (†): Primary -zda- stems
 * Type 6e (†): Primary -zda- augmentations
 * Type 7: irregular
 * - ablauting raw stems
 * - primary raw stems
 * - suppletive
 * - irregular raw

The timeless classes are:
 * Type 8 (four classes)
 * Type 8a: Primary -s- augmentations
 * Type 8b (†): Primary -žd- stems
 * Type 8c (†): Primary -žd- augmentations
 * Type 8d (†): Primary -žd- extensions

The distinguishing feature forms of these classes are:
 * infixed
 * augmented
 * stem
 * extended
 * prefixed
 * suffixed
 * primary vs. secondary

Infixations are single consonants inserted into a verb's root to provide inflectional information. Stem additions are morphemes added directly to roots to form stems onto which morphosemantic suffixes are added. Augmentations are inserted at the end of the morphosemantic slot. Extensions attach to either stem additions or augmentations, both already previously extant as active parts of the inflectional system. Prefixes and suffixes are, respectively, in the prefix and suffix slots; temporal prefixes are usually closer to the root than particles or postpositions, whereas suffixes are usually the first in the suffix slot. The difference between primary and secondary inflecions is that secondary inflection involves ablaut as part of its inflectional process; this ablaut is irregular and varies from verb to verb.

Verbs are (primarily) arbitrarily sorted into classes; verbs can be in several classes at the same time, with semantics that may or may not have since shifted.

The aspect inflection morphemes have four different shapes, based on whether they are surrounded by consonants or vowels:

Aspect affixes are attached to the verb, with a general tendency to attach to the telic momentane and atelic progressive when deriving/changing aspect. Many verbs have differing telic and atelic stems that may diverge quite wildly from one another. Additional affixes exist, but they usually carry semantic, derivational information along their aspectual information -- the ones in the table are primarily inflectional.

Person and Voice
Mestian is pecular in that it inflects for voice using the order of affixes for person marking in the verb. Thus, person marking and voice are intrinsically linked. The distinction between the active and passive is determined by the ordering of affixes into the two slots:

Here the red letter represents the primary affix and the black letter the secondary affixes. Applicatives attach a <-žde-> as augmentations onto the moprhosemantic slot, after the last morpheme (including temporal augmentations) and use the same affix order as passives. Mediopassives use the order of indicatives, whilst requiring the reflexive pronoun <ãmaj> to inflect and come after the verb.

Mestian verbs agree with the nominative, accusative, partitive and oblique arguments with unique affixes, in turn grouped into primary (nominative) and secondary (accusative, partitive, oblique). Verbs can take at most three at a time; they cannot have both the accusative and partitive at the same time. The secondary affixes, when occuring together, can fuse into one unseparable and partially opaque hyper-affix, with a less frequent separate form. The person affixes have four different shapes, based on whether they are surrounded by consonants or vowels.

To improve legibility and reduce table size and number, the following matrix is used:

Mood and Polarity
Grammatical mood in Mestian has a simplistic realisation: it is plainly marked with an inflectional suffix at the end of the verb. The modal suffix is conflated with polarity.

Mestian moods are divided into three morphophonological classes:
 * D-Moods
 * Indicative
 * Gnomic
 * Potential
 * Interrogative
 * S-Moods
 * Subjunctive
 * Permissive
 * K-Moods
 * Volitive
 * Imperative
 * Optative

The mood affixes have two forms: a postconsonantal and postvocalic form:

Syntax
Mestian is very strongly a head-initial language, with the natural word order varying between VSO and VOS. Verbs always come first in verb phrases, while noun phrases can have their heads be in almost any position amongst their modifiers, with a strong preference towards being initial. The model of syntax used for constructing Mesian syntax is a form of constituency grammar with elements of a constructional grammatical framework that assumes a sharp, but not binary, gradient between the lexicon and syntax.

Noun Phrases
Mestian noun phrases (NPs) are patterned after a typical lowlands phrase model: despite its relative freedom in word order, Mestian has a very strong preference for head-initial NPs. Mestian NPs are made up of a head and a sequence of modifiers. The head of a NP can be a noun, a nominalised adjective, a nominalised number, a third person pronoun, a RC or an NP, and modifiers can be adjectives, NPs in possessive cases, possessive pronouns, numbers or AdPs.

The sequence of modifiers is built of blocks of consistent sequences that can be swapped around one for another but cannot have the order of their contents altered. The blocks are:
 * shape-size-colour
 * moral.opinion-quality.opinion-age
 * ethnicity/nationality-material
 * number
 * unclassified adjective types

Possessive NPs and pronouns always come directly after the head, regardless of where the other blocks are located. Blocks can either all follow or precede the head. The general tendency is for numbers to come the farthest from the head: if the NP is head-initial, numbers will tend to be the last block in the sequence; if it is head-final, they will tend to be the first block in the sequence.

Verb Phrases
Mestian verb phrases (VPs) are built fairly simply: they are made up of a head in the form of a verb, and modifiers in the form of AbPs. VPs are divided into finite and non-finite VPs; finite VPs are fairly rigidly head-initial, and non-finite VPs can be either head-initial, head-medial or head-final, with a tendency towards a head-initial configuration.

Adjective Phrases
(AdPs)

Comparative Adjective Constructions
Constructions that involve adjective comparison are a special form of adjective phrase and are generally grouped into a concept termed comparative adjective phrase (CAdP). A CAdP is made up of one comparative adjective and its complement; the complement may either be a noun phrase or nominal relative clause on one hand — making it a direct complement that takes the dative case on its head and all modifiers that refer to the head — or an adverbial clause on the other — making it an indirect complement.

An indirect comparative complement may be introduced only with the interrogative adverbs sîxis (how much) and sîkkari (to what extent). This adverb is then always preceded by the conjunction īka (than).

Relative Clauses
Mestian relative clauses (RCs) are a subtype of subordinate clause that functions as a NP head. They themselves are made of two parts: a relative head in the form of an NP, and a modifier clause. The modifier clause (MC) is thus said to be headed by the relative head (RH). MCs come in two different versions: a normal nominal MC (NMC), and a possessive MC (PMC).

NMCs, directly referring to the RH, are introduced by the relative pronoun, ûttaj, that inflects corresponding to its role in the sentence; as MCs are very starchly head-initial, this then forces prepositions to come right after the pronoun and act as postpositions.

PMCs, referring to possessions of the RH, are instead formed as transformations of possessive NP head-modifier pairs where the modifier becomes a RH and the head then becomes a part of the MC. They are introduced with the relative pronoun ûttaj that instead agrees in case and number with the head of the original NP (aka referent), now transformed into an NP constituent of the clause. This referent is then linked to the MC with a possessive pronoun that agrees in gender and number with the RH and takes the corresponding possessive case.

Verbs in RCs undergo deranking in certain contexts: if they refer to a possible or hypothetical situation, or carry adverbs of uncertainty, renarration, infirm evidentality and their like, they are transformed by the following mapping function:

Tense-mood combinations that do not map cleanly (primarily those involving irregular remnant tenses) either switch tense to a cleaner and more regular form that is then able to undergo such a reductive transformation, or force a suppletive stem replacement. Main clauses that feature imperatives, hortatives and interrogatives cannot undergo relativisation.

Vocabulary
Mestian/Lexicon Mestian/Texts