Monday, March 26, 2018

Detail #377: Number vs. Collectiveness meets Morphology, and a Verb Voice is Accidentally Born

Consider a language which has fusional marking for number and something else. A familiar situation is of course case, but we could also consider, say, definiteness or something else. I will, for easiness' sake, use case for the example given below.

Now, consider nouns that are sort of collective by nature - family, team, group, tribe, etc. Case morphology could maybe conflate some cases' number, or even make them behave morphologically quirkily in such a way that, say, oblique cases are marked like plural nouns, but nominative and accusative are marked like singulars (or even, with the potential for distinguishing plurals and singulars in those cases). 

Now, we could go on a bit and come up with ways of distinguishing many families/groups/etc from one family/group/etc in the mandatorily plurally marked cases: maybe the number 'one', maybe adjectives have singular congruence for singulars, maybe doubling the plural case marker makes for an explicit plural (but a single plural case marker leaves the grammatical number open), maybe rephrasing so the noun is expressed as a direct object permits for the accusative to distinguish singular vs. plural. One could imagine yet another voice there, one that reduces the emphasis on the object, and simply shifts it to the obliques. This could be an interesting voice!

See, we didn't necessarily turn the oblique into an object by a voice operation, but rather by rephrasing. We might've changed the verb entirely, from, say, 'carry a thing (+ an oblique 'towards X' )' to 'approach X (with a thing)'. Now, what this voice would do - and could be used for even in other contexts where this particular rephrasing is not used to enable distinct number marking - would simply consist of making 'approach a place with a thing' be more about 'a thing' than about 'a place'. Which particular oblique takes on 'object-like significance' is somewhat fluid, and depends on contextual cues - definite nouns are more likely to do so than others, animate nouns more so than others, and maybe some oblique hierarchy like instruments > places > times > ...

Could such a 'voice-like' thing have a participle of its own? Possibly, but that'd be weird.

Wednesday, March 7, 2018

Detail #376: Quirky Adjectives

There are some obvious 'quirky' things an adjective could be made to do, e.g. itself be in a strange case or cause the noun to be in an unusual case. However, we can also imagine other things.

1. Case
Certain adjectives could force their noun to be in some case, or at the very least block one case from being marked. Maybe something like
nom → acc
acc → acc
dat → dat
abl → abl
gen →gen
Another option would simply be that the adjective never occurs in NPs of a certain case. This has two possible interpretations: 1) such NPs mark a different case instead, or 2) such NPs simply are never used in positions/constructions that take the forbidden case, and some synonymous phrasing is used instead.

Further, we can imagine quirks in certain case marking positions. Consider, for instance, object complement adjectives, e.g.
I painted the house red
One could imagine that certain adjectives require special marking that other adjectives do not. And here, we could have a slight bit of alignment shenanigans appear - maybe some adjectives, such as 'dead' follow an ergative alignment,meaning that they take the nominative|absolutive when applied to objects or to  intransitive subjects. It's not common in English for transitive subjects to take adjective complements, but maybe your language does.
The example clause of painting a house red suggests to me a different thing; we'll stick to painting for now, I hope the reader is able to re-apply the idea to other topics. Maybe basic and non-basic colours (for some way of dividing colour-space up) take different markings:
I painted the house redI painted the house of orange
The question with regards to the language then is whether this is specific to the combination of the verb "paint" and a set of adjectives, or whether it's just specific to adjectives in that position in general, e.g. would something along this line also have 'of' or not:
I found it (to be) orange
Both ways are reasonable in a language with this kind of marking, and one can probably imagine different subtypes of complements that a conlang could have acting differently, classified by aspectual or volitional or kinetic features or whatever.

2. A Vaguely Alignment-like Thing for a Marker
In e.g. Sami languages, adjectives have a thing that isn't quite congruence, but is not far away from it either. As attributes, they take a suffix, as complements they do not. Thus "the red house" has the marker, "the house is red" does not. We could now imagine situations where this is broken, or even having some adjectives go the other way around, maybe even introducing some kind of 'alignment-like' way it works.

Consider, for instance, adjectives that take this marker whenever they have any kind of NP as complements. I'll use -X as a shorthand for the morpheme in the examples:
the big-X man
"the big man", because big is an attribute.

the man is big
"the man is big", because big is not an attribute
Now, 'afraid' is not really used as an attribute much in English afaict, but let's pretend:
the man is afraid - not an attribute, no nominal complement

the man is afraid-X of spiders - not an attr, but does have nom. compl.

the afraid-X man - a compl.

the afraid-X of spiders man - attr., as well as nom. compl.

to be afraid-X of spiders is a common phobia - nom.compl, but not an attr.

to be afraid is counterproductive - not nom.compl, not attr.

3. A Dummy Head for Adjectives
Let us imagine a situation where adjectives take case marking when attributes, but never when complements. Now, this language has, historically, developed a need for case marking on attributes on occasion, but the restriction still exists. A dummy NP head has turned up that does carry case, though. However, this dummy head is defective, and lacks the nominative. Now, we can imagine situations where adjectives still are needed as subjects, and we can further imagine that, say, the accusative form of the dummy head turns into a nominative-accusative case form. However, we can also imagine a situation where voice operations are used to turn the adjective into a permissible subject without introducing an additional case form to the dummy head.

Monday, March 5, 2018

Detail #375: A Weird Pairwise Voice Construction

This post is adapted from a comment I almost made in a facebook conlanging group, but has been reworked a bit. The first half of the idea was just meant to answer a question about marking something else but subject vs. object on the pronouns, but I went whole hog for full NP marking, and ended up taking a twisty turn towards the end:
So, normally, in a nominative-accusative language, the transitive subject and the intransitive subject have the same marking, and the object a different one. In an ergative-absolutive language, the intransitive subject and the object have the same marking and the transitive subject a different marking.

In some split-s languages, viz. the fluid-s* ones, most intransitive verbs can go either which way - either have subjects marked as transitive subjects or as objects. In different fluid-s languages this is used to mark different things, but volitionality seems popular.

However, this only permits the feature to be marked for on intransitive subjects. Workarounds? Well, voices! With the passive voice, you can mark whether the patient volitionally got acted upon, and with the antipassive, you can permit for the subject of a transitive verb to mark whether the action was carried out volitionally or not.

What if you want both? Well, maybe there could be some kind of "split-voiced" verb, where you turn a verb into two, each with an opposite voice, and each intransitive. Maybe using a special conjunction or a special verb controlling them both, or a special form of the main verb with two independent intransitive auxiliaries:
fight-SPECIALFORM Mark-nominative do-3sg Tom-abs do-3sg
Mark (volitionally) fought Tom (against Tom's will)
Maybe the object and subject have different verbs to make it clearer which is which:
fight-SPECIALFORM Mark-nominative do-3sg Tom-abs stand-3sg
Mark (volitionally) fought Tom (against Tom's will)
 Maybe other arguments as well have dedicated verbs? A different solution already hinted at could be this:
verb-splitter Eric-nominative fight-antip. and Samuel-antip. fight-intr
Eric fought Samuel
Now, you may not always have volitionality implying subjectness:
verb-splitter Eric-absolutive kill-intr and Samuel-absolutive kill-pass
Eric reluctantly/accidentally killed Samuel
The verb splitter may be more conjunction like or more verb like or whatever.

*The other kind of split-s language has the verbs being lexically determined as to whether they take the nominative or the absolutive.

Friday, March 2, 2018

Detail #374: Cases and coordination blocking

Coordination is often a relevant thing with regards to identifying the syntactic structures underlying a language. A question like 'are indirect and direct objects the same thing' can in some languages be decided to be negative because the two cannot be coordinated:
*I gave him and a gift
We also find that we can't coordinate subjects and objects in English:
*I and him hit
However, we can also imagine semantic restrictions on objects, e.g. forbidding something like
he kicked the wall and the man
due to the difference in animacy (or whatever) being too large. Sometimes, we find that different cases can coordinate, e.g. with locatives or certain other obliques:
he is in jail and out of luck
However, these usually are of similar phrase-types, e.g. adverbials or complements or whatnot.

We could go and do weird things with this though. What if the singular and the plural version of, say, the object case, could not be coordinated?
*I saw her and them
*I saw the teacher and the pupils
This would force some periphrasis, or alternatively there'd be morphological cheats - maybe possessive markers hide the plural accusative marker (like in Finnish), and permits it again, so you could say 'I saw the teacher and his pupil(s)'. Or maybe a 'fake pluralizing' strategy emerges, where singulars can have a formally plural but semantically empty pluralization going? A depluralizer of sorts: maybe the numeral 'one' with plural congruence on it?
I saw the one-s teacher-s and the pupil-s