Subject adpositions in Dairwueh and Bryatesle

This is a draft for some work that 'ties' together one detail in Dairwueh, Bryatesle and Sargaĺk. Details may still change.

Proto-Bryatosargaĺk had a subject postposition, probably something like dant. It's most clear descendant is the pegative marker -ta in Sargaĺk, the two -at-suffixes in the Dairwueh case system are probably descendants as well but less obvious examples. Some masculine and feminine nouns in Bryatesle have also incorporated it into their lexical forms.

However, subject adpositions seem to have been a sufficiently important trait in the syntax of PBS that new ones took its place in branches that lost 'dant'.

From the evidence, we can gather that it is unlikely the subject postposition was mandatory in every subject NP - rather, subjects with certain particular semantic or syntactic properties called for it. In all descendant languages, the subject postposition leaves some traces, but the traces are somewhat spread out.

1. Sargalk

In Sargalk, the postposition became the pegative case. This case marks the subject of a ditransitive verb, but also appears in some other contexts:

• with some particular verbs
• before some postpositions
• to mark intensity of actions
• for subjects of generic statements about a class (generally in singular)

A different subject preposition 'ved' also emerged, which was used thus:

•  agent-like, or agent-associated comitatives
• an optional marker for transitive subjects that had been displaced from their expected place in the clause
• entire subclauses and gerunds standing as subjects of a verb
• after a demonstrative that is a subject, when its referent is the previous sentence
• contrasted subjects
• whenever an explicit subject is present with an imperative
• the agent of passives

2. Dairwueh

In Dairwueh, dant has no direct descendant outside of morphology, but the preposition 'bur' has taken up a similar role. It appears in the following contexts:

• as a marker of resumptiveness in subclauses
• emphasized 'continued same subject'-pronouns
• sometimes to mark definite intransitive subjects
• with loan words that cannot be inflected for case, and with NPs that have no case-bearing marker (e.g. a subclause or similar)
• subjects in subclauses, and as an introducer of subject-oriented subclauses
• in comparisons of subject to subject, the subordinate subject has this marker
• in clauses with a vocative, to distinguish which NP is subject and which is vocative.
• in some dialects, atypical subjects such as mass nouns take it
• enables subjects to take a preprepositional, which usually used to convey in what capacity someone does something.

3. Bryatelse

Modern Standard Bryatesle lost 'dant', but gained a postposition 'uid', in some dialects 'uib'. There are also dialectal traces of 'dant'.

The 'uid' postposition:

• relative clauses that relativize the subject mandatorily have this marker on the subject - even if it strictly speaking is outside of the subclause.
• comparisons of subjects
• subjects in subclauses mandatorily take 'uid'.
• generic statements about a class of things (generally singular)
• a handful of verbs require 'uid' for the subject.
• similar vocative distinction as in Dairwueh
• a few dialects have adapted 'uid' to be a copula.

Traces of 'dant':

• in northeastern dialects, the definite form is a result of *dant → end.
• in southeastern dialects, *dant → ''tẽum" functions as an existential verb.
• In some dialects, *dant → -tɨnt is the secondary subject marker. The majority of dialects have -nisr, which probably derives from a different postposition that also operated as some kind of subject-postposition-like thing.

Dairwueh: The Recipient

There are restrictions in Dairwueh on when recipients or benefactives can be marked by the dative - whereas perceivers and experiencers can be marked by the dative nearly anytime.

The requirements for a recipient to be in the dative, rather than marked by the əre preposition, are as follows:

• The verb either has a nominative subject or a pro-dropped subject.
• If the subject is in the genitive (due to being the definite subject of a transitive verb), the verb must describe a concrete exchange of possession or control of the object. If it is in the nominative, no such restriction exists.

This restriction seems to come out of a conflict for control of the verb phrase by genitive subjects and dative recipients (see the first table here for reference), but this notion of control seems to be 'transferable' if there is a clear enough vector of transfer of the control of a third actant. In some speakers estimation, a dative is also acceptable if an instrumental is present.

With a few verbal constructions - causatives, for instance - the dative cannot mark recipients or benefactives for similar reasons: there is a subject that has too much control, and this subject does not cede the control to the recipient. Nominative subjects in such constructions belong in the +control +subject cell of the subjecthood-control scheme.

Another such restriction seems to be whenever a dropped, implicitly nominative subject is coordinated with a genitive subject - even if the verb is not transitive:

*man.gen lit a candle and sang her.dat
the man lit a candle and sang for her

It seems a coordinated genitive subject is enough to force the genitive-like requirements onto non-transitive verbs as well.

The Syntax and Semantics of Gender in Dairwueh, Bryatesle and Sargaĺk

In the DBS family of languages, gender/noun class, much like in Indo-European, Semitic, Northeast Caucasian and Bantu languages, is an important feature that has remained present over a long span of time. Only one dialect of Sargaĺk has lost the gender system, viz. Inraj Sargaĺk, which has done so under strong influence from a gender-less substrate.

For a conlanger, it's very easy to say 'my conlang has so-and-so many genders' without thinking much about the semantics of the system. This is a post that tries to set a different tone.

Syntactical and Morphosyntactical Properties of Gender

In both Dairwueh and Sargaĺk, there is an independent morphology for adjectives and pronominal determiners that conveys agreement with the head noun. This congruence extends to adjectival complements of verbs, and may also affect verb congruence.

Sargaĺk seems to have an innovative feature in its having dropped the neuter, shifting the nouns over to masculine and feminine.

In most of Bryatesle, syntactically gender only affects pronouns and determiners, and to a limited extent verbs: neuters and non-neuters have different verb forms in the third person. Morpohologically it also has a strong effect on nouns.

Only Sargaĺk retains the gender distinction in the plural.

Semantics

The obvious semantic split in the gender system of all three is one corresponding to biological sex. However, this only applies to humans: the biological sex of 'generic' animals does not strongly correlate with the gender used. In part, this disconnect has a practical epistemological basis: if you see a horse in the distance, you may not be able to tell what sex it is. (And indeed, animals for which it is likely that a speaker can distinguish the sexes, it is more likely that no generic noun at all exists. This especially holds for some tamed birds, some livestock and dogs, but also in the case of Sargaĺk with regards to some really huge seals.)

When speaking of individuals of the species you either have morphological devices for clarifying the sex of the generic noun, or gender-specific nouns at your disposal.

In Bryatesle and Dairwueh, there also exists a neuter gender. It would be easy to say the neuter gender is for inanimates, but this does not hold; the only strict rule in regard to the neuter is that it does never refer to humans.

For many of the tendencies described here, conflicting tendencies sometimes appear due to the Sargaĺk gender system.

The animals that are neuters are generally not very well "respected" - in both Bryatesle and Dairwueh there's a surprising agreement: certain small predators that kill poultry, stinging insects (except the bee in areas where beekeeping is practiced), frogs, lizards, snakes, and inedible tiny kinds of fish.

Regarding inanimate things, it turns out that there is a rather elaborate distribution between the genders: nouns that provide a context tend to be of the same gender as the gender associated with the context - i.e. men and fishing boats in all of the languages, women and the house in all these languages. A context mostly is a place (natural or a building), or something within which a person can be and act. However, any prevalent and topical tool smaller than a boat is likely to be of the opposite gender (but also possibly neuter in Dairwueh and Bryatesle), whereas objects that appear in many contexts tend to have a rather equal distribution over the available genders. Thus, the hammer is feminine, the spinning wheel is masculine, a net is feminine, a bread paddle is masculine.

Dairwueh:

teməri ( hammer, feminine )
yənera ( spinning wheel, masculine )
muliri ( net, feminine )
eskəna (bread paddle, masculine )

Bryatesle

demby ( hammer, feminine )
yinvinu ( spinning wheel, masculine )
nvuly ( net, feminine )
iska ( bread paddle, masculine )

This distribution probably has emerged to increase the likelihood that a thing and a person spoken about will be of distinct genders, and thus have distinct pronouns. In the case of Dairwueh and Bryatesle, of course, the use of neuters facilitates distinct references to an additional noun if it happens to be neuter.

This also extends to verb nouns: a verb noun that is associated with women will generally be masculine (or neuter), and vice versa, thus

gistas ( breastfeeding, masc )
zexnas ( cooking, masc )
bagrasi ( the act of hunting, fem )
vyrnasi ( participation in a battle, fem )

from the verbs

gistër, gistai, breastfeed
rexnër, zexnai cook
begrer, bagrai hunt
vyrner, vyrnai fight a battle

In both D and B, verb nouns that are not particularly associated with either of these genders tend to go with neuter, but so do verb nouns that are associated with non-human subjects regardless of gender.

1. Size

Things of extreme sizes - very small or very big - tend to be neuter. In Sargaĺk, they tend to be masculine unless they have some kind of cultural prominence, in which case they may be feminine. If there is a progression of sizes (e.g. mountain - hill, river - ditch, tree - bush), the order tends to be neuter > masculine > feminine.

Bryatesle:
burts - mountain
xyles - snowflake
ulid - valley
sarn - the sky
lyvmat - the world
mal - a grain of sand
nibit - a drop
yfir - a louse
ebik - a birthmark
ryts - a scale (of fish or reptiles)
sinis - a very tiny amount

Dairwueh:
pəltən - mountain
əls - a valley
xorm - the night sky
yunt - the world
ʒiks - the 'details' in a coarse surface
nəf - a drop
ifrəl - a louse

2. Edibility

The more edible a thing is, the more likely it is to be neuter (or in Sargaĺk, masculine). Flavour, nutrition and texture all are factors in this. Live animals are obviously an exception to this. Things that are entirely outside of the scope of even consideration for edibility are not affected by this tendency.

An inedible root or plant is likely to be feminine, as is animal considered to be inedible.

3. Social roles and professions

In all these languages, nouns denoting social roles and professions will follow the gender they are most closely associated with, even if the particular referent is of a different sex: thus, it is grammatically ok in all of these languages to say something like

she is a fisherman

or

he is a seamstress

However, pronouns that are not determiners will adhere to sex.

In Bryatesle, there is an analogous extension in the function of buildings of different types (i.e. a wooden hut used for storage or for living in or for doing some kind of work in). A storage room for firewood is a

yvalk

which is masculine , but if the building is a 'hut', it is a  

sira

which is feminine. In this case, the 'actual sex' of the building follows the building type rather than the building 'profession', and so you could have a feminine pronoun referring to an yvalk, if it is known that the yvalk is a sira.

Named buildings also have genders, which tend to be building-specific.

Dairwueh: Personal Names and Cases

An Indo-European-like trait that Dairwueh and Sargaĺk share (but that Bryatesle lacks!) is case congruence on adjectives. What makes this particularly Indo-European-like (and not, e.g. Baltic-Finnic like or Kayardild-like) is the existence of distinct sets of morphemes for the adjectives and for the nouns. 

I have understood that the origin for this in Indo-European is that morphemes originally used with pronouns for some reason migrated onto the adjectives, although this of course leaves open the question of the origin of the distinct morphemes on the pronouns. (However, typologically I doubt whether that's very unusual. We find small examples of similar things elsewhere, such as the Finnish -t accusative for personal pronouns, or the comitative requiring possessive suffixes on the noun - the latter possibly leading, over time, to a situation where the noun and adjective have distinct forms.)

What's this to do with names? In Dairwueh, personal given names can behave both like nouns and like adjectives, depending on the presence of a clan name. A patronymic can behave like a noun if no other part of the name is present.

Thus,

nom: Doras
acc: Doranna
dat: Doraar
gen: Doraat
loc: Doraŋa

would be the noun-like forms, and if Doras' father was Elti, you get

nom: Eltikar Doras
acc: Eltikan Doranna
dat: Eltikarz Doraar
gen: Eltikarz Doraat
loc: Eltikari Doraŋa

however, if a clan-name was involved, Doras too would - except in the nominative - inflect by an adjectival paradigm:

nom: (Eltikar) Doras Marzi
acc: (Eltikan) Doran Marzinna
dat: (Eltikarz) Dorarz Marziar
gen: (Eltikarz) Dorarz Marziat
loc: (Eltikari) Dorari Marziŋa

In the nominative, the patronymic has its own feminine form e.g. Eltikama (derived from the father's name, though), but in all other forms, it basically used feminine congruence instead on the masculine patronymic stem.

Nick-names of course exist, and tend not to adhere to this pattern. However, a nick-name nearly never is used in apposition with patronymics or clan-names.

As linguistic history goes by, other forms of 'family names' besides clan-names start appearing, with the usual suspects: professions, places of origin, remarkable attributes, etc.

// TODO: I should definitely finally get around to getting those adjective case markers done for Dairwueh.

A Number and Numeral-Related Thing in Sargaĺk and Dairwueh

There's no need for a language to have a 'perfect' analogy to the word 'both' (despite the fact that it exists in several subfamilies of European as well as in several Uralic languages, and these are only the ones I've been able to verify that they are not direct cognates or derivatives of the lexeme for 'two'). However, Sargaĺk manages to double that, by having two words with similar meanings but different morphosyntactical behaviours as well as slight differences in meaning.

In Sargaĺk, two is yor. 'Both' is either vrir or lyəs. Vrir takes a formally singular noun after it:

vrir	tame	-ta
both	man	pegative sg
both	men

There are some complications: vrir does not distinguish the absolutive and pegative. For nominative or pegative nouns, it is always itself unmarked, but has the pegative singular marker on the main noun. For all other cases, the noun is in the singular case, and vrir takes the singular oblique case congruence:

vri(r)	-ə	tame	-rne
both	singular masculine oblique	man	singular lative
both		men	to

Lyəs however, takes plural congruence with all cases and the main noun too is consistently plural. As subjects the verb for both of lyəs and vrir take plural congruence, except if vrir is used with certain words like 'hands', 'eyes', 'ears', 'nostrils', 'scissors' or 'the side of a boat'.

Both of these can also be used as pronouns, much like English 'both'. They can also be used for a dual reflexive construction which can be used with any subject numbering two, regardless of morphological number.

The semantic difference lies in the extent to which the two referents are seen as separate units ('lyəs') or a concerted group ('vrir')

In Dairwueh, a cognate of vrir exists, ŋrəz. This particle has a few uses that have developed out of an original meaning of 'both': in NPs it goes before any number to mark 'all N of', but without any explicit number present it signifies 'both'. In numeral complements it serves to mark the number as that of a group, rather than as a number of independent individuals.  This it also does with plural, indefinite determiners and pronouns, thus:

guni	ŋrəz	tirs
are.3pl	(both)	six
they are		six

They are (a group of) six.

v.s.

guni	tirs
be.3pl	six
they are	six

they number six, there are six of them, (but as individual things)

ŋrəz is the only 'numeral' in Dairwueh to take case. It roughly follows the plural paradigms:

masculine:

nom: ŋrəzo
acc: ŋrizna
dat: ŋrizit
gen: ŋriŋa / ŋridin
loc-instr: ŋriŋa / ŋrider

feminine:

nom: ŋriri
acc: ŋrizar
dat: ŋrizit
gen: ŋrizin
loc-instr: ŋrizar

neuter:

nom: ŋrəza
acc: ŋrəza
dat: ŋrizit
gen: ŋrizit
loc-instr: ŋriŋa

Ćwarmin, Bryatesle: Proper Nouns and Definiteness

In this post, "definite" refers, with regards to noun-phrases, to the quality of being a referent whose identity is known to the listener of an utterance. "Specific" refers to the quality of being a referent known to the speaker.

Ćwarmin and Bryatesle seem to have evolved definiteness marking in their case systems during times of contact. Ćwarmin's indefinite/specific/definite-distinction seems to go back fairly early, though, but we find dialects in contact with Bryatesle sometimes missing the specific/definite or the specific/indefinite distinction. The stage at which definiteness started becoming a thing in the Ćwarmin branch and Bryatesle-Dairwueh must be around the time of proto-BD, but later than Astami began diverging from the rest of the Ćwarmin languages.

Different languages in these groups have, however, dealt somewhat differently with definiteness marking on proper nouns. Proper nouns are most often definite by nature. In Ətimin, proper nouns are not marked for definiteness at all, with a few toponyms as exceptions. Rasm'in' and Ćwarmin, however, tend to use definite case marking for proper nouns in cases other than the core cases nominative, accusative, genitive and dative. The genitive and dative are flip-flopping in both, though. In these languages too, some toponyms have names where even the core cases are definite. However, exceptionally, Rasm'in' has a toponym that is always specific rather than definite, viz., mworanyus ädʒiniis', '(these/some) property markers of the belly', a historical border marker between the Kəlkəj and Moduwt tribes. The 'belly' refers to this being the most 'central' of the border markers between the tribes.

In Ćwarmin, personal names sometimes may appear in the specific. This, oddly enough, tends to indicate that the person referred to is a very recent acquaintance of the speaker, and that the speaker therefore is not sure whether the listener knows the referent.

With non-nominatives in Bryatesle, the definite case is almost always used with definite proper nouns except toponyms (and even there, many toponyms are definite - e.g. Zgakintën, 'the hilltop', a very definite hilltop near the capital. Omission of the definite marker, if no other secondary case takes its place, indicates specificity in non-nominatives. For nominatives, the partitive secondary case serves to indicate specificness.

One final language to consider is Dairwueh. It too has a very limited definiteness marking in the differential object marking of transitive subjects - genitive for definite, transitive subjects. It turns out a similar pattern holds there - nominative indicates specific, genitive indicates definite.

Dairwueh Case System: Cases as Bundles of Features, pt II

Dairwueh has five cases,

• nominative
• accusative
• dative
• genitive
• locative-instrumental

The following treatment only deals with the use of these cases when they are not accompanied by prepositions. Prepositions basically just ignore this stuff.

Basically, all cases except the locative-instrumental sometimes are subjects, giving us the following feature: potential subject? Of course, this is a very limited feature of most cases - the accusative being subject of a few verbs, the dative likewise. Thus it is not a very strong feature; however, as for subjects go, nominatives and genitives both very frequently occur as subjects. We basically can assign those two the value +subj.

Before going on with this, I need to explain quickly the use of the genitive for subjects in Dairwueh: the genitive stems historically from an ergative case. It is used for definite subjects of transitive verbs.

Now we have (nom, gen) vs. (acc, dat, loc-instr). We want for the next feature to pick out one or more out of both these sets. Obviously the feature needs to distinguish nom from gen. Genitive and loc-instr both do adnominal things, but we can also consider how dative and genitive both imply some kind of control over something else: the dative receives control, the genitive has control. The nominative, however, also can have control over something - in the case of an indefinite subject it has control of a transitive verb, so, this particular feature would only serve to divide up the (acc,dat,loc-instr) set: (acc, loc-instr) vs. (dat) which pairs with (nom, gen) as far as this feature is concerned.

	control	¬control
subject	genitive (nominative)	nominative
¬subject	dative	accusative, loc-instr

This is not even really all that ineffective, but helps us envision how to split the next pair of pairs: (nom, gen) and (loc-instr, acc). Alas, I cannot come up with any feature that would distinguish loc-instr in particular from accusative while also distinguishing genitive from nominative, except for the dialects where the locative-instrumental marks all possessums. For such dialects, a "possessor/possessum" diagonal split would work. However, this diagonal only splits the cells it passes through, not including the two other cells of the table at all.

Here's a different option:

	¬adnominal	adnominal
subject	nominative, genitive	genitive
¬subject	dative, accusative	loc-instr

The two tables above should suffice together to distinguish all the cases except for the nom-gen definiteness distinction. This distinction is explained above, and is unique to that pair.

This featural decomposition basically hints at how these cases are used beyond the implications their names imply: recipients, locatives and instrumentals, possessors, subjects, objects.

A Bryatesle-Dairwueh-Sargaĺk Adjective Type

In the BDS family, certain particles with adjective-like distribution are present in all languages. These adjective can be used with any NP in almost any position, even though the most similar translation in English cannot: 'what about/how about ...'

In Bryatesle, the adjectives are kyne- and sudu-, in Dairwueh xən- and orə-, and in Sargaĺk you find cin- and asku-. The pairwise distinction between these differ between the languages: the Dairwueh pair differ by animacy (xən- being animate, orə- being inanimate), the Bryatesle kyne- and sudu- differ by number (sudu- being plural), and the cin- and asku- pair of Sargaĺk differ by definiteness (asku- being indefinite).

It should be clear that kyne-, xən- and cin- are cognate. One rather typical vowel correspondence pattern in monosyllabic roots can be spotted here.

Bryatesle /ɨ/ <y> : Dairwueh /ə/ <ə> : Sargaĺk /i/ <i>

The sound changes on the initial consonants are a bit more complicated though. The cin- in Sargaĺk suggests the original sound was /k/ rather than /k'/, since /k'/ is more stable against sound changes in Sargaĺk. This fits well with Dairwueh, where k > x, k' > k. The three other forms do not seem to be cognate at all.

Some examples of use would be these:

Bryatesle:

tvem kynë mindë gavari livytri

tvem kyn-ë mind-an gava-ri livyt-ri

tvem		kyn	-ë	mind	-ë	gava	-ri	livyt	-ri
you		how about	def.acc.fem	girl	acc.fem	meet	2sg.atelic	went	2sg.atelic
you		how about	the	girl		that you meeting		were

how about the girl you were meeting

Dairwueh:

xənŋa srotoŋa misandeb

xən	-ŋa	sroto	-ŋa	misand	-eb
how about	instr/loc	small boat	instr/loc	arrive	2sg.past
how about	with	the boat		arrived	you

How about (with) the boat (with which) you arrived 

We notice in both Bryatesle and Dairwueh that the case function of the noun is somewhat ambiguous - it is never quite clear in past tense expressions whether the case pertains to its relation to the past tense verb or how it relates to the inquiry. Therefore, the case forms often may get somewhat confusing.

To Be Right in my Conlangs

To express the sentiment of being right in my conlangs, a number of slightly different constructions are used. However, first a quick overview of being right in the languages I speak and a few more.

In Finnish, the usual way would be olla oikeassa, essentially 'to be in (the) right'. In Swedish, however, the expression is ha rätt, 'to have right'. This also seems to be the way of expressing it in German, recht haben. Russian has прав, 'right, true' as a complement, so e.g. ты прав, "you (are) right/true". Onto languages I don't speak: The French and Italian expressions seem to be analogous - although 'to have reason' seem to be more literal translations. Spanish is a bit like Finnish, although with 'in truth' (en cierto) rather than 'right'.  Bear in mind that I might be wrong on any of those beyond Finnish, Swedish, English and Russian. As a side note, for Finnish, English, German and Russian, there's a clear connection to right (the direction or side).

So, onto the conlangs!

Bryatesle

In Bryatesle, being right is expressed using a verb that is closely related to the verb 'carry', verg, in the perfective aspect, with a quirky case subject, viz. the ablative.

tërty virg-a
you.abl carry-perf.3sg
you're right

In isolation, virga sometimes is used to express 'you're right'.

Ćwarmin

In Ćwarmin, being right is expressed by the demonstrative adverb olba/elbə ('this') in an adverbial form, olbaru, elbəri (essentially 'like this, like so, thus') or in the nominative definite olbutu, elbiti.  The construction with olbaru/elbəri uses the reflexive possessive accusative:

bec olbarsun
you thus-refl.poss.acc
~you have thus
you are right

The olbutu/elbiti form comes with the dative of whomever is right:

un olbutu
(s)he.dat right.nom.def(s)he is right

Dairwueh

Unlike Bryatesle, Dairwueh uses the verb əduin, 'hold', but like Bryatesle, it uses a quirky case subject: the genitive. The verb is in the 3rd person II.

vedin ŋe ədu-ar
I.gen was.3sg hold-past.prtcpl
"my was held"
I was right

Ŋʒädär

Ŋʒädär has a reflexive, locative expression. 

(sint) prä-ŋä-bürs-äz ŋul-ər
(they) right-at-3pl/3pl-direct self-plur
they are right

Prän does in fact have a slight typological similarity to 'right', although it signifies a different type of direction: in fact, it signifies the east. prä-ŋä thus also signifies "in the east", but the only situation in which this appears with transitive marking is when signifying 'being right'. For those who haven't read how the reflexive works in Ŋʒädär, here's a post.

Sargaĺk

Sargaĺk uses a reduplication-like construction:

(ne-tta) tvadas tvadas yəra-si
I-peg truth truth put-1sg
I put truth (to) truth
I am right

Clearly, the pegative of the subject suggests that the two instances of "truth" are considered different constituents, however - one being a direct object and one an indirect object. Hence the (to) in the word-for-word English translation. Here, as well, the subject pronoun can be omitted.

Dairwueh: Relative Clauses

Relative clauses in Dairwueh reuse a certain idea I quite recently had – the relativizing auxiliary. In Dairwueh, it is kadal. Its main forms are

present tense:
3sg I: kada (masc, neut) | kado (fem) | kaduni (plur)
3sg II: kadal (also infinitive)
past tense:
3sg I: kadiŋ | kadari (plur)
3sg II: kale
irrealis:
3sg I: kadəyi (masc, neut) | kadəvo | kadəŋan (plur)
3sg II: kaləy
negative present and irrealis:
3sg I: kadešne
3sg II: kaleš
negative past and irrealis:
3sg I: kadeyš | kadeyšin (plur)
3sg II: kaleš
passive:
3sg I: kadeŋa
3sg II: kaleŋ

The difference between 3sg I and 3sg II is somewhat different than elsewhere in the language - 3sg I is restrictive, 3sg II is nonrestrictive. The main verb appears in an infinitive or participle form. If the relativized noun is some form of oblique or non-nominative argument of the subclause, a resumptive pronoun appears in the clause in the relevant case (or with the relevant adposition). 

If the relative clause lacks an external head, the verb will almost always be 3sg I. If the implicit external head is not nominative or accusative, there has to be some external resumptive pronoun in that case. Often, this is a demonstrative, but a personal pronoun is also possible, and comes immediately after or before the subclause.

Whenever the external head is a personal pronoun, there may be a person discongruence - and there may even be a personal subject person in the subclause itself, e.g.

ekadašor ver samotəg, (ver) samotas
rel_verb.pass_neg I win.inf, win.1sg
I, who is not won over, win 

Intonationally, the subclause is often introduced with a slight drop in pitch for first word of the subclause - which is not necessarily the relative auxiliary verb, followed by a relatively quick rise, followed by a slow descent to the end of the subclause.

Dairwueh: A Subject and Object Preposition

Dairwueh also has prepositions marking the subject, object and the indirect object, although these prepositions are very optional. They serve a few roles, however, and are mandatory when the conditions for some of these roles are fulfilled:

• enabling preprepositional arguments for subjects, objects and indirect objects
• intensification of the verb (optional)
• topicalization (optional for fronted objects)
• gerund formation for whenever the gerunds are subjects or objects
• a variety of things with regards to transitivity-changing operations
• introduces direct objects with normally intransitive verbs
• introduces datives for verbs that normally do not take them
• marking resumptive pronouns that are objects (mandatory)

The subject preposition ne cannot take genitive subjects in most dialects, but does tend to imply definiteness - and is also the main way of marking definiteness for intransitive subjects. The object preposition ne takes whichever case the direct object of the verb takes, and thus is one of the few prepositions to take many different cases. Obviously, the prepositions for subjects and objects are identical except for the case they take. The indirect object preposition ser takes the dative.

Preprepositional arguments of subjects and objects serve certain roles:

For subjects, it represents the role a subject is or is imagined to be in:

sparage	-ŋa	ne	Armus	side	erb	-aŋ
farmer	masc sg gen	subj prepos	Armus (proper name)	good	is	3sg I
as a farmer			Armus	good	is

as a farmer, Armus is good

For objects, it represents, likewise, the role an object is in; there can be a causal relation - I ate something because it was food. I gave someone something because it was their inheritance. For indirect objects, ne can also be used for causal relations: I gave someone something because he was a cleric. ser does not permit causal relations: the preprepositional marks the use for which a direct object was given the indirect object. Notably, ser cannot stand with all dative arguments, only with those that are indirect objects.

A Dairwueh Preprepositional Construction

In Dairwueh, having is expressed using the preposition edə in combination with the third person II copula -

	pos	neg	pos irrealis
pres	erb	dirš	ŋey
past	ŋe	ediš	(ŋe)

Thus, baud ŋe edə vena  - I had a farm (farm was by me).

The preposition edə occurs in some temporal expressions and some fossilized locative expressions, but is otherwise not used much.

However, edə can also take a preprepositional noun or adjective (or even adverb). In combination with the possessive construction, the adjective corresponds to the bolded elements below:

• to have object as/for complement
• X's object is complement
• object is X's complement 
• object is complement by X/for X
• owner perceives object as complement

In a few dialects, the complement can also be an infinitive, in which case it's a causative structure, or an active(!) participle, in which case it's a passive.

Dairwueh: Noun Morphology in Depth, pt I

Providing the noun case and number suffixes for Dairwueh hides some of the things that are going on. We find, for instance, nouns like xei, whose plural stem is i-, giving us the following paradigm:

xei   ir
xena ijivna
xear ijivit
xeat ijedin
xeŋa ijeder

For most nouns, two stem forms suffice, but some nouns require up to four stem forms. However, there is no single system of four stems that is sufficient to account for all morphological variation – one needs five stems, of which no word uses more than four distinct ones:

1. singular nominative stem
2. singular oblique stem
3. plural nominative stem
4. plural oblique stem
5. derivative stem

Few nouns have only one stem - these are basically only nouns that lack plurals or singulars altogether.

Nouns with two stems come in two main flavours: 2a: 12 / 345 (almost all being feminine), or 2b: 1234 / 5 (almost all being neuters). However, 2c: 1235 / 4 also exists for a few nouns, (e.g. erha, 'king') and 2d: bits, 'direction', exceptionally follows an 1 / 2345 pattern. Whenever a noun has three distinct stems, usually the lines of division are 3a: 1 / 234 / 5 (e.g. dor, 'man'). A small minority of nouns has 3b: 123 / 4 / 5. For nouns with four stems, the two that are merged are always either 4a: 2 and 4 or 4b: 3 and 4. Part 2 of this post will go through the historical reasons behind the morphophonological changes, this post only provides examples of the classes and explains the basic structure.

Further, the derivative suffixes that are given as examples below are of course also members of such classes; -res, for instance, is a member of 3a, -res, -rto-, -rr, while -pan is a member of 3b, with -pan-, -po-, -pla- as the stem forms.

2d: bits, 'direction'

	sg	pl
nom	bits	bətil
acc	bətna	bətivna
dat	bətar	bətivit
gen	bətat	bətŋa
loc-instr	bətŋa	bətŋa

Derivative example:

bətres: a signpost

3a: dar, 'man'

	sg	pl
nom	dor	daran
acc	darna	darivna
dat	dar	darivit
gen	darat	daredin
loc-instr	darŋa	dareder

Derivative stem: dri-

dripan: manliness, masculinity

2c: erha, 'king'

		sg	pl
nom		erha	erhan
acc		erhana	erivna
dat		erhar	erivit
gen		erhat	eredin
loc-instr		erhaŋa	ereder

Derivative stem: erha-

erhaksa: kingdom
erhapan: royal legitimacy, inheritance of kingly title

4a: soŋe, 'noble title'

		sg	pl
nom		soŋe	soša
acc		soke	soka
dat		sokn	sokivit
gen		sokŋa	sokivit
loc-instr		sokŋa	sokŋa

Derivative stem: sot-

sotres: banner
sotukri: a woman whose nobility passes by female inheritance
sotsek: a nobleman

The Anatomy of the Description of a Conlang

My previous post got a question that got more difficult to answer the more I was thinking about it.

Is this meant to be a simple quirk of the language or is there a shade of meaning implied by the genitive marking of tired, as opposed to, say, sleepy?

This was meant to be a quirk of the language, much like there are words that lack congruence, have incomplete congruence or have extra congruence in a variety of languages. Consider, for instance, English this/these vs. the, or Swedish trött which conflates neuter and common gender forms (but seems to rather maybe lack a neuter form altogether?).

However, I here need to divert attention to what I am describing when I describe Dairwueh (or Sargaĺk, Bryatesle, etc ...). Of course I am describing some kind of partial language whose only 'real' existence is in the descriptions I have created and the ideas I have in my mind. However, these ideas have some kind of structure to them, and I don't just mean that the ideas belong together. In some sense, my descriptions of these languages are synchronic - they pertain to one particular timespan in an imaginary timeline. There's some ideas as to what came before, often in a rather vague form, and sometimes there's ideas as to what is coming to come later.

Of course, any real language has a lot of individual variations - consider, for instance, the difference between partitive and accusative objects in Finnish. Can we be sure that every native speaker encodes the exact same distinction by that? I am not so sure of that! I am inclined to think that, in fact, we have a lot of greatly overlapping distinctions, most of which don't differ by much, but the occasional outlier exists. Any 'properly realistic' conlang should also have this, I think. And thus I will usually end up with slightly fuzzy ideas of what's going on - I'll try and fuzzily permit for variation in my idea.

The tired bear ate the honey. (the bear possesses tiredness?)
The sleepy bear ate the honey. (the bear has the property of sleepiness)

I don't think there's any such meaning-difference distinguished here in the Dairwueh-period I am describing, but the origin of the construction might be an almost-onset of such a development. Consider "a strong man" vs. "a man of strength" as a similar distinction, but with strength in the morphologically marked genitive instead, and as an adjective, not a noun of the same root. So, the origin of this construction might've had some such meaning, but only the form got conserved, not the meaning. This is also why it did only remain in the nominative - being that it was more often used with nominatives than with other nouns (due to such distinctions being more often made for the subject than for other NPs), which is why the normal congruence won out for those.

Would a Dairwueh speaker notice the genitive marking or just think it 'proper Dairwueh-ish'?

They would notice it as the normal thing to do with those adjectives for those adjectives they are used to using that way, i.e. other dialects would feel weird if their set of such adjectives differed significantly. It is possible that they wouldn't perceive the -at marker as the same morpheme as the genitive masculine marker, though, but rather as an exceptional gender-indifferent nominative marker.

However, I can imagine a Dairwueh scholar writing something like

erb

e-

-tsihka(l)-

-šor

rəmg

dar

,

tsihka

-ŋ

rəmg

-at

dar

3sg negative passive auxiliary

neg pcpl prefix

write

neg pcpl suffix

tired

man

write

passive

tired

sg masc gen

man

is not

written

tired

man

is written

tired

man

erb

e-

-tsihka(l)-

-šor

rəmg

-i

tol

-i

,

tsihka

-ŋ

rəmg

-at

tol

-i

3sg neg passive auxiliary

neg pcpl aux

write

neg pcpl aux

tired

fem nom

woman

fem nom

write

passive

tired

sg masc gen

woman

fem nom

is

not

written

tired

woman

is written

tired

woman

Or in translation "it's not (to be) written 'rəmg dar', it's written 'rəmgat dar', it's not to be written 'rəmgi tol', it's written 'rəmgat tol', correcting uneducated speakers who are generalizing the adjectival congruence patterns even to these, or conversely saying to use an adjective in the regular pattern instead of the genitive pattern.

I ask because, perhaps frighteningly, I use your blog to learn linguistics and so I always look for a meaning in the bits of grammar you create and I can't see what marking 'tired' differently does to the word.

Usually, if such a different marking does something to the meaning of the word, I will point it out; in this case, it's just a morphological deviation with regards to these few nouns. As to using this blog for learning linguistics, I guess there's worse places, but I really really suggest you also use some kind of complement. This is of course just a sample of ideas from typology, filtered through potential misunderstandings on my part and then brewed in a vat where my imagination acts as the fermenting agent.

This will obviously result in a somewhat unbalanced diet when it comes to learning linguistics.

Dairwueh: Semi-genitive Adjectives

In Dairwueh, there is a set of adjectives whose congruence pattern is slightly off. Their behavior differs in one particular spot: for nominative nouns, their marking is masculine genitive. (The one exception is the word sabrin, "unusual, foreign, strange" which is in the feminine plural genitive with nominative nouns).

Not only do these behave slightly oddly within noun phrases (and fail to have comparative marking altogether), they also behave weirdly with copulas.

The main examples that are widespread through Dairwueh dialects include

koŋsat: last, final, complete
rəmgat: tired
lodat: correct, right, entitled
satpat: sturdy, firm
nestat: ancient
bartat: partial, incomplete, one among many, a few among many

Notice that the -at suffix in all of these is the genitive masculine morpheme, so the roots will consist of the adjective without that suffix.  Less widely distributed examples that are found in the capital area prestige dialect are

sxundat: wafer-thin
ropsat: bloody (from 'rapəs', blood, from *rrabx, bleed)
julkat: lazy
silgat: rank, rancid
xugat: brave
xsəlrat: avid, skilled, obsessed
tagrat: content
nalkat: sad
rusnat: smooth

With the copula, this type of adjective is marked by the preposition 'lo', when marking having that quality, and by the preposition 'əre' when marking acquiring that quality, and no congruence marking with regards to gender or number appears. The adjective then is in the masculine instrumental (-ŋa) with 'lo', and masculine accusative (-na) with 'əre'.

Dairwueh demonstratives

Dairwueh has two and a half level of demonstratives.

The two basic ones are 'this', av, and 'that', xev. The third, ŋev, is only distinguished in the nominative and genitive, and otherwise conflates with xev. The basic inflections are:

nom	acc	dat	gen	loc-instr
xev, xej	xena	xenar	xenat	xeŋa
ŋev, ŋej			ŋenat
av, aj	ana	anar	anat	aŋa

There are some further related adverbials derived from the same roots:

xeke - there
ŋeke - there (even further away)
ajke - here

xekem - thither
ŋekem - thither
ajkem - hither

xeŋesa - from there
xeŋesa - from even further away
aŋesa - from here

There are also verbs deriving from these roots, although ŋe- entirely is absent from these.

xevin - to go towards there
avin - to come towards here

xevki - to arrive there
avki - to arrive here

xeŋsin - to leave from there
aŋsin - to leave from here

xeski - to exit an enclosed volume, such as a house
ŋaski - to exit an enclosed volume, such as a house, in which the speaker is located

These four infinitives ending in -ki are the only Dairwueh infinitives to end in vowels. Beyond the verbs above there are also causatives, which are the main ways of expressing 'bringing' something:

xevlik - to bring there
avlik - to bring here

Avlik is mostly common as an imperative, avlu, avlisu, but appears in other forms as well. Xevlik in all its forms is by all counts significantly more common.

Numbers in Sargaĺk

In the initial sprachbund of which Sargaĺk (and proto-Sargaĺk, and the whole proto-Sargaĺk-Bryatesle-Dairwueh) was part, a few tendencies with regards to numerals held.

1. Numerals were morphologically invariant. (Except for morphemes that only served to 'structure' compound numbers)
2. Base twenty with a subbase of ten.
3. Numerals could be adverbials of repetition with intransitive verbs without any adpositions or other marker.
4. Ordinal-like meanings could be obtained either by combining cardinals with adpositions (Bryatesle 'dedak', after, giving Bryatesle its off-by-one ordinal system; Dairwueh 'yil', about, like, made from, consisting of, during; Sargaĺk 'jaxru' in a sequence (or path)) or by combining the number with a variety of verbs ('stand at', 'reach', etc)
5. Cardinals were not commonly used as attributes, but almost invariably as predicates.
6. Ratios were not really used much; half, third, fourth and fifth had dedicated roots; larger denominators probably never appeared until the Bryatesle expansion.
7. Ordinals for members of small sets (set sizes somewhat below 10) often were formed analogously to finger names; more general ordinals were formed by cardinal + adposition.
8. 'all' is a numeral, and has ordinal as well as other numeral-like uses; There are two 'alls' - 'all in a relatively manageable set' and 'all (in some set of unknown size)'
9. "none", "no", etc are not numeral-like in distribution

This has not survived entirely unchanged into Sargaĺk. Numerals, with the exception of 'one', are still morphologically invariant. The base still is twenty, and a subbase of ten still is clearly visible. Sargaĺk retains both adpositional/adnominal and verbal ways of expressing ordinals. Five is somewhat abandoned, in that numerals now freely appear as attributes. Recently ways of forming ratios of greater complexity than N/5 have developed. Ordinals by finger names can be used for orders in sets smaller than five, starting with the thumb. 'All' no longer distributes like a numeral, but one can form ordinals from it: from the greater all, 'od', the ordinal construction simply signifies the final, finally, the last one, from the set-specific all, 'odar', it signifies the most recent, the latest, the last one (of a queue or line or set).

In Bryatelse and Dairwueh, these properties have developed slightly differently: Bryatesle has given numerals case (and gender), both as independent heads of NPs and as attributes and determiners. There are but a few traces of base 20 left (the words for 200 and 400 among them). Much like in Sargaĺk, ordinals are formed by a postposition (but it has its 'off by one'-quirk). Cardinals are used as determiners and attributes, and ratios are commonplace. Names of fingers for ordinals have largely been abandoned. 'All' operates more like an indefinite determiner or pronoun.

The numbers of Sargaĺk are these (with some gaps where the forms should be predictable):

1: dər
2: yor
3: xrik
4: knər, bilon*
5: mil, salp*
6: mur
7: ćəx
8: ksen
9: lini10: təŋbar ('half-twenty')
11: dərdər
12: yoryor
13: xrikəxrik
14: knərəknər
15: xriksalp ('three hands')
16: murmur
17: ćəxćəx
18: ksenəksen
19: linilin
20: baran
21: baran dər
22: baran yor
.
.
30: xriktəŋbar ('three half twenty', 30)
31: xriktəŋbar dər
40: yorbar
50: yorbartəŋbar
51: yorbartəŋbardər
60: xrikbar

200: təŋkuton: half of 400
340: ćəxćəxbar
400: kuton

Numbers much larger than 400 are seldom used, and formed somewhat haphazardly.

As mentioned, ordinals are formed by the dummy noun jaxru. The number of repetitions something has been performed is marked by the bare numeral for intransitive verbs, but with the dummy noun k'abar (derived from the verb 'sidan', take) for transitive verbs.


The noun following the numeral is in the singular for any case but the pegative, for which the number is plural after numbers greater than one.

* bilon and salp are words denoting the four non-thumb fingers or the full palm; sometimes, they are used as numbers.

Dairwueh: Person-Specific Quirky Case

Certain verbs in Dairwueh have no forms for some set of persons, generally one or both of first and second person. Most of these in fact only have third person singular forms. Whenever the subject is "missing" for a verb, it can be formed by having that person's pronoun as a quirky case subject.

The entries below are given as root (meaning) (persons missing) case.

Examples:

mogar (rot) (I, II) acc
ebas (lose, misplace) (I) dat
konav  (bequeath to*) (I) acc
atisal (have sufficient stature to reach something) (I, II) gen
embor (appear to be) (IIsg) loc-instr
adval (displease) (Isg, IIsg) acc
saŋəst (remain) (Isg) loc-instr
lohak (fear) (Isg, Ipl, IIpl) dat

All of these do permit having first and second person subjects. These, however, require oblique marking. The oblique marking will also extend to nouns coordinated or apposite to such a subject. Some examples:

je	moga	-Ø	|	moga	-r
it	rot	3sgI	|	rot	3sgII

it rots

ve	-na	moga	-r
I	-acc	rot	3sgII

I rot

Je, it, is third person, and therefore does not behave in any extraordinary way. Ver, I, however, cannot stand as a nominative subject of mogar, but mogar permits accusative subjects for first and second person, and therefore we get the accusative vena.

These kinds of subjects, unlike the nominative ones, require the 3sg II verb. As mentioned, coordination and apposition do also get affected:

ve	-na	erha	-na	kona	-š	mik	-e
I	-acc	king	-acc	bequeath	-3sgII_neg	nothing	neut_acc

I, the king, bequeath nothing

Normally, erha kona(v/š) would be permissible, but since it's in apposition with ver, which cannot stand in the nominative as subject of konav, it must agree in case therewith.
Contrast with the situation where only the third person subject is present:

erha	-(a)t	kona	(-:v)	kauda	-nu	gir	jera	-lir
king	-gen	bequeath(3sg)	(3sgII)	power	-fem.acc	along	daughter	-dat_fem with dis- similation of -rir

the king bequeathed (his) power to (his) daughter

Note that erha is in the genitive because it's a transitive verb with a definite subject - a slightly ergative pattern in Dairwueh. Contrast with the following, where both a first person and a third person NP is present - the first person pronoun that has to take accusative as subject of this verb also makes the other noun do so:

ve	-na	sa	erha	-na	emb	-e	dəŋak	-un	-əi
I	-acc	and	king	-masc_acc	appear	3sgII_past	fight	active present participle	masc_plur

 I and the king appeared to be fighting/enemies

Contrast to the next clause, where both subjects are third person, and therefore do trigger person/number congruence, and do not have any curious case marking:

erha	sa	kunər	-ta	emb-	ari		dəŋak	-un	-əi
king	and	tribe	-fem_plur	appear	3pl past, indicative		fight	active present participle	masc_plur

the king and the tribes appeared to be enemies

Since embor is intransitive, erha is in the nominative despite being definite.

* The noun to whom something is bequeathed is marked by the preposition gir, 'along, through'.

Case in Answers in my Conlangs

Single-noun answers to questions sometimes in some languages do not necessarily preserve case. If we think of English prepositions as case markers, the two possibilities can be seen below:

Case Preserving:
Q: To where are you going?
A: To Närpes.

Non-case Preserving:
Q: With what are you going to build it?
A: Blood, sweat, tears, nails, wood and a hammer.

My conlangs have slightly differing approaches to responses. Ŋʒädär permits case preservation, but also permits some non-preservation - oblique cases that are not preserved are replaced by the locative, whereas dative, genitive-comitative and complement cases are replaced by the absolutive.

Q: vär xogon t'e-k bürü-ŋö-z
Q: you house what-instr build-fut*-direct
with what do you intend to build a house?
A1: altaŋ-rak
A2: altaŋ-ŋa
A: (with|at) brick

* the meaning of -ŋö- varies with the verb root and with surrounding morphemes.

Ćwarmin requires case preservation, except with direct objects and quirky case subjects: for all of these, an answer in the nominative is permitted.

Q: u kar-ar?
Q: (s)he what-from?where is he from?

A: kirəc-ər
A: from far away

Q: bec kar-ac źarkus-amca
Q: you who-acc meet-recent_past
who did you meet?
A1: Garan
A2: Garan-uc
Garan-(acc)

Sargaĺk preserves case except the pegative, which is replaced by the nominative in short answers. 

Dairwueh preserves case except in situations involving quirky case, where the nominative or the accusative can appear instead, depending on whether the noun asked for is subject or object. One minor exception is that nominative interrogative pronouns with transitive verbs can take genitive nouns as answers, if the noun given for the answer is definite.

Finally, Bryatesle does not preserve secondary case ever in short answers. Subjects and objects can preserve case, but may also be marked by nominative (regardless whether it's subject or object, or even quirky case subject or object that is being asked for). The two other cases can be preserved, or be given in the answer in the accusative case.

Ćwarmin, Sargaĺk, Bryatesle, Dairwueh, Ŋʒädär: Verbs for Speaking Foreign Languages and Babbling

The main approach to forming words for 'babbling' is using a syllable that sounds inherently nonsensical, often with some measure of reduplication:

English: babble
Georgian ლუღლუღი ‎(luɣluɣi), ბუტბუტი (buṭbuṭi)
Greek: βάρβαρος (barbaros)

Note that 'barbaros' basically meant 'someone who speaks incoherently' or somesuch. Similar words can be found in Bryatesle and Dairwueh:

Bryatesle:

klaklan (type II verb)
klaklasi (noun, masculine, 'babbler', 'barbarian')
klaklara (noun, feminine, 'babbler', 'barbarian')
klakyli (noun, masculine, 'babble')

Dairwueh:

xrəxlə- (verb, 'babble')
xrəxlaŋo (noun, masculine, 'babbler', 'barbarian')
xrəxli (noun, feminine, 'babbler', 'barbarian')
xrəxle (noun, neuter, 'babble')

Being less culturally dominant, Ćwarmin, Ŋʒädär and Sargaĺk do not associate speaking a different language with being a barbarian. Klaklas has been borrowed into Ćwarmin as a word for barbarian. 

The Ćwarmin word for babbling (and also speaking foreign languages) is the verb dindin, and similarly Ŋʒädär has the verb vörvör. Both are formed by reduplication of some slightly 'nonsensical-sounding' syllable. In Ŋʒädär, this verb is especially interesting since it is the usual verb for describing the ability of speaking a foreign language:

saɤ brıətəs-rık vörvör-dü*-s
I-abs bryatesle-instr speak_foreign-POT*-1sg_intr
I (can) speak (in) Bryatesle

* this morpheme isn't, strictly speaking, a potential marker throughout the verbal system; the TAM&c system in Ŋʒädär is fairly complicated, with different markers acquiring different meanings in different combinations and with different verbs.

Vörvör can be made transitive - the object then is whoever one speaks to. In the intransitive with a language as instrumental it is seldom used without the potential mood, although it is possible to just omit the potential marker. If the instrumental argument is missing, it just means 'babble'.

Ćwarmin does not have a similar use for dindin - it merely signifies babble or speak an incomprehensible language. Ćwarmin therefore uses the regular verb for 'to speak' when discussing particular linguistic competences, with the language in the instrumental.

Sargaĺk has unique verbs for each of their close neighbours' languages:

tvemarej - to speak Bryatesle (from Bryatesle tvem, 'you')
becarej - to speak Ćwarmin (from Ćwarmin bec, 'you')
soŋarej - to speak Lamen (from Lamen sõq, a very common particle)
erbarej - to speak Dairwueh (from Dairwueh erb-, 'to be')

The Sargaĺk have such verbs for three other minor languages of the area. For general babbling, Sargaĺk has sormoj, which is unusual in not having any reduplication in it.