Adjectives as Nouns in Ćwarmin

Ćwarmin's equivalent to 'the X one' (where X is an adjective) has a little split - half the time, it's as easy as inflecting an adjective for a case and definiteness; the rest of the time, it's less easy. 

Ćwarmin only permits adjectives to stand independently as nouns in the definite and specific forms; an indefinite NP requires a nominal head. For the definite and specific forms, just take an adjective and add the def/spec case marker on it. For indefinite adjectives, there must be a noun that is a dummy head of the NP.

There are several nominal heads that can be used for this, and here the patterns that determine their usage is described.

The main words used for this purpose are

kalć - stick, branch, plant
mirgə - board, cloth, skin, cover
ogmo - stone
taxkar, taxŋar - assembly of parts, construction, something that is built out of smaller parts
voram - belly, torso (of persons)
garnun - body part
səkve - land animal
kic - fish or sea animal
loma - bird or bat
verći - account, words, story, sum, plan
mokmo - action, story, outcome, result
yulzər - assembly of people, bunch, gang, congregation
yulzvonar - horse-mounted assembly of people
semtə - flock of wild, non-carnivorous animals
roŋ - flock of wild, carnivorous animals
kiŋre - flock of tame animals, fortune, boon

There is also a pair of very vague words, erkar, erter  (lit. one-who, one-what) which can be used for almost anything. However, many of the words above are extended by some form of metaphor, so kalć, for instance, covers nearly any tool or implement, ogmo likewise covers many tools and implements, but also building material (even wood!), kic covers anything spotted moving in water (including boats!), loma covers anything in the sky, and yulzvonar covers any group of people ready to fight.

With case forms that lack definiteness, leaving these out is basically syntactically well-formed, but for most Ćwarmin speakers, if the nominalized adjective is semantically indefinite, it requires a noun, if it's specific or definite, it does not. Thus, definiteness marking with independent adjectives is obtained by absence of a dummy noun, which might be typologically unusual.

Detail #258: A Really Weird Voice

To come up with bizarre voices, one needs to come up with bizarre restrictions that make them reasonable. One imaginable restriction that could make such a thing reasonable would be this: subject pronouns can only refer to previous subject nouns, whereas object pronouns can refer to subjects as well as objects.

Consider then something like:

I wrote a book. It flopped.

"It" cannot refer to 'book', due to the previous restriction. So, there's a voice that moves subjects to objects, and leaves the subject spot empty. Thus:

I wrote a book. flopped-BIZARROVOICE it.

This voice could of course also take on other uses: change the scope of negation:

a student did not arrive can only mean a specific student did not arrive. However, did not arrive-BIZARROVOICE a student can mean 'not a student arrived', i.e. no student arrived.

The voice could imply lack of volition, or somesuch?. It could further have the regular object either be demoted to some kind of oblique position, or maybe permit for double (and even triple?) objects. Maybe a subj→obj→ind.obj→obl hierarchy exists. A quirk on that could be that things only are demoted if they are pushed, so e.g. an indirect object only turns oblique if there's a direct object that is demoted.

Detail #257: 'Final' Numbers

Consider a special marking on numbers that is used when finishing counting. This replaces both ordinals and cardinals.

Thus

... twenty-two, twenty-three, twenty-four, twenty-five, twenty-six.[end-morpheme].

Possible origins: markers for perfect or perfective aspects, or maybe reduced verbs (a structure expressing the idea that now, 'twenty six (pieces have been) counted').

Another source could be as simple as 'and' or a word like 'last'. And finally, maybe congruence with the noun class of the counted things only appear on the final number, and thus serves to signal completed counting.

Finally, some kind of superlative marking could work? Potentially, 'last' could already have a superlative marker in it (consider how English 'last' ends in what probably is the superlative morpheme, much like how the un-cognate word 'sist', with the same meaning, has the superlative morpheme in Swedish) - so a reduced 'last' could bring with it the superlative morpheme anyway.

But without bringing 'last' into it, maybe the final number is considered the greatest of the bunch, and just therefore it gets the superlative marker. This would be pretty neat.

Keep in mind that not all languages have comparatives and superlatives, though.

ANADEWs #1: Chukchi

In Chukchi, the absolutive case - which is pretty much what you'd expect from the label - is formed by root reduplication.

I guess if a conlanger did that, we'd all be having at him or her for being unrealistic.

Personal Pronouns: Ŋʒädär

Ŋʒädär has a relatively simple personal pronoun system:

	I		II		III
	sg	pl	sg	pl	sganim	sginan	sgobv	pl	plobv
abs	saə	dat	vär	ür	ehi	is	əqı, isqi	sint	ənqı
poss	saəxom	darom	växem	ürim	ehem	is	əqlıs	sinim	əŋrəs
dat	saəm	darot'	väm	ürüt'	ehem	isim	əqım	sinit'	ənqım
loc	saŋa	daŋo	väŋä	üŋö	ehma	isiŋa (-iŋa)	əqŋa	sinŋä	ənŋə
abl	sana	dalo	välö	ülö	ehnä	isli	əqla	siŋlä	əŋla
lat	saus	daus	väüs	ülüs	ehlüs	islis	əqlıs	silis	əŋlıs
com	saəs	daras	väräs	üräs	ehles	islis	əqlıs	silis	əŋrəs
instr	sarək	darək	välik	ülük	ehlük	islis	əqlık	silik	əŋrək

I should probably ask for forgiveness regarding the colour combinations I have used for the table headers.

There are a number of minor irregularities with regards to the usual nominal case morphology. There's also a case distinction that is missing from most other nominals - the possessive vs. comitative distinction. The interested reader might try to ferret out cognates to the Ćwarmin pronouns.

Detail #256: Onwards with #255 – "Inverse Imperative" due to Social Rank

Consider, again, the imperative. In a highly hierarchical society, lower classes would very seldom give commands to higher classes. It would be economical to use the same markers though - so, for instance, 

direct imperative mood:
high → low: imperative
low → high: asking for permission to do something

reverse imperative mood:
high → low: granting permission to do something
low → high: imperative

One point I am trying to convey here, in part, is that asymmetrical inverse systems might be pretty cool.

Detail #255: Inverse Mood (for a few modal auxiliaries)

Let us consider a mood like 'indicative'. For subjects of certain persons, the indicative might be the most common mood, but for others it might be less so – simply put, you don't often need to ask questions about your own doings in the past, whereas you more often might need to ask questions about what the second or third person did in the past. With the second person, it is somewhat common to give commands, but it's less common to command oneself. A lot of similar observations about moods not being equally distributed for each person can probably be made.

While having a full-on inverse mood system might get a tad unwieldy and maybe even unnatural, there could be some default modal implications with regards to person in some circumstances (e.g. maybe certain subordinating conjunctions trigger the modal implication). Let's have "do" be the inverse auxiliary:

I have a thing	(indicative)
you have a thing	(indicative)
... that I have a thing	(indicative)
... that I do have a thing	(optative)
... that you have a thing	(optative/imperative)
... that you do have a thing	(indicative)

Of course, it could also be possible for the verbs themselves to have a lexically determined default mood, or even for verb-person pairs to have lexically determined default moods.

Unlike systems such as inverse alignment or inverse number, this wouldn't thoroughly permeate the system, though, but only appear with certain pairs of moods and maybe, as mentioned previously, in certain limited contexts.

A final potential thing here - maybe do in the imperative would imply non-second person. The reasoning would be that second person is the usual recipient of commands, and so first and third person receiving commands would require the inverse auxiliary. From this, we get the inverse auxiliary also acting as a person marker in certain moods.

Reflexives in Ŋʒädär

Ŋʒädär has a reflexive pronoun, 'ŋul'. Its inflection is as follows:

	sg	pl
absolutive	ŋul	ŋulɤr
dative	ŋulum	ŋulɤt'
locative	ŋulŋa	ŋuŋa
lative	ŋulus	ŋuvus
ablative	ŋuluk	ŋuvuk
gen-com	ŋuruk	ŋuruk
complement	ŋulȝuv	ŋuvȝuv

The direct-inverse system of Ŋʒädär of course involves itself in the whole affair. Now, it should be fairly natural that ŋul can never be a subject - much like you cannot say 'myself saw me' for 'I saw myself' in English. Ŋul occupies a spot in the hierarchy, and thus nouns that are lower than ŋul need to have their verbs marked for inverse, whereas nouns that reside higher than ŋul have their verbs marked with the direct marker.

1st singular >
1st plural >
2nd singular >
2nd plural >
1st, 2nd, and 3rd person animate dative subjects >
3rd animate proximate >
dative objects and ŋul >
3rd animate obviative >
3rd inanimate prox >
other non-nominatives subjects and objects >
3rd inanimate obv

A complication arises in the plural, however. It turns out ŋulɤr appears in two places. You get things like

sint ŋulɤr p'arab-z
sint ŋulɤr p'arab-jut
they selves protect-[direct/inverse]

It turns out these two reside at opposite ends of the whole hierarchy.  When it's at the top of the hierarchy, the verb has to have inverse marking, and of course vice versa when it's at the bottom of the hierarchy. (Thus, we can know which of them it is by looking at the marking on the verb.) There is a difference in meaning:

sint ŋulɤr p'arab-z: they protect themselves
sint ŋulɤr p'arab-jut: they protect each other

Thus, ŋulɤr signifies reciprocality with the inverse, and reflexivity with the direct. Thus:

ŋulɤr (reciprocal)
1st singular >
1st plural >
2nd singular >
2nd plural >
1st, 2nd, and 3rd person animate dative subjects >
3rd animate proximate >
dative objects and ŋul >
3rd animate obviative >
3rd inanimate prox >
other non-nominative subjects and objects >
3rd inanimate obv
ŋulɤr (reflexive)

 As for the case forms, ŋulɤt' and other obliques rank as other obliques, and do not distinguish reciprocal from reflexive. The plural genitive, however, moves its whole noun-phrase to the top or bottom of the hierarchy, and makes the NP behave analogously to ŋulɤr - just rephrasing the possessive structure either as 'each other's Xs' or 'their(refl, shared) Xs'.

Sargaĺk: The Present Tense

Sargaĺk verb morphology, like its other morphology, is agglutinating. The person-number-gender markers are as follows:

	1sg	2sg	3sg m	3sg f	1 pl	2pl	3pl m	3pl f
Conjugation I	-əi	-i	-ju	-(ĺ)tua	-(k)in	-va	-mu	-nu
Conjugation II	-i	-	-i	-(s)	-i	-u	-n	-n
Exceptional Morphemes	-əta	-ita	-juta	-suta	-(k)in vs. -sim

The difference between conjugations I and II depend on the position of the secondary stress in the verb; if it's on the person morpheme, the verb belongs to conjugation I, otherwise, it belongs to conjugation II. Normally, the stress sits on the first syllable, thus making the third syllable carry secondary stress. However, if the first syllable is closed, the second syllable carries stress, and the final syllable is reduced.

Sargaĺk has a handful of verbs that distinguish inclusive and exclusive first person in their verb forms. Here is an overview of some verbs and the forms that are used for the distinctions.

jarnap – to inherit

jarnakin = we (incl) inherit
jarnasim = we (excl) inherit

karb - to be forbidden to do

karkin = we (incl) are forbidden from doing
karsim = we (excl) are forbidden from doing

jesal - to be permitted to

jeslin (...)
jesim

vak'am - to gain from

vak'in
vak'sim

rosk'ir - to be related

rosk'in
ros:im

The exceptional morphemes for 1sg and 2sg provide forms for when the verb is ditransitive. The third person exceptional forms are only used when the object or the indirect object also has higher animacy.

A Pattern for Case on Infinitives in Ćwarmin, Ŋʒädär, Sargaĺk, etc

In the languages of the arctic regions of the continent on which my conlangs reside, the following trait holds, except for the Tatediem tribes that have moved into the area fairly recently.

nouns:
subjects → mainly nominative
arguments → other cases
obliques → other cases

infinitives:
subjects → mainly nominative
arguments → nominative
obliques → other case

With other argument types, a number of arguments may have non-nominative cases. The most obvious examples would be the accusative, which in Ćwarmin marks the direct object, the pegative, which in Ćwarmin marks the subject of a ditransitive verb, and the dative, which in Ćwarmin and Ŋʒädär marks the indirect object. Several of the normally oblique cases also can appear as arguments of specific verbs, such as:

kəc-ite ostanc-ap dart-u
wall-DEF.NOM storm-instr withstand-3sg
the wall withstands storms

From Ćwarmin. 'ostanc-ap' is not an object - it does not pass objecthood tests; however, it is not in an adjunct relation to the verb either, as it does not pass such tests either. For instance, 'ostanc-ap' is not a permissible answer to 'kəc-ite terce dart-u', 'how does the wall withstand?', whereas adverbs of manner, instrument and similar are permissible as answers to 'terce'. For one, it's integral to the verb – although the withstood thing can be omitted, it still is implicitly there, and without an implicit noun phrase to which the listener can bind a null pronoun there, any finite sentence with dartan is malformed.]

äinäyi tunt'a-: k'ıʒo-lus
old.woman enjoy-intr game-to
(the) old woman enjoys the game

From Ŋʒädär. The intransitivity marker on the verb should indicate that k'ıʒo is not an object or subject, nor an indirect object. The case marking suggests a spatial reading but this is obviously not the case either. This also neither passes all adjuncthood or all objecthood tests, although it passes more objecthood tests than it passes adjuncthood tests. Similar things apply here as with the verb dartan in Ćwarmin.

Ərges simi p'arima-mai tuxa-ju
father son tradition-with teach
the father teaches his son traditions

From Sargaĺk. P'arima-mai is not a regular object, since in that case, Ərges would be in the pegative, and you'd have *Ərges-ta simi p'arima(-mai) tuxa-ju. Similar restrictions as in the previous two examples apply.

However, we find that all three of these can take naked infinitives:

kəc-ite atosćun dart-u – the wall withstands (for someone to) attack (it)
äinäyi tunt'a-: kılıs – the old woman enjoys singing
ərges simi garjir tuxa-ju - the father teaches his son to ask permission (before entering places)

However, other circumstances where these cases would be used, the infinitive would require the  case as well, such as:

pər źil-ic kopon-ap śapr-u
man nail-acc hammer-instr strike-3sg
[Ćwarmin. The instrument is not an argument of 'śapran'. It would be a reasonable answer to the question "pər źil-ic terce śapr-u".]

pər-ite kəc-itiś arnjan-ap əmnit-i
man-def wall-acc.def hurry-instr build-3sg
the man builds the wall quickly

Now, we need to consider the notion of argument here. I distinguish, for these languages, two kinds of noun phrases in a clause - arguments and adjuncts. To understand the difference, we need to look at the mental record that a "verb template" is. Here's an example for some randomly imagined verb:

	Action:	(verb)
	Subject:	Agent
	(Object:)	Patient
	Oblique object:	Destination
	⋮

So, the verb as different types of arguments, and each argument is a thing in this template. A location can be part of the template, but it needn't - all verbs can be carried out at locations, but only for some is the location part of the template. In some, the object might be filled out with a location, etc. Some verbs may have multiple permissible templates.

What happens in these languages is that an infinitive is assigned to whichever semantically reasonably likely slot in the template that it could imaginably fill, and there you go. Nouns, however, get case assigned by the slots. Adverbials are not assigned case by any system like this, but by a more 'universal' template, and therefore, any noun phrase that adds information in a clause beyond what the templates enable will follow more regular patterns. This 'universal' template has a wider range than does the verb template, and ends up forcing case marking onto infinitives outside of the range of the verb's 'gravitational pull'.

Ŋʒädär: Accusativity in Subclauses

Although Ŋʒädär mainly deals with relative subclause-like structures by means of participles, complement clauses are not dealt with by using non-finite verbs, but rather by having particles that introduce subclauses. The verb in such a subclause generally uses some subordinate moods, e.g. subjunctive or conjunctive, depending on some semantic factors. Although the conditional keeps the inverse alignment going, the subjunctive and conjunctive have a more accusative-like thing going for them.

First observations: the subjunctive and conjunctive keep the same tense markers as the indicative mood. Subjunctive affixes -p'An-, conjunctive -čEm- at the spot where the direct/inverse marker would otherwise have gone. The direct object takes the dative case, whereas the subject remains in the absolutive.  Generally, the word order will be SOV, although SVO or VSO is attested. Passives are formed using the intransitive marker -lU/-lE after the mood markers. For the passive, either of the subject or object can be omitted, but they will retain their case marking - thus it's not a real passive. The object is not promoted to syntactically subject-like status.

Two Papers: Indexical Shifting

A friend made me aware of this article on 'indexical shifting' in indirect speech. E.g. contexts where 'I' can mean '(s)he'.

Another paper that deals with some other examples, and some even more exotic behaviors regarding index shifting can be found here.

Dairwueh: The Passive

In addition to its active verb forms, Dairwueh also has passives. In the present indicative, it is formed by affixes, but in other tenses and moods it is formed by auxiliaries with participles, much as in English. This table has been inserted in the Dairwueh tables of morphology:

passive, present	-ŋor	-ŋor	-ŋa	-ŋ	-ŋan	-ŋan	-ŋa
Periphrastic Passive forms
passive, neg. pres.	erb- + passive neg participle
passive, irrealis	ŋey- + passive irrealis participle
passive, past	ŋe- + passive affirmative participle
passive, neg.	erb- + passive negative participle

The form given for the periphrastic verb form's auxiliary is given as the 3rd person II form, although each form of the corresponding row in the table below can be used depending on the person of the passive subject. However, the 3rd person II form can also be used with other person subjects under certain circumstances, such as

Beyond promoting the object to subject, the passive has a few other uses as well: some intransitive verbs can be passivized to mark lack of volition. Some intransitive verbs, such as shine, stand out - əkšat, appear to be - aruas, deserve - kivankan, sweat - unhən, guess - iltad   only have passive forms.

Finally, most verbs of emotion are passive, and have the "object" as an oblique form, e.g.

keŋa tsayŋor
him-instr hate-1sg.pass
I hate him

xovit eirŋor
you-dat love-1sg.pass
I love you

The copula (erb-, ger-, dir-)

The auxiliaries in the table are forms of the copula, which has not yet been described in any post. The copula is highly irregular, having lots of different stems popping up in rather different positions. In tabular form, we would have:

	1sg	2sg	3sg	3sgII	1pl	2pl	3pl
present indicative	bras	er	erbaŋ	erb	gan	guše	guni
present irrealis	ger	ger	giŋ	ŋey	gran	gras	gran
present negative	dir	divne	dirne	dirš	gruš(en)	gruš(en)	gruš(en)
past positive	gis	gerb	giŋ	ŋe	grad	grabe	gari
past negative	edirš	edirš	edirš	ediš	greyšin	greyšin	greyšin

These are all the finite forms of the copula in Dairwueh. The participles are formed from the form diral, which also is the infinitive.

Detail #254: A Restricted Case, with Noun Transitivity

Let us imagine a case that appears on personal pronouns and which signifies 'according to whose opinion'. So, e.g. me-accord = by my opinion.

Let's then go on to some special situations. The third and first person are oftentimes indicative, whereas the second person generally speaking is interrogative unless very specific cues as to its indicativeness are present. One such cue is if it directly follows (alternatively precedes) a perfective verb:

I wrote your-accord the letter
I wrote the letter as per your opinion regarding how it was to be written

your-accord this is good?
is this good, in your opinion?

If there is a missing argument, there is sort of an implicit 'what':

your-accord we do now?
what should we do now, in your opinion?

In this circumstance, the by-your-opinion is always clause-initial (if the language does wh-fronting). 

Further, these can serve as a somewhat adpositional things, giving, for instance,

plan-acc your-accord is?
what is your opinion of the plan?

elections-acc his-accord are superfluous
by his opinion, elections are superfluous

However, pronouns are  incorporated, giving a double possessive structure:

its-his-accord: by his opinion of it

Detail #253: Mixing Alignments - Split-Ergative/Inverse

I remember back when ergativity was sort of new and unfamiliar to most conlangers – today, split ergativity is almost old hat, and people are getting into alignments like inverse, secundative and so on.

Let's however consider split-alignment again for a moment. How do we resolve subjects when we coordinate a transitive and an intransitive verb?

man wolf see-DIRECT and run-INTRANS
[assuming man > wolf in the animacy hierarchy)

Is it the man or the wolf who runs? If the language underlyingly is accusative, it would be the man. If the language underlyingly is ergative, it's the wolf who is running. Now, the split-inverse/ergative option sort of appeals to me - and in this case, we'd have the man running if we did

man wolf see-INVERSE and run-INTRANS.

Now, how about some other options? We could have direct and inverse apply to intransitives in a way similar to switch-reference (but limited to intransitives after transitives), thus giving us:

man wolf see-DIRECT and run-DIRECT
man sees wolf and runs

man wolf see-DIRECT and run-INVERSE
man sees wolf and it runs

Another option could be that a subject of a previous verb automatically is temporarily shifted upwards in the hierarchy. Depending on where in the hierarchy it lands, different results obtain, e.g. if 1 p > 2 p > previous subject > 3p, then

I wolf see-DIRECT and run-DIRECT
I see the wolf and I run

I wolf see-INVERSE and run-DIRECT
the wolf sees me and I run

I wolf see-INVERSE and run-INVERSE
the wolf sees me and runs

I wolf see-DIRECT and run-INVERSE
I see the wolf and it runs

he wolf see-INVERSE and run-DIRECT
the wolf sees him and runs

he wolf see-INVERSE and run-INVERSE
the wolf sees him and he runs

Detail #252: Having (Optional) Morphemes follow Alignment-like Distributions

A thing that could be fun is having (optional) markers for some modal information (or some other thing pertaining to the whole verb phrase), whose distribution follows alignment-like patterns (these markers affix to nouns). A language with several different patterns like those could be pretty cool, e.g. some markers are ergative-like in distribution, some accusative-like, some nominative-like, some absolutive, some maybe secundative, etc.