Monday, August 31, 2015

Detail #200: Conflated Participles and Culture

Imagine a language where you have your usual passive and active participles. However, for some verbs, these are morphologically conflated (due to sound change, semantic shift or whatever). Some of these conflations are given cultural significance, i.e. judges are called 'decided/deciding', and this is parsed as them having been decided for their role, as well as their role being that of deciding what is right. 

Some other participles might exist, and if the change that has led to conflation is not sound change-related but rather some odd semantic-grammatic change, you might have other conflations as well, and maybe those too are parsed as telling something significant about the action itself as well as those who the participle pertain to, e.g. maybe the active participle and the recipient participle of 'hunt' is the same, and therefore, hunting is understood as something that is done by the hunter for himself more than for the others that benefit from the game he catches. 

Maybe an instrumental participle and the passive participle of 'strike with a hammer' are conflated, and the fact that tools can be used to make tools is somehow seen as embedded in this.

Detail #199: An Ergative Subsystem in a Language with Absolute Directions

Let us consider a language that otherwise is fairly accusative in its alignment. Intransitive verbs of movement or transitive verbs that impart movement, however, have a somewhat ergative structure.

The absolutive is split up into a few subcases: each one correlating to one of the absolute directions (four or eight or whichever number the language has). The ergative is either the same as the nominative for transitive verbs or an instrumental or somesuch.

Thus, "I go north" is "I-abs-north go", "I hunted it northwards" is "I.erg hunted it-abs-north".

The directional cases are also used with verbs of location; however, with first and second person, and sometimes third person in proximity to the first person, this is omitted. However, if the speaker and listener are exceptionally far in some direction, the directional cases may be used with intransitive verbs (as well as when first or second person is the object of a verb of movement, that sets out to explain how come the person is exceptionally far from his usual geographical range.)

Friday, August 28, 2015

Detail #198: Some Uses for a Case Restricted to Certain Possessums

Consider the following statement:
I have a book
 We could consider having a special case that appears in this, giving us instead a clause of the structure
I am book-POS
Translating this into English as 'I am book-having' could make sense, to some extent. In fact, English almost permits this with a few particular nouns:
she is armed
he is tender-hearted
they are all peg-legged
In English, suffixing -having, or using the past participle(!) of a noun tends to either parse as indefinite or somewhat inalienable possession. In the language we're now creating, we ignore definiteness. We might even exclude -POS from being used with inalienable possession altogether.

However, having a case that is specifically restricted to forming ways of saying 'have' might seem a tad wasteful, so let's extend its use a bit. We might use it whenever possession is changed in some way or other.
I gave you-ACC?/DAT? house-POS
The possession of the house has changed, and thus we mark the house with -POS. If the possession per se doesn't change, we use the accusative instead:
I gave you-DAT key-POS
In this circumstance, the case on the object distinguishes "lend" from "give", or "borrow" from "take". This case could of course also be used with specific verbs like 'marry' (maybe the subjects both are marked by -POS?), and maybe that's the one exception where the -POS case can go on subjects? Maybe subjects can be in the -POS case under a number of circumstances?

Wednesday, August 26, 2015

Detail #196: Pronouns pertaining to Contrary Interests

Consider situations where people together do something, all in order to safeguard their own interest – in opposition to each others' interests. Although it seems unlikely that nouns would have such forms - except maybe a few specific nouns - plural pronouns could imaginably have forms for this, especially subject forms. 'We negotiated a truce' could then have two different meanings depending on which form is used:
weregular negotiated a truce → we had negotiations with them that concluded with them and us reaching a truce
wecontrary negotiated a truce → we1 and 2 negotiated a truce between us1 and 2
This could extend to more collaborative verbs as well:
weregular got married → both of us found spouses whom we married
wecontrary got married → we married each other
Basically the name of the pronoun type could be something like 'preemptively reflexive/reciprocal pronoun' or something. However, since the type of reflexivity/reciprocality is not specified – i.e. there's no, for lack of a better, more specific term, pseudo-resumptive pronoun to tell us the actual role that the subjects take with regards to one another – this creates some nice ambiguity while also dissolving some ambiguity.

Detail #195: Contrastive Pronouns for Enumeration

Consider a fishmonger asking a customer which fish the customer wants:
"this one? or this one? how about this one?"
There could easily develop a slightly productive way of marking listed pronouns:
this, or this, this then, even this, ... → this, oris, thsen, evnis
The same morphemes could affect personal pronouns when ordering several different people, for instance:
you do this, oryou do this, youn do this, evnyou do this
Same goes with third persons. With first person pronouns, however, it can be used to mark sequences of events:
I went to town, ori found a sweet girl, Ithn bought her a drink, evni never heard of her again
After 'evn-', that form is repeated (or maybe the sequence restarts at or-). Derive the pronouns from adverbs, conjunctions and pronouns in your own language, of course. One possible source could also be intensifiers and comparatives, even superlatives.

Monday, August 24, 2015

Detail #194B: Contrastive Dummy

A dummy pronoun with the contrastive case mentioned in the previous post could be a pretty intriguing thing: instead of saying "but X verbs" you'd have "X verbs dummy.CONTR". Which particular syntactic position - subject, object, indirect object, more general adverb - that the dummy takes depends on the argument structure of the verb itself.

Detail #194: Contrastive Marking

Let's consider differential object marking (DOM) along the Baltic Finnic type, i.e. one of the cases is universally used with negative verbs, and also quite often with affirmative verbs - basically, the other possible case marks a combination of things, of which affirmativeness only is one. Now, imagine constructions with contrastive gapping:
I don't eat pork but venison
Let's further assume that the DOM over time is weakened, and the more general case - the one used both with negatives and many affirmatives - turns into a more general accusative. The other case remains, but is used for whatever usages there might be.

However, we return to the form above: the case might have been used to contrast the two objects even when not enough of the usual requirements were fulfilled. So, one usage that gets tied up with it is contrast.
I don't eat pork.ACC venison.CONTR

This might extent to other roles: contrasting with subjects, locations, etc. Contrast might go even beyond negative-affirmative:
I solve problems.ACC and also cause them.CONTR
even beyond that, you can get subjects:
I solve problems.ACC and he.CONTR creates them.ACC
in both of these examples, 'and' could be more idiomatically translated into English as 'but'. However, in the language we're dealing with, 'and' might not even necessarily be marked. In a way, CONTR marks the central argument of a coordinated, contrasting VP.

It might even develop further, to get things like:
I got 99 problems.ACC a bitch.CONTR
 where the first construction we saw also has an affirmative structure - i.e. we're not just contrasting verb phrases with each other, or a negative verb phrase with a contrasting object coordination, we're also doing the same for affirmative verbs.

Sunday, August 23, 2015

The Ćwarmin-Ŋʒädär Family of Languages

Ćwarmin belongs to a geographical span of about two dozen languages with transitional zones between them. Ćwarmin itself is the southeasternmost of these languages – there is a geographical discontinuity from Ćwarmin to the east, but after some nearly uninhabited regions as well as some regions where languages related to Tatediem, as well as some local relict populations still exist, a few non-contiguous zones of languages related to Ćwarmin appear. Ŋʒädär, being the language with the largest population in the whole family beside Ćwarmin, can reasonably be used as the other language by whose name to form a compound term for the whole family. It also belongs to a clearly different branch - one of the three principal branches - of the family.

In some senses, Ćwarmin is somewhat exceptional, in having lost several features characteristic of the family:
  • the inverse system
  • the lack of a distinct accusative case (however, the reflexively possessed suffix does go on reflexively possessed objects in related languages as well), but also an inverse alignment
  • the proximate-obviative system (or rather, it is drastically reanalyzed in Ćwarmin)
  • the front rounded vowels
  • the requirement that non-pronominal roots be at least bisyllabic
  • some prefixing (only traces remaining in Ćwarmin)
  • a rich system of participles
  • morphological differentiation between transitive and intransitive verbs
  • loss of ejectives (due to Dairwuo-Bryatesle influence)
We can notice in fact, that Ćwarmin's unrelated neighbour - Dairwueh - acquired the beginnings of a proximative-obviative system from other, now extinct languages of Ćwarmin-Ŋʒädär stock, but the differences between the systems in early Ćwarmin and early Dairwueh influenced Ćwarmin more. It seems that Dairwueh's population largely derives from groups that have spoken languages closely related to Ćwarmin.

Ćwarmin has kept many features intact though:
  • vowel harmony
  • palatal consonants
  • mostly suffixing (although Ćwarmin has indeed nearly maxed this feature out)
  • a rather simple tense-aspect system (other branches seem to have created more complicated things) with regards to finite verbs
  • separate cases for complements (a separate object complement case is known in roughly a third of the Ćwarmin-Ŋzädär stock, nearly all sub-branches have languages in them that have it, and nearly all languages have traces of it)
  • mainly dependent-marking
  • the paucal number
  • a rich case system (which it in fact has also almost maxed out)

In the farther eastern branches, a few interesting developments have occurred:
  • Parts of the Dagurib branch (insular) has abolished consonant clusters almost thoroughly, and has lost a lot of the case system, and extended the inverse to some rather odd constructions
  • The Ŋʒädär branches only have traces of the paucal
  • The Dagurib have turned the paucal into a dual, but have also restricted it to only appear on nominatives and accusatives and pronouns and a small set of nouns.
  • Ŋʒädär languages have developed a greater amount of adjective congruence
  • Ŋʒädär languages have increased the morphological complexity of the verb significantly, with significant numbers of voices, aspects, moods and so forth.
  • the non-insular Dagurib languages have developed a gender system, and also a gender-based congruence system; the inverse system has been quite strongly affected by the gender system as well.

The Ćwarmin-Ŋʒädär languages cover pretty much all of the arctic region of the world they inhabit. The only exceptions are incursions along the southern border of the arctic region, where groups related to Dairwueh and Bryatesle, Tatediem as well as Barxaw intrude. In addition, four small isolate languages persist in small pockets among the Ćwarmin-Ŋʒädär areas. Three of these are typologically very similar to the ĆŊ languages, whereas the fourth is quite exceptional.

Friday, August 21, 2015

Dairwueh: Is the Diminutive a Derivative Form?

In most of the Dairwueh dialects, the answer to this is clearly affirmative. However, in some dialects in geographical proximity to Bryatesle, a development has occurred.

Bryatesle as noted has a dual case system, with primary and secondary cases stacking on top of each other. The primary and secondary cases are two disjunct sets, so you don't get things like "double accusative" or "accusative dative". You do, however, get things like "accusative partitive" or "dative secondary subject" and the like.

Dairwueh dialects in close contact with Bryatesle have started to acquire traits along the lines of the dual case system of Bryatesle. Let us compare the typical case systems of the two languages:

Dairwueh: nominative accusative dative genitive loc-instr
 This very straightforward system contrasts with the Bryatesle system of
{nominative, accusative, dative, ablative} *
{ definite, possessum, secondary subject, reciprocal object, negative agreement, partitive, suggestion marking}
 However, not all the elements of the product of these two are permitted, and many elements are merged - for many forms, the dative and ablative are not distinguished, for instance.

Generally speaking, the Dairwueh reliance on the genitive to mark possession has not been weakened in any dialects. Some extent of double marking can be found in a few dialects, where the possessum generally gets a somewhat simplified morphology: accusative throughout. Secondary subject and reciprocal object-like markings are slowly appear. Generally speaking, the lexemes that correspond to "themselves", "each other" and so on are getting more syntactically bound than in mainstream Dairwueh and they have also migrated to somewhat unusual positions.

Definiteness is more often marked by determiners than in mainstream Dairwueh, but this has not taken on any case-like behavior at all.

Negative agreement does appear through semantically bleached negative indefinite determiners. These are far from mandatory, and more clearly not syntactically bound just yet. 

The peculiar thing, though, is the partitive: these dialects reuse the diminutive as a partitive case. This also reverses the order of the morphemes with respect to the order in Bryatesle. It also makes the diminutive in these dialects somewhat odd: it has properties more akin to derivative morphology, and properties more akin to inflectional morphology.

Monday, August 17, 2015

Detail #193: A Grammaticalization Path for the Translative Case

In Finnic languages, the translative comes from a kind of locative case (a type of lative). Another locative case (a type of more strict locative) became the essive.

For people unfamiliar with these cases, I'll provide some examples:

The translative marks a noun or adjective that something turns into or is turned into.

the tree

This can also be used in transitive sentences, and in that case, the translative usually pertains to be qualities that the object attain through the action: maalasin talon siniseksi (I painted the house blue).

The essive case marks a role something already has or has assumed:
lääkärinä suosittelen kuntoilua
doctor-essive recommend-1sg exercise-partitive
as a doctor, I recommend exercise
 But this can also have a more temporal significance:
teiniikäisenä harrastin tanssia
teenager-essive practice*-1sg dance-partitive
as a teenager I had dance for a hobby
* no really suitable verb exists to translate it, you need a phrase like "have X for a hobby". 

Counterfactual significance is also possible:
sinu-na e-n hyväksy-isi tommos-ta
you-essive not-1sg accept-conditional such-partitive
I would not accept such a thing if I were you
As mentioned, these developed out of locatives in Finnish, and such locative use can still be found: ulkona, outside, huomiseksi, for tomorrow, tännemmäksi, closer towards here (formed with 'here + comparative + translative'), kaukana, far away, lähempänä at a location closer to something, lähemmäksi towards closer to something.

So, what's another conceivable origin beside locatives? I'd go for verb derivation from adjectives and later nouns. Swedish has -na for turning adjectives into verbs of transition. It's not entirely productive any longer. English has a cognate form, but this is usually transitive: fasten, lengthen, redden, etc. Now, the Swedish type of verb could start being used as infinitive or participial complements of a verb: I got hit reddened: I got hit (and turned) red.

After a while, their use as actual verbs might be lost, and the suffix might start appearing on nouns as well.

Detail #192: Inverse Possession

Give (some) nouns a default possessor (say, 1sg for terms of immediate family; 1sg for 'wife' if 1sg is a married man, else, 2sg if 2sg is married, etc). Have two markers - the direct and the inverse marker.

The direct marker means the noun is possessed by the expected pronoun or noun; the inverse marker moves it to the second rung in the hierarchy. Any move downwards in the hierarchy requires an explicit noun or pronoun somewhere close by in the clause.

Inalienably possessed nouns take either an empty marker or the inverse marker.

Explicit topics are actually ranked higher than other third person nouns or pronouns, but lower than 1st and 2nd person pronouns, no matter what the noun is, so the inverse possession is also somewhat reflexive with regards to topic.

Finally, of course certain nouns whose possible possessor is restricted by biology or social context will vary by speaker - wife being an example I previously gave - whereby depending on who makes the utterance, the interpretation will be different. To make a written language make sense then, the language also will need to have other grammatical cues as to gender and social standing of the speaker, or some learnt assumptions about how to express possession in writing even when the speaker's identity is unknown.

Friday, August 14, 2015

Detail #191: Inverse Aspect

Have some verbs by default be 'narratively subordinated', i.e. they tend to happen during the time of other verbs when accounting for events (and express manners by which something is done, parts of a larger procedure, etc – 'setting the sail' as part of the more general 'travel by boat', etc). Have some verbs be 'narratively leading', i.e. they tend to introduce new events that might not necessarily follow on the previous event - but primarily, the verb is of interest as a salient part of the narrative, a "turning point" or the like. Finally, some verbs are just 'narratively subsequent', i.e. effects of the previous verbs.

Now, each verb is by default one (or two) of these ­ ­ ­– when the verb is used in another aspect, it takes the inverse aspect marker. This does not specify which of the two other (or one other) that the verb is – this has to be deduced from context.

Thursday, August 13, 2015

Person and Noun Class Congruence on Verbs

One thing with person and noun class congruence that could be interesting to develop a bit would be to weaken the link between congruence on one hand and subject and object on the other.

In part this might be due to the subject not really having emerged in the language yet. So, we get a topic-comment like language where congruence might imply that the topic and the verb are "somewhat" closely aligned - i.e. agent, patient, recipient, beneficiary, instrument.

Congruence with another noun implies that the "topic" is some other type of topic - location, associate, "dangling topic", etc. The congruence picks the most salient noun with regards to the discourse that falls within the most 'core' semantic roles.

Different verbs in the same verb complex can have the same agent, yet have wildly different verb congruence:
I think a bear mauled him → I think.3sg.masc a bear  maul.3sg.masc him

do you believe man will ever walk on the moon? → you believe.3sg.masc man walk.inanimate moon?
In the case with the bear, he is the topic, not "I", so he beats out the first person with regards to think. "He" is also the topic with regards to maul, so congruence is masculine. In the second example, "man" is the initial topic, but "the moon" kind of becomes significant enough to cause inanimate congruence for the second verb.

It seems possible that "important" non-agents would be likely to attract congruence in non-discourse-initial verbs, if they are likely to reappear or their identity is very important to the discourse. Thus, verb congruence in such a language serves rather to attract attention to a noun than to inform us of its role.

Tuesday, August 11, 2015

An Auditory Demonstrative

Some news from the world of real linguistics - and the freshness of the information can be gleaned form its content.
Auditory demonstrative in Khaling
 This can easily inspire a bunch of ideas: other senses, how strong the perception is, potentially even "demonstrative tense", i.e. "the thing we just heard", "the thing we heard a while ago", "the thing we're hearing now", etc.

Friday, August 7, 2015

Detail #190: Non-Subject Discourse Topics

Yehuda Falk's book (previously mentioned here) points out that subjects often are discourse topics, and discourse topics often are subjects. Without further defining what a discourse topic is, let us just go with some vague notion of what they are and come up with some ideas for a conlang:
Non-subject discourse topics can appear as null pronouns after prepositions; some syntactical positions have special prepositions (that can go elsewhere than the actual word-order-wise position of arguments that it stands in for) that only ever appear for discourse topics (or rather, there are adverbs that refer to the discourse topic that can stand in for NPs in certain positions)

Another thing one could do is to have mandatorily possessed nouns that have no explicitly stated possessor be assumed to belong to the discourse topic. 

A third thing could be to weaken the link to subjects and strengthen the link to objects, and let any zero subject or zero object be parsed as referring to the discourse topic.

Wednesday, August 5, 2015

Detail #189: Non-subject Imperatives

Let us consider the imperative:
you read the book → [_] read the book!
Usually, if you want to imperativize some other constituent, you need to go by way of voices and moods and so on, and this might be a universal.

I.e. "let me see you", "show yourself". There's no way of going
I see you → I see(objective imperative) [_]
 We could of course distinguish or conflate indirect objects in such a system as well. However, we could do something even more interesting: some verbs, in their imperative form, have the same morpheme for subject and for indirect object imperatives, and a distinct one for direct objects:
you read it.acc → read.IMP_1 it.acc = read it
I read it to you → I read.IMP_1 it.acc = let me read it to you / listen as I read
The thing that distinguishes these two is of course the presence or absence of an overt subject. The second person goes into the "highest" ranking open slot - subject > object, etc.
I read you  → I read.IMP_2 = let me read you (in case the language uses read also for various soothsaying practices, or in case reading some bodily decorations or whatever is a thing in the culture)
Other verbs conflate the  direct and indirect objects, but distinguish the subject:
I build you it → I build.IMP_1 it ≃ let me build it for you
I build you it → I build.IMP_1 ≃ let me build for you
*you build it → build.IMP_1
you build it → build.IMP_2 it = build it!
This is of course a bit like a passive, but differs in not having a patient or recipient as its subject - the agent remains as the subject.

Subjects and their Properties, pt I

This is a really badly structured post, it's essentially just some points taken without much digestion from Yehuda N. Falk's Subjects and Universal Grammar. Turns out there's a lot more to cover than I first thought, so I'm going to give this in two or three rather lightly written installations.

This mainly is based on Falk's aforementioned work. Other works on subjects exist, but for conlanger purposes, a really concise summary of one book might be helpful. I have not seen much about subjects written for a conlanger audience, except the obvious things about ergativity and quirky case.

In the first chapter, Falk first presents a first approximation of subject, and this approximation is quite informative, although he goes on to refine it later on. We start out with this bit:
if a verb has an agent, that agent is the subject
Notice that this does not disallow non-agent subjects: this only disallows that in cases where the verb does have an agent.
the adressees of imperatives are subjects
I.e. if you first order someone, and then state that the person you ordered carried it out, you should be able to do so with that person as the subject of the verb you ordered him to do:

A (to B): Give an example, please!
B: I am giving an example now.
The first approximation also contains this:
in lots of languages, it is more likely for the subject not to be explicitly stated (i.e. a null pronoun) than other constituents
Further, in many languages, reflexive pronouns can only refer back to the subject of the same verb, c.f.:
I saw myself in a mirror
*She received the letter that I sent herself
(if the post office did a mistake, however, "she received the letter that I sent myself" is possible.)
In English, apparently it might be possible to have reflexive pronouns referring to non-subjects:
she made him embarass himself
And even more clearly:
noone can stop him but himself

Control is the term used for the restriction that appears in these examples:
they persuaded the starship captain to kiss the alien woman
* they persuaded the alien woman for the starship captain to kiss
they persuaded the alien woman to be kissed by the starship captain [Falk, pp. 4, 5]
There is no way - short of rearranging the nestled verb phrase so that the non-agent is a subject* - to make certain verbs (believe X to Y, persuade X to Y, etc etc) have X be a non-subject of the nestled verb. ( * this does violate part of the first approximation, but that will be dealt with in the second approximation)

And even further, raising seems restricted to subjects:
It seems that lions eat zebras
Lions seem to eat zebras
*Zebras seem for lions to eat
Zebras seem to be eaten by lions [Falk, p. 5]

Falk also points out that in most languages, subjects are more prone to wh-movement than other constituents are (but English is a clear counterexample!)

Further, many languages only permit sentences with subjects, and in certain frameworks the subject itself is not part of the verb phrase but exists outside it. Subjects tend to be definite.

Things like negation tend to operate differently with regards to scope over subjects than over other types of constituents. The example provided is
a. A student didn't take my course (a particular student did not take the class, and we may expect the speaker to specify why this is a problem or why this student is relevant)
b. I didn't see a student (unclear as to whether it's a specific student unknown to the listener, or whether there's no students at all)

I will continue this with a summary of the refined approximation, which might be followed by a more tidied up summary of the first entire chapter. We can notice there's already some space for conlanging, which I will showcase in the very next post on this topic.

Falk, N. Yehuda, Subjects and Universal Grammar - An Explanatory Theory (Cambridge Studies in Linguistics 113), 2006. Basically everything is from the first seven pages.

Sunday, August 2, 2015

Saturday, August 1, 2015

Concultural Cuisine

Have marinades based on different species of ants "chewing" a little bit on meat, vegetables or food. Adding honey both affects the actual sweetness and how many ants are attracted