- 1 Phonetics
- 2 Morphonology
- 3 Morphology
- 3.1 Nouns
- 3.1.1 Declensions
- 3.1.2 Spatial cases
- 3.1.3 Substantivized genitive
- 3.1.4 Pronouns
- 3.1.5 Quantity words
- 3.1.6 Declinable conjunctions
- 3.2 Adjectives and adverbs
- 3.3 Adverbial particles
- 3.4 Verbs
- 3.1 Nouns
- 4 Syntax
- 4.1 Nominal groups
- 4.2 Main clause
- 4.3 Negation
- 4.4 Adpositions
- 4.5 Reported speech
- 5 Lexical derivation
|Stops / affricates||p||t||ts [t͜s]||k|
Affanonic distiguishes the following monophthongs: a o u e i ø y (IPA values), and the following diphthongs: ai ei oi au.
Every consonant phoneme can be used as a single-consonant onset. Vowel onsets are allowed too, although common only word-initially. The consonants j- and h- do not occur before the narrow front vowels i and y, while w- cannot stand before the narrow labialized vowels u and y.
Besides, a syllable can begin in a cluster, Cj or Cw. However, h r l w (and j) are not allowed before j, while p f m s n l j (and w) are not tolerated before w; intervocalic combinations like -rj- or -lw- that do occur word-internally have a syllable boundary between the two consonants, and thus do not form a syllable onset. Therefore the complete list of onset clusters looks like the following: pj- tj- tsj- kj- fj- sj- mj- nj-, tw- tsw- kw- sw- rw-.
The following single-consonant codas are allowed: r l n f s and gemination of the onset stop/affricate (clustered or not) of the next syllable. None of these can follow a diphthong. The coda -n remains [n] only before n- t- ts-; before m- p- it assimilates to [m] (which is reflected in spelling: -mm-, -mp-), whereas word-finally (except when followed by a vowel-initial word without a pause) and before all the other consonants it becomes a velar [ŋ] (still written n).
The coda -n often assimilates to a following l- (to produce a geminated -ll-), but this alternation is not automatic anymore and therefore is better considered part of morphonology rather than of synchronic phonotactics.
The position of stress is not contrastive.
Disyllabic words always have the stress on their first syllable. In words of more than two syllables, the stress falls on the penult if it is heavy (i. e. uncludes either a diphthong or a coda consonant) and on the antepenult otherwise.
In declarative utterances pronounced with neutral intonation, the stressed syllale is characterized by its falling tone; the pretonic syllables tend to form a high-pitch platform, while the tone of the posttonic ones is low. The dynamic aspects of stress are not prominent in Affanonic.
An important process often occurring in inflectional forms is historical palatalization: *lj → j, *rj → h. When the the change *lj → j produces a consonant combination that is allowed as an onset cluster, a resyllabification takes place, which is sometimes important for stress position. Thus, a form like rarkenjur 'brothers' is syllabified rar.ke.njur and therefore has the stress on its first syllable, despite being historically *rar.ken.ljur.
[more stuff to be added later]
The morphology of Affanonic is peculiar in combining a sophisticated inflection in nouns with near-total lack of inflection in verbs (with a few exceptions pertaining to auxiliaries).
Affanonic is traditionally described as having 31 cases. Six of these (nominative, accusative, genitive, dative, benefactive, and instrumental) are called “primary cases”; the rest are referred to as “secondary cases”. Secondary cases include 24 “spatial cases”; besides, one more form traditionally viewed as a secondary case, the so-called “substantivized genitive”, is properly a derived noun which itself inflects for case.
The forms of secondary cases are derived from certain primary cases in a rather uniform manner. For this reason, the overview of declensions below will confine to primary cases, while the formation of secondary cases will be explained a bit later.
Affanonic distiguishes between two numbers, singular and plural. The use of plural in countable nouns is rather consequential. Uncountable nouns are formally singular, virtually without any exceptions (i. e. pluralia tantum do not exist as a category). A so-called “counting form” of countable nouns which is derived from the plural stem is used with words that denote quantities (including numerals).
Every case ending, as well as the plural stem marker, have two forms depending on the so-called stem class of the noun. The stem classes are purely formal and are conventionally named simply “class 1” and “class 2”. Within each stem class, there are further subgroupings determined by the final sounds of the stem; the most basic divide is between “vocalic declensions” (comprising the nouns whose stems end in a vowel) and “consonant declensions” (which include stems ending in a consonant).
With most nouns, the declension type can be predicted from the nom. sg. form (more specifically, its final sounds).
Class 1 declensions
The prototypical endings used with class 1 stems are: -i or zero, nominative; -ys or -wis, accusative; -al, genitive; -njun, dative; -lu, benefactive; -ma, instrumental; -jur, plural marker. Various alternations on morpheme boundaries make the formation of certain forms somewhat obscure.
Class 1 vocalic declensions
|Declension type||-ai: somai 'stone'||-oi: appoi 'cloud'||-ei: kei 'mountain'||-u: kommu 'husband'||-i: liffi 'bird'||-y: fomy 'fish'||-au: nau 'turn, iteration'|
The nominative ending used to be uniformly -i, but only -ai, -ei and -oi survive as diphthongs synchronically.
The genitive somol is historically soma-al; the appearance of -f- and -s- in genitives looks synchronically unmotivated, but actually is predictable from the quality of the final sound of the stem. Cf. similar processes in class 2 declensions, as well as in secondary case formation.
The contraction in the accusative kommus is phonetically regular and used to be active in other declension types as well, but ultimately was mostly eliminated by analogy (except in nouns in -u and a few fossilized forms).
Class 1 consonant declensions
|Declension type||-in, -yn, -øn: soffjunsin 'louse'||-on: non 'person'||-un: mumun 'seat'||-jun: mjun 'rain', hun 'heart'||-en: rarken 'brother'||-an: tswan 'bitter vetch'||-l: tjupol 'beech tree'||-r: soror 'rat'||-af, -of, -uf, -if, -ef, -yf (i. e. -Vf except -øf): rwetsif 'fox'||-es, -is, -ys: tsarkes 'battle'||-os, -us, -øs: tswistjos 'belly'|
|- nominative||soffjunsin||non||mumun||mjun, hun||rarken||tswan||tjupol||soror||rwetsif||tsarkes||tswistjos|
|- accusative||soffjunsinys||namys||mumonys||mimys, rimys||rarkanys||tswonys||tjupolfys||sororys||rwetsinfys||tsarkesys||tswistjonsys|
|- genitive||soffjunsinal||nanfal||mumonal||minfal, rinfal||rarkanal||tswonal||tjupolral||sororal||rwetsinfal||tsarkesal||tswistjonsal|
|- dative||soffjunsinnjun||nonnjun||mumunnjun||mjunnjun, hunnjun||rarkennjun||tswannjun||tjupolhun||sorornjun||rwetsifnjun||tsarkesnjun||tswistjosnjun|
|- benefactive||soffjunsillu||nollu||mumullu||mjullu, hullu||rarkellu||tswallu||tjupolru||sororu||rwetsiflu||tsarkeslu||tswistjoslu|
|- instrumental||soffjunsimma||nomma||mumumma||mjumma, humma||rarkemma||tswamma||tjupolfa||sororma||rwetsinfa||tsarkesma||tswistjosma|
|Nom. pl.||soffjunsinjur||nonjur||mumunjur||mjunjur, hunjur||rarkenjur||tswanjur||tjupolhur||sorohur||rwetsifjur||tsarkesjur||tswistjosjur|
The assimilation fm → (f)f is a common process also found in other formations in the language.
The historical stem-final -m and -n differently affect the quality of the preceding vowels where they are in coda position. In the type hun, the latter process produced a combination -ju- which complicated things further by causing a historical palatalization of the preceding stem-internal -l- and -r-.
The types rwetsif and tswistjos include stems which historically ended in -nf- and -ns-; these clusters simplified when in coda position.
The only peculiarity of the type soror is the historical assimilation -rl- → -r- in benefactive and -rlj- → -rj- ( → h) in the pl. forms.
The type tjupol is the most complicated: these stems end historically in -lr-, with various assimilations and cluster simplifications.
Class 2 declensions
The prototypical endings in this class are: -e or zero, nominative; -øs or -wes, accusative; -ja, genitive; -njøn, dative; -lø, benefactive; -mja, instrumental; -jør-, plural marker.
Class 2 vocalic declensions
|Declension type||-a: øfja 'horse'||-e: rele 'lizard'||-o: jo 'ice'||-ø: fø 'wind'|
All the alternations involved are the same as with class 1 stems.
Note that the type rele includes only such stems with nominative in -e that cannot be interpreted as one of the class 2 consonant declensions described below.
Class 2 consonant declensions
|Declension type||-øne: fømøne 'death'||-ene: pintene 'garden'||-ane: tejane 'sky'||-ame: janjame 'leaf'||-ome: tjome 'juice'||-lre: punfølre 'wolf'||-øf: røf 'mouth'||-as: jas 'snake'|
Most alternations found in these types are the same as with class 1 declensions; only the type jas is peculiar.
Plural declension and the counting form
|Declension type||class 1: somajur 'stones'||class 2: øfjajøre 'horses'|
The counting form of class 1 nouns is identical with nom. pl., but in class 2 it is a separate form built by dropping the nominative ending.
A few nouns are class 1 stems in sg. but class 2 in pl.; this shift is accompanied by altering the stem vowel. Examples include moi 'eye', woi 'mushroom', roi 'horn', and koi 'bone', whose plurals are møjøre, wøjøre, røjøre, and køjøre, respectively.
Many irregularities characterize the declension of pronouns, as described in the respective sections below.
Inventory and meanings
The 24 spatial cases can be presented as combinations of eight “localizations” with three “orientations”.
Localizations describe the position of a reference point relative to the object denoted by the noun being inflected:
- adessive, 'at';
- inessive, 'in';
- superessive, 'on' (usually, 'in contact with the upper surface of');
- apudessive, 'near, close to';
- subessive, 'under, below (with or without contact)';
- postessive, 'behind or after';
- antessive, 'before or in front of';
- extraessive, 'outside'.
Orientations specify the type of movement or static position relative to the point referred to by the localization:
- locative: '(resting) at the point specified by localization';
- lative: '(moving or directed) towards the point specified by localization';
- ablative: '(moving or directed) away from the point specified by localization'.
With the forms of the inessive localization and the noun mjunki 'water' taken as an example:
- inessive-locative: mjunkiwiry 'in the water';
- inessive-lative: mjunkinimy 'into the water';
- inessive-ablative: mjunkimoi 'out of the water'.
As can be seen from the above examples, the name of each spatial case consists of the name of the localization followed by the name of the orientation (e. g. “inessive-lative”, “subessive-ablative”, etc.). However, the name of the locative orientation can be omitted where no ambiguity will arise, so e. g. “superessive-locative (case)” can be referred to as simply “superessive (case)”.
Besides referring to spatial relations proper, the spatial cases have quite obvious temporary usages (especially common with gerunds). Thus, the locative cases can convey meanings like 'at', 'before', 'after' (a specific moment), 'within' (a timespan); the ablative cases add the meaning of 'since', and the lative cases, of 'until'.
In addition, individual spatial cases have various metaphoric usages, including grammatical ones.
As already mentioned, the forms of spatial cases are based on certain primary cases. Specifically, the locative cases are always derived from the accusative; the lative cases, from the dative; and the ablative cases, from the instrumental.
In groups of words that must agree in case, normally only the last one is inflected for spatial case, while all the others are put in the respective primary case (accusative, dative or instrumental). In particular, declinable conjunctions and equative markers have no forms of spatial cases.
The formation of spatial cases can be conveniently presented as replacement of parts of the endings of the respective primary cases by the endings of spatial cases. The endings of spatial cases used with class 1 and class 2 stems are different. In the tables below, the formation of spatial cases is illustrated with the words mjunki 'water' (class 1) and fjas 'sand' (class 2).
Spatial cases of class 1 nouns
|Primary case||Accusative: mjunkiwis||Dative: mjunkinjun||Instrumental: mjunkima|
|Replaced part of primary case ending||-s||-njun, -(*l)-jun, -hun||-a|
|adessive||-ras: mjunkiwiras||-nimas (-limas, -rimas): mjunkinimas||-os: mjunkimos|
|inessive||-ry: mjunkiwiry||-nimy (-limy, -rimy): mjunkinimy||-oi: mjunkimoi|
|superessive||-run: mjunkiwirun||-nimun (-limun, -rimun): mjunkinimun||-on: mjunkimon|
|apudessive||-skjummo: mjunkiwiskjummo||-njunkjummo (-junkjummo, -hunkjummo): mjunkinjunkjummo||-akjummo: mjunkimakjummo|
|subessive||-spis: mjunkiwispis||-njumpis (-jumpis, -humpis): mjunkinjumpis||-affis: mjunkimaffis|
|postessive||-raffa: mjunkiwiraffa||-njunfaffa (-junfaffa, -hunfaffa): mjunkinjunfaffa||-offa: mjunkimoffa|
|antessive||-ra: mjunkiwira||-njunfa (-junfa, -hunfa): mjunkinjunfa||-o: mjunkimo|
|extraessive||-hu: mjunkiwihu||-nimju (-limju, -rimju): mjunkinimju||-aju: mjunkimaju|
Most alternations found in spatial case formation have already occurred in primary cases; in particular, those affecting the dative -(n)jun are of the same nature as those found in the declension of hun 'heart'.
Some new ones include: -s (in codas) :: -r- (intervocalic) – in all cases derived from the accusative; -p- (after consonants) :: -ff- (intervocalic) – in the subessive series.
Spatial cases of class 2 nouns
|Primary case||Accusative: fjaføs||Dative: fjojøn||Instrumental: fjaffja|
|Replaced part of primary case ending||-s||-njøn, -(*l)-jøn, -høn||-a|
|adessive||-has: fjaføhas||-nemjas (-lemjas, -remjas): fjolemjas||-os: fjaffjos|
|inessive||-rø: fjaførø||-nemø (-lemø, -remø): fjolemø||-ø: fjaffjø|
|superessive||-røn: fjaførøn||-nemøn (-lemøn, -remøn): fjolemøn||-øn: fjaffjøn|
|apudessive||-skjømmø: fjaføskjømmø||-njønkjømmø (-jønkjømmø, -hønkjømmø): fjojønkjømmø||-akjømmø: fjaffjakjømmø|
|subessive||-spes: fjaføspes||-njømpes (-jømpes, -hømpes): fjojømpes||-affes: fjaffjaffes|
|postessive||-haffja: fjaføhaffja||-njønfjaffja (-jønfjaffja, -hønfjaffja): fjojønfjaffja||-offja: fjaffjoffja|
|antessive||-ha: fjaføha||-njønfja (-jønfja, -hønfja): fjojønfja||-o: fjaffjo|
|extraessive||-hø: fjaføhø||-nemjø (-lemjø, -remjø): fjolemjø||-ajø: fjaffjajø|
All the alternations are already familiar.
|Stem class||class 1: kommufalli 'what belongs to (the) husband'||class 2: føffesjai 'what belongs to (the) soul'|
|Cases (sg. or pl.):|
|Substantivized genitive (2nd)||kommufalsalli||føffesjasjai|
The forms of substantivized genitives make no distinction between sg. and pl. of the possessed item. However, all the secondary cases are formed freely, by regular substitution of case endings. Curiously, substantivized genitives of substantivized genitives are also used, as e. g. in the following example:
E₁ tassijural₂ munnai₃ ristiwiras₄, fy₅ tatsanalsalli₆ astiwiras₇ nini₈. 'The₀ clothes₃ of₍₂₎ my₁ children₂ are₈ here₄, and₅ those₆ of₍₆₎ (my₀) sister's₆ children₀ there₇.'
(Tatsanalsalli being the nom. of the subst. gen. of tatsanalli, which is itself the subst. gen. of tatsen 'sister'.)
The substantivized genitives are sometimes lexicalized and then moved to the regular substantive declension types, with regular pl. forms. With cl. 1 stems, such derivates are invariably the nom. forms reinterpreted as regular nouns of the liffi type. With cl. 2 stems, there are two options: the derivate inherits either the oblique stem (in -ja) and declines according to the øfja type, or the nominative (in -jai), and then is reinterpreted as class 1 (!) stem of the somai type.
[Details of case usage remain to be added]
Personal and demonstrative pronouns
The personal pronouns of 1st and 2nd persons display a number of peculiarities, compared to the regular nouns, in both case forms and the plural stems.
|'I'||'you' (sg.)||'we'||'you' (pl.)|
The personal pronouns of the 3rd person are essentially demonstratives. The forms of ri refer to objects that are close to either the speaker or the listener; the objects that are at a distance from both, or are invisible to the speaker, are denoted with forms of ai. These usages of the two pronouns are called “deictic”; in non-deictic use, ai is the default anaphoric pronoun (i. e. referring to something already mentioned by one of the interlocutors), while ri is mainly used cataphorically (that is, to denote some object the speaker is going to introduce).
Note that plural forms of these are built like with 1st and 2nd person pronouns only when they refer to animate beings (not necessarily sentient); otherwise their plurals are built like those of the regular nouns.
|'this one; he/she/it'||'that one; he/she/it'||'these, they' (animate)||'these, they' (inanimate)||'those, they' (animate)||'those, they' (inanimate)|
When used as attributes of other nouns, the demonstratives require a construction which is essentially the same as with appositions (see the sections on the equative marker and on nominal groups below). The more common position for the pronoun is after the noun; in this case, there are some peculiarities not found with appositional nouns, incliding free use of genitives and contractions with the equative marker. When placed before the noun, the demonstratives require the usual appositive construction (with no contractions and no genitives available); in this case they are semantically somewhat out-of-focus and may resemble definite articles, although are thus used infrequently.
|'this (these)', postpositive||'that (those)', postpositive||'this (these)', prepositive||'that (those)', prepositive|
|Nominative||li ri||j-ai||ri li||a li|
|Accusative||pys hus||pys søs||hus pys||øs pys|
|Dative||njun rinjun||njun sanjun||rinjun njun||anjun njun|
|Benefactive||lu rilu||w-alu||rilu lu||alu lu|
|Instrumental||ma rima||m-oma||rima ma||ama ma|
It is interesting to note that prepositive and pospositive demonstratives sometimes occur simultaneously with one noun; in such cases, at least one of the demonstratives is interpreted as non-deictic:
Ai₁ li₂ e₃ rarken₄ li₅ ri₆. 'The said brother of mine whom you see here' (lit. 'That₁ one₁ who₂ is₂ my₃ brother₄ who₅ is₅ this₆').
Other demonstratives that are mostly used independently include risti 'this place, here', asti 'that place, that direction, there', rinau 'this time, now', and anau 'that time, then' (note that risti and rinau, like ri, are often used cataphorically and then cannot be translated as 'here' and 'now'). They are declined like regular class 1 nouns and mostly used in spatial cases, although primary cases are also common, as e. g. in Ristiwis₁ ei₂ tesi₃ 'I₂ like₃ this₁ place₁' (or 'I like it here'). They can form plurals to refer to multiple locations or instances, as well as to broader areas or more prolonged timespans.
The main interrogative pronouns include: y 'what? which one?'; yrky 'who? which one?'; riffikky 'what? who? what kind of object(s)?'; ysti 'where? which place? which direction?'; riffitti 'where? what kind of place?'; ynau 'when? which time?'; and riffinau 'when? what kind of circumstances?'.
The pronouns y and yrky imply that the object of inquiry belongs to a set which is already known or obvious from the situation; yrky implies an animate object (but not necessarily sentient). In contrast, riffikky implies that the speaker knows nothing about the object, and may need its description rather than identification; this pronoun is not sensitive to animacy. The semantic difference in the pairs ysti vs. riffitti and ynau vs. riffinau is partly similar with that between y / yrky and riffikky.
The pronouns ysti, riffitti, ynau, and riffinau are mostly used in spatial cases, as e. g. in Ystiwiras₁ ai₂ nini₃? 'Where₁ is₃ he/she/it₂?', although primary cases are also common, as e. g. in Ystiwis₁ toi₂ tesi₃? 'Which₁ place₁ do₃ you₂ prefer₃?' and in similar contexts.
Of the interrogative pronouns, only y, yrky and riffikky display some peculiarities in declension (namely, in acc. sg.); the rest belong to the standard class 1 declension types.
The plurals of y and yrky are used rarely; usually they imply not only that the objects are already known to be more than one, but also that the speaker wants them to be enumerated; the plural of riffikky does not exist.
In contrast, the plurals of ysti, riffitti, ynau and riffinau are rather common; ystijur and riffittijur imply that there are several relevant places, while ynaujur and riffinaujur often refer to prolonged periods or multiple instances rather than to single moments.
The reflexive pronouns are based on the auxiliary noun ry 'self' which normally is not used without a genitive of a pronoun or noun that specifies it. A genitive of a personal pronoun and ry are written as one word and pronounced accordingly: ery 'myself', tory 'yourself', ekwiry, tokkwiry, riry, ary, rikwiry, rijuralry, akwiry, ajuralry. The genitives of other words remain separate, as e. g. in e₁ mamol₂ ry₃ 'my mother herself' (lit. 'my₁ mother's₂ self₃').
The word ry has no plural and is declined like the interrogatives y, yrky, riffikky.
The genitive of a reflexive pronoun can often be translated with 'own', as in erysal₁ fei₂ 'my₁ own₁ house₂', e₁ kispjufal₂ rysal₃ øfja₄ 'my₁ wife's₂ own₃ horse₄'.
The equative marker
“The equative marker” is the traditional term for the auxiliary word that links a noun (or pronoun) with an apposition, as e. g. in to₁ mamai₂ li₃ e₄ tsuma₅ kispju₆ 'your₁ mother₂ who₃ is₃ my₄ future₅ wife₆' or elu₁ lu₂ to₃ tatsellu₄ 'for₍₁₋₂₋₄₎ me₁, (as₂) your₃ sister₄'. The head noun, the equative marker and the apposition must agree in (primary) case, and therefore the equative marker is inflected for case; however, it is insensitive to number distinctions and has no secondary cases. Besides, it has no forms of genitive (primary or substantivized), as for appositions to nouns in genitive there is a different construction, where the head genitive is replaced with a substantivized genitive: to₁ mamollu₂ lu₃ e₄ tsuma₅ kispjufal₆ felu₇ 'for the house of your mother who is my future wife' (lit. 'for₍₂₋₃₋₇₎ your₁ mother's₂, which₃ is₃ my₄ future₅ wife's₆ house₇').
This pronoun plays a very important part in Affanonic syntax. Prototypically, it indicates that its head predicate has the same topic as a preceding predicate, but in a different role. Its other uses are described in the section on syntax below.
The role-switching marker does not differentiate singular from plural, but has all cases including secondary.
The relative pronoun fylli is used in relative clauses when they follow their head nouns. It inflects for case which denotes its role in the relative clause (rather than in the main clause), e. g. nontalu₁ fynnjun₂ ei₃ fati₄ 'for₍₁₎ the₀ man₁ to₍₂₎ whom₂ I₃ told₄ (that₀)'. There are all case forms, both primary and secondary, but no number distinction.
This category includes numerals (e. g. tsei 'three') and a few words that denote quantities of countable objects: fumøne 'many'; pimoi 'few'; kjunsøf 'several, a few, some (of)'; supu 'all (of)'. They are declined like the ordinary nouns in singular and carry no number markers of their own. The noun whose quantity is specified is put in counting form and precedes the quantity word: øfjajør₁ tsei₂ 'three₂ horses₁'.
Where nouns or pronouns linked by the conjunctions fy 'and', la 'or', or pau 'but not' stand in the same (primary) case and form the same phrase constituent, the conjunctions must agree in (primary) case with the words they link, e. g.: e₁ mamalu₂ fylu₃ elu₄ 'for₍₂₋₃₋₄₎ my₁ mother₂ and₃ me₄', e₁ mamanjun₂ fynjun₃ enjun₄ 'to₍₂₋₃₋₄₎ my₁ mother₂ and₃ me₄'.
The case forms of the three declinable conjunctions are shown in the following table.
|fy 'and'||la 'or'||pau 'but not'|
Adjectives and adverbs
These are basically one part of speech in Affanonic. Thus, fje 'beautiful' is mostly used as an adjective, but can also mean 'beatifully' when it stands before a verb. Similarly, ese 'sometimes' is mostly used as an adverb, but can also be used with a noun to convey a meaning like 'occurring sometimes'.
Both adjectives and adverbs must immediately precede the word they specify.
The words belonging to this class have no inflectional forms.
This is another class of words that often correspond to avderbs (most commonly, sentential adverbs) in other languages. However, they cannot be used as prepositive attributes of nouns and, unlike the usual adverbs, can be (and mostly are) separated from their head verb by any number of other constituents.
An example of adverbial particle is affaini 'one day (in the past)'.
It is interesting to note that spatial cases of nouns are syntactically more similar with adverbial particles than with adverbs proper. Indeed, many adverbial particles resemble case forms derived from obscure stems not used as ordinary nouns; an example is mølø 'visually, visibly, by eye, before one's eyes' which looks like a benefactive derived from some *mø (cf., however, moi 'eye' and its irregular plural møjøre 'eyes').
All the verbs are divided into two major classes: perfectives and (lexical) imperfectives.
Perfective verbs are semantically telic. In other words, they denote processes that imply a change of state or achieving a result, and therefore cannot be normally viewed as allowing for an indefinitely long continuation (i. e. as continuing after the point where the change is complete or the implied result is achieved). Examples of such verbs in Affanonic are soffoffi 'to fall asleep', mumin 'to sit down', fati 'to say, tell', føref 'to kill', and njun 'to become'.
In contrast, lexically imperfective verbs are semantically atelic, i. e. denote static conditions or cyclic processes that do not imply a specific endpoint by themselves. Examples include roffoffi 'to sleep', lilami 'to stand', piffati 'to talk, speak', and nini 'to be'.
Perfective verbs have two standard imperfective derivates that can be considered part of their paradigm: durative and resultative (jointly referred to as “secondary imperfectives”, as opposed to “lexical imperfectives”).
Duratives are formed by the auxiliary verb nyral placed after the unaltered stem of the perfective verb. A durative can have either of the following two interpretations. In one interpretation, it denotes an indefinite number of repetitive events which individually would be denoted by the original perfective; this interpretation is called “iterative”. The other interpretation implies a shift in the time scale: the process or activity that should eventually result in the change of state denoted by the original perfective is viewed as long enough to serve as the “indefinitely long” background for some other, shorter events, and thus as requiring an imperfective to describe it. This interpretation is termed “continuative”.
Resultatives are formed by the auxiliary ja put after the stem of the original perfective. They denote a continuous static condition that results from the event denoted by the perfective.
Lexical imperfectives do not have any standard perfective derivates, since the associated semantic shifts are more diverse and idiosyncratic. Therefore, all processes that derive a perfective from an imperfective are considered as belonging with lexical derivation.
The finite forms of imperfectives (both lexical and secondary) taken without any additional marking are the standard forms of the present tense. The term is purely conventional; while the default meaning of the present taken outside any specific context is indeed that of an action or condition which continues at the moment of speaking, it can also denote an action or condition that was simultaneous with, or served as the background for, another action which can relate to present, past or future. In particular, adverbs and adverbial expressions that denote time, e. g. affaini 'one day (in the past)', can freely combine with the present tense, and do not require the verb to carry any additional marking to express non-present. In other words, the characteristic of this tense is purely negative: it does not mark a continuous action or condition as preceding or following anything else.
On the contrary, the perfective verbs taken as bare stems indicate by default that the action in question has been, or had been, completed by the present moment or by some moment in the past. Therefore, their bare stem is considered a past tense which is termed “aorist”. However, if there is a direct indication to future in the same sentence, the aorist form can mean that the action will have been completed by some moment in the future.
Imperfectives can be explicitly marked for past tense by adding the postpositive auxiliary njau (which is actually a perfective verb 'to turn' in its non-auxiliary use). This combination is termed “imperfect tense”.
The explicit reference to future is made differently with perfectives and imperfectives. Imperfectives use a special postpositive tense marker, jawa. This form is referred to as “imperfective future”.
Perfectives, on the other hand, are combined with the word tsjuma which is properly an independent adjective / adverb meaning 'in the future, imminent(ly)'; therefore, it is put before the verb and is sometimes separated from the latter by another adverb. Nevertheless, such combinations with tsjuma are often treated as a special tense called “perfective future”.
The subjunctive of the perfective verbs is formed by adding the postpositive marker nifføf. This form refers to the past; however, it can combine with tsjuma to convey future.
The copula nini 'to be' has an inflectional form of subjunctive, ninifføf. This form refers to the present. The other lexical imperfectives form their present subjunctive by adding ninifføf postpositively. However, duratives do not use this form, but replace their nyral with nyralpøf; similarly, the resultatives substitute jafføf for their ja. The past subjunctive of all imperfectives (including duratives, resultatives and nini) is formed by adding a pospositive nufføf, and the future subjunctive, by a postpositive nawafføf.
Gerunds represent the only type of regular non-finite forms of Affanonic verbs. They can be considered substantivized verbs which make it possible to use a verb as an argument of another verb, or of any other word. On the other hand, a gerund can retain all the arguments of the original finite verb without altering their form.
The benefactive of gerunds has some usages similar to those of infinitive in English.
Perfective verbs form their gerunds with the postpositive auxiliary noun nau (which in its non-auxiliary use means 'turn, iteration, time').
Lexical imperfectives have gerunds formed by the postpositive jo.
The gerunds of duratives are formed by replacing the finite durative marker nyral with naujur (i. e. the plural of nau).
The gerunds of resultatives are formed by replacing the finite resultative marker ja with the auxiliary noun jai, which declines like a class 2 substantivized genitive.
The gerund forms listed above are basically tenseless; however, due to the vague temporal meanings of the Affanonic present and aorist, the above types of gerunds can be viewed as their substantivized equivalents.
The use of the gerunds equivalent to the other tenses is optional. The perfective gerunds with nau can combine with tsjuma to refer to a future action. The finite markers of all the tenses of imperfectives other than the present are historically perfective verbs, so gerunds can be derived from them by adding the postpositive nau as well.
The gerund equivalents of all forms of the subjunctive are rare, but can in principle be derived freely by adding the same nau as the most universal gerund marker.
The nominal group template can be presented as follows:
|A prepositive demonstrative||A genitive (with or without dependents) or a prepositive relative clause||An adjective (with or without dependents)||The head noun||The postpositive demonstrative (and/or other appositions)||An adverbial particle or a noun in oblique case used attributively||A postpositive relative clause|
It is very usual to construe a noun denoting a possessor not as the usual genitive but as substantivized genitive put in apposition. This option is especially common when the genitive has its own attributes, and obligatory if any of such attributes are postpositive.
All dependents of an adjective are put before it:
Mjunkima₁ njunta₂ pir₃. 'A₀ ditch₃ full₂ of₍₁₎ water₁.'
Liloi₁ mjunkima₂ njunta₃ pir₄. 'A₀ ditch₄ full₃ of₍₂₎ black₁ water₂.'
This is also true of comparative constructions:
Sajurnjun₁ tsera₂ sunol₃ nontai₄. 'A₀ man₄ stronger₃ than₂ others₁.'
An adjective cannot be separated from its head noun except by another adjective, which cannot normally have any arguments of its own.
To obviate the restrictions on word order, an adjective is often replaced with a relative clause based on the adjective and the copula verb nini 'to be':
Nontai₁ fylli₂ sajurnjun₃ tsera₄ sunol₅ nini₆. 'A₀ man₁ (who₂ is₆) stronger₅ than₄ others₃.'
Note that personal pronouns have relative freedom of taking all the types of attributes allowed with the regular nouns, e. g.:
Situn₁ ei₂ wo₃ tituli₄ njau₅. 'Stupid₁ me₂, I₂ did₍₄₋₅₎n't₃ know₄₋₅ (that₀)'.
Since in their internal structure the relative clauses of Affanonic present no peculiarities compared to main clauses, it is more convenient to consider them in the section on nominal groups.
Prepositive relative clauses
The easiest way to construe a relative clause in Affanonic is simply to put the predicate (with or without its arguments) immediately before the head noun. This option is especially common when the predicate has no dependent nouns. The head noun fills an argument valency of the predicate. However, by itself this construction provides no information on the specific argument role of the head noun, which therefore remains to be retrieved from the context. With intransitive verbs, there is a tendency to interprete the head noun as filling the subject valency, e. g.:
Tsjurin₁ nontai₂. 'The₀ man₂ who₀ came₁ (has come)'.
Cf.: Nontai₁ tsjurin₂ 'The₀ man₁ has₍₂₎ come₂'.
With transitive verbs, the head noun often fills the valency for direct object:
Føref₁ øfja₂. 'The₀ horse₂ that₀ has₍₁₎ been₍₁₎ killed₁'.
Cf.: Øfjawes₁ føref₂. '(Somebody₀) killed₂ the₀ horse₁'.
However, the head noun can in principle correspond to any other role, e. g. expressed by a spatial case:
Ei₁ jeke₂ fei₃. 'The₀ house₃ where₀ I₁ live₂'.
Cf.: Ei₁ a₂ fewiry₃ jeke₄. 'I₁ live₄ in₍₃₎ his₂ house₃'.
Some disambiguation is achieved if the relative clause has an explicit argument; then the head noun can fill only a remaining argument valency:
Øfjawes₁ føref₂ nontai₃ li₄ ri₅. 'This₄₋₅ man₃ who₀ killed₂ the₀ horse₁'.
Nontai₁ li₂ ri₃ føref₄ øfja₅. 'The₀ horse₅ killed₄ by₀ this₂₋₃ man₁'.
Cf.: Nontai₁ li₂ ri₃ øfjawes₄ føref₅. 'This₂₋₃ man₁ killed₅ the₀ horse₄'.
A more powerful means of disambiguation is provided by the use of the role-switching marker whose case indicates the role to be filled by the head noun:
Lili₁ føref₂ nontai₃. 'The₀ man₃ who₍₁₎ killed₂ (somebody₀)'.
Cf.: Nontai₁ li₂ ri₃ lili₄ føref₅. 'As₍₄₎ for₍₄₎ this₂₋₃ man₁, he₍₄₎ killed₅ (somebody₀)'.
Pipyry₁ ei₂ jeke₃ fei₄. 'The₀ house₄ in₍₁₎ which₍₁₎ I₂ live₃'.
Cf.: Fei₁ li₂ ri₃ pipyry₄ ei₅ jeke₆. 'As₍₄₎ for₍₄₎ this₂₋₃ house₁, I₅ live₆ in₍₄₎ it₍₄₎'.
However, the use of the role-switching marker may cause some problems, too. Specifically, the role-switching marker can be understood as part of the main clause rather than the relative clause. Thus, a sentence like:
Lili føref nontawis hunti.
- has two possible interpretations:
(a) Lili₁ ((føref₂ nontawis₃) hunti₄).
'(Somebody₍₁₎ already₍₁₎ mentioned₍₁₎) saw₄ the₀ man₃ who₀ had₍₂₎ been₍₂₎ killed₂'.
(b) ((Lili₁ føref₂) nontawis₃) hunti₄.
'(Somebody₀) saw₄ the₀ man₃ who₍₁₎ had₍₂₎ killed₂ (somebody₀ else₀)'.
Another problem common to all types of prepositive relative clauses is that basically they occupy the same slot in the nominal group template as genitives. Indeed, if put before the genitive, they are naturally understood as depending on the noun in genitive rather than its head:
Føref₁ nontol₂ øfja₃. 'The₀ horse₃ of₍₂₎ the₀ man₂ who₀ was₍₁₎ killed₁' (rather than **'The man's killed horse').
For this reason, prepositive relative clauses depending on a genitive's head noun are allowed to precede only the genitives of personal pronouns, e. g.:
E₁ myffawis₂ rissjun₃ a₄ miti₅. 'His₄ daughter₅ who₀ washed₃ my₁ cloak₂'.
On the other hand, if the relative clause has any explicit nominal arguments, a preceding genitive will be naturally understood as part of the relative clause:
E₁ øfjawes₂ haf₃ u₄. 'The₀ dog₄ that₀ bit₃ my₁ horse₂' (rather than **'My dog that bit the horse').
For this reason, genitives depending on the head noun can precede a relative clause only if the latter has no explicit nominal arguments, e. g.:
To₁ fje₂ tsalami₃ nyral₄ miti₅. 'Your₁ daughter₅ who₀ sings₃₋₄ beautifully₂'.
To avoid the restrictions associated with prepositive relative clauses, other types of relative clauses can be used.
Postpositive relative clauses
These are formed using the relative pronoun fylli which indicates the role of the head noun explicitly, e. g.:
E₁ rarkanal₂ kispju₃ fylli₄ e₅ myffawis₆ rissjun₇. 'My₁ brother's₂ wife₃ who₄ washed₇ my₅ cloak₆.'
To₁ kommufal₂ fei₃ fyppyry₄ toi₅ jeke₆. 'Your₁ husband's₂ house₃ in₄ which₄ you₅ live₆.'
Postpositive relative clauses often have a pronoun (especially the cataphoric ri) as their head:
Ri₁ fylli₂ ajurys₃ suffi₄ tsjurin₅ lili₆ e₇ kommu₈ nini₉ jawa₁₀. 'He₁ who₂ has₍₄₋₅₎ brought₄₋₅ them₃ will₁₀ be₉ my₇ husband₈'.
The above example illustrates another peculiarity of postpositive relative clauses: to avoid their predicate's being perceived as forming a serial construction with the predicate of the main clause, the role-switching marker (lili in this example) is inserted after the relative clause; it refers to the topic of the sentence, which may coincide or not with the head noun of the relative clause.
A restriction associated with postpositive relative clauses is that nouns in genitive cannot take any postpositive attributes, including relative clauses. This problem can be solved, as usual, by substantivizing the genitive and construing it as an apposition:
Fei₁ li₂ to₃ mitisalli₄ fylli₅ fje₆ tsalami₇ nyral₈. 'The₀ house₁ (which₂ is₂) of₍₄₎ your₃ daughter₄ who₅ sings₇₋₈ beautifully₆.'
The postpositive relative clauses can be extraposited, i. e. put after the main predicate of the sentence in which their head noun is an argument. In such cases, the head noun must be the last nominal constituent that precedes the main predicate:
E₁ rarkanal₂ kispju₃ tsjurin₄ fylli₅ e₆ myffawis₇ rissjun₈. 'My₁ brother's₂ wife₃ has₍₄₎ come₄ who₅ washed₈ my₆ cloak₇.'
Head-internal relative clauses
In this type, the head noun must either be one of the pronouns ri, risti, rinau, or have ri as its attribute. The head noun remains within the relative clause in its usual form and position, according to its syntactic role; its role in the main clause is conveyed with the role-switching marker:
Toi₁ ajurys₂ ristiwiras₃ øftsi₄ lili₅ e₆ fei₇ nini₈ jawa₉ 'The₀ place₃ where₃ you₁ found₄ them₂ will₉ be₈ my₆ homeplace₇'.
Cf. Toi₁ ajurys₂ ristiwiras₃ øftsi₄. 'You₁ found₄ them₂ here₃'.
This type of relative clauses is widely used in Affanonic. However, it is normally required that the relative clause should constitute the topic part of the sentence.
Head-internal clauses are often used to single out a nominal group constituting the contrastive (or otherwise emphasized) rheme of the sentence. This construction is similar in structure with the English "It is X that Y" (only with the order of components reversed):
Hus₁ e₂ tatsen₃ rissjun₄ nyral₅ lili₆ to₇ myffai₈ nini₉. 'It₍₁₋₆₎ is₉ your₇ cloak₈ that₍₁₋₆₎ my₂ sister₃ is₍₄₋₅₎ washing₄₋₅'
Cf. Hus₁ e₂ tatsen₃ rissjun₄ nyral₅. 'My₂ sister₃ is₍₄₋₅₎ washing₄₋₅ it/this₁.'
The verb must be put in the end of the main clause (as actually of any other type of clause); it can be followed by discourse particles, but the latter can be viewed as a special type of auxiliary verbs. Only extraposited subordinate clauses (either postpositive relative or conveying reported speech) can be put after the main predicate.
Adverbs must immediately precede their head verb and cannot be separated from it except by another adverb.
The order of all the other constituents is “free”, since morphology provides sufficient information about their syntactic roles. The most prominent regularity with their placement is a tendency to consider one nominal constituent as “topic” and the rest of the sentence as a comment describing its condition or an event that happened to it. Normally (i. e. in unemphatic speech), the topic is put first, and all the other constituents follow it.
There is a tendency to put contrastive (and other emphasized) topics in nominative, irrespective of their intended syntactic role, and to put the role-switching marker immediately after them.
Any constituent can be dropped when clear from the context. In principle, a verb taken alone can be a grammatical sentence.
[Stuff to be added later]
The most common copula is nini v. ipf. 'to be'; another example is njun v. pf. 'to become'.
A noun used as nominal predicate must be in the nominative case.
Adjectives are used in their usual uninflected form.
Nouns in spatial cases (mostly locative) and some adverbial particles are put before the copula without any modifications.
The negative particle wo can be put before any part of the sentence to negate it.
Adpositions can be derived from verbs or nouns.
Postpositions derived from verbs
Some verbs in the position of the first part of a serial construction can be used to introduce their object as an additional nominal argument of the main predicate, i. e. like a postposition. The verb tsomin v. pf. 'to take' is often used in this way, as e. g. in the following sentence:
Nontai₁ sunollowis₂ tsomin₃ nontisal₄ mitserawis₅ ni₆. 'The man made the woman's small image of wood' (lit.: 'The₀ man₁, having₍₃₎ taken₃ wood₂, made₆ a₀ small₅ image₅ of₍₄₎ the₀ woman₄').
[More PI lexical stuff is needed to expand this section]
Adpositions derived from nouns
Typically, an adposition is derived from a case form of noun with an attributed genitive, and thus is a postposition governing the genitive case. For example, from the inessive of tswis 'inner side, part; interior, inside':
Fesal₁ tswisyry₂ 'inside the house' (lit. 'in₍₂₎ the₀ interior₂ of₍₁₎ the₀ house₁').
Fesal₁ tswismoi₂ 'from inside the house' (lit. 'out₍₂₎ of₍₂₎ the₀ interior₂ of₍₁₎ the₀ house₁').
However, the genitive can be substantivized, and put after the adposition as its apposition:
Tswisys₁ pys₂ fesallyry₃ 'inside the house' (lit. 'in₍₁₋₂₋₃₎ the₀ interior₁ which₂ is₂ of₍₃₎ the₀ house₃').
Tswisma₁ ma₂ fesalmoi₃ 'from inside the house' (lit. 'out₍₁₋₂₋₃₎ of₍₁₋₂₋₃₎ the₀ interior₁ which₂ is₂ of₍₃₎ the₀ house₃').
This pattern may produce e. g. a preposition that governs an essive or ablative case of substantivized genitive, as above.
Besides, the lexical noun can be put in a spatial case which is used as a postpositive attribute to tswis in the same or another spatial case:
Tswisyry₁ fewiry₂ 'inside the house' (lit. 'in₍₁₎ the₀ interior₁ which₀ is₀ in₍₂₎ the₀ house₂').
Tswismoi₁ fewiry₂ 'inside the house' (lit. 'out₍₁₎ of₍₁₎ the₀ interior₁ which₀ is₀ in₍₂₎ the₀ house₂').
This produces a preposition again, but now governing a spatial case of the lexical noun itself rather than of its substantivized genitive.
[More PI lexical stuff is needed to expand this section]
The reported speech clauses (which actually can be used to convey the content of knowledge, thought etc.) can be construed in one of two ways.
One way is to use the clause as the topic of a sentence and to mark its syntactic role with the role-switching pronoun:
Toi₁ tsjuma₂ tsjurin₃ pipys₄ ai₅ fati₆. 'He₅ said₆ that₍₄₎ you₁ would₂ come₃'.
The other way is to mark its role in the main clause with the cataphoric ri in appropriate case, and to put the subordinate clause after the main predicate:
Ai₁ hus₂ fati₃, toi₄ tsjuma₅ tsjurin₆. 'He₁ said₃ that₍₂₎ you₄ would₅ come₆'.
The latter type is called “extraposited”.
Derivation of nouns
[Stuff to be added later]
Derivation of verbs
[Stuff to be added later]
[Stuff to be added later]