Monday, September 29, 2014

An a posteriori idea

I am not a big fan of a posteriori conlangs as a concept, although some of the best conlangs out there are of that kind (Novegradian, I am looking at you). There is one such idea that increasingly has haunted me ever since I visited Iceland.

I have come across claims that I have been unable to relocate (and thus I now doubt it) that in the early 20th century, there were plans to attract more settlers to the eastern half of Iceland. According to my now lost source, one option they considered was inviting a few thousand (or even tens of thousand, not sure on that) Finns.

Now, let us imagine this had happened a bit earlier - Sweden had in fact similarly attracted Finns (then of course subjects of the Swedish crown) to settle various uninhabited forests in mainland Sweden (and even in non-integral parts of the Swedish empire) as far back as the late 16th century. In the 17th and 18th centuries, such groups also settled in Norway, then under Danish rule. Some Finns probably have lived in northern Norway since times immemorial, but a large migration occurred between 1820 and 1890.

So, what if Denmark had invited Finns to settle on the eastern part of Iceland in the middle of the 19th century? What would their language be like today? What would Icelandic be like today?

A few things may lessen the weirdness of the pairing:

  • Both have quirky case, although Icelandic has more of it for objects than Finnish has. Finnish has differential object marking, though!
  • Both have rich morphology, although Finnish is more agglutinating and Icelandic more fusional.
  • Both have rounded vowels, both have diphthongs, and Icelandic preaspiration does not differ all that much from Finnish hC clusters.
Differences abound, though.

It is not a conlang I will ever make, but here's the idea up for grabs! If you decide to go for it, do tell!

Sunday, September 28, 2014

Detail #102: First and Second Person and a Noun Class System

A language with a noun class system such as that of the Bantu languages could imaginably have the first and second person fall outside of the noun class system in some way - quite obviously, first and second person are somewhat exceptional referents.

Thus, one could easily imagine a language with Bantu-style verb congruence to entirely lack congruence for first and second person:

Ta-kulu ta-pade - IV.chieftan IV.rule - the chieftan rules
Ne-kume ne-kpeŋdu tu-hirin - III.village II.prophet - the village hosts a/the prophet
É gbile tu-hirin - I doubt II.prophet
Su ragan ta-hixlo - You hate IV.king
However, first and second person may sometimes interact with noun class marking. This is especially common with constructions amounting to 'as a ...':
gi ta-kulu ta-é inki-i ta-humesxa : as IV.chief IV.I not.(intense) IV.tolerate/approve
'as a chieftain, I do not approve'.
If the identification is hypothetical, the pronoun does not, however, acquire class marking - the verb in these cases acquires an irrealis marking, however:
gi ta-kulu é inki-i an-humesxa : as IV.chief I not.(intense) irrealis.approve
'as a chieftain, I would not approve'
This type of marking can also appear on objects, whenever the object is an object due to some specific thing it is:
ba-kpappi tu-tí  tu-muda f-tu-pade tu-ahid : 3pl(human)-beat II.son gen.II.chieftain II.other
they beat you (because you are the) son of another chieftain

Tuesday, September 23, 2014

Detail #101: Case marking hierarchy

A system whereby which word in a noun phrase carries the case marking for the entire noun phrase is determined by a hierarchy. In this language, all adjectives are essentially nouns in apposition.

Which word goes where in the phrase is determined by various patterns. Examples include:
age immediately precedes title or occupation
terms of kinship are titles, but if another title appears in the same phrase, the kinship term goes after it.
adjectives denoting shape, material or colour (for inanimate objects) indirectly precede titles (i.e. if other things precede them, other things go closer to the title)
personal appearance (beard, hairiness, hair colour, skin colour,) indirectly precede descriptions of age, but go after descriptions of shape
non-title agentive nouns/adjectives go first 
and a bunch of similar rules

Now, let's further come up with a hierarchy were basically each noun and adjective has a rank, and the highest ranked  noun or adjective gets the case marking:

old captain Stanford.acc
beardy old.acc  captain 
beardy old chieftain.acc 
dancing/-er beautiful.acc young
dancing/-er.acc skilled young
The main challenge in making such a language would of course be keeping track of the rules for ordering the words and then structuring words into the hierarchy. I imagine the results could be pretty cool, though.

To really break this idea out of what might seem even remotely realistic, how about having each case have its own hierarchy?

Sunday, September 21, 2014

Barxáw: Auxiliaries

Barxáw has auxiliaries. Some of them may be surprisingly specific in what they mean, and some of them behave in syntactically somewhat odd manners.

Let us first consider the most common syntactical pattern for auxiliaries in Barxáw. 
Nounsubj Aux Verb (Nounobj) (Oblique arguments)
Some cause the main verb to be marked in some manner - a preposed é is not unusual. Considering this extra particle part of the aux would make sense, but some transformations show that é's location is relative to the main verb rather than to the aux.

A number of auxiliaries that follow this pattern are:
áɲ̟è - to be able to
t'íð -  to begin to
jùm (é) - to stop 
kìr (é) - to  do repeatedly or habitually
 err (é) - to stop due to some external factor (the external factor being preceded by the preposition if it is a noun, and the subordinating conjunction ísnu if it is a clause - the external factor need not be specified, however).
 génð [definite human subjects], gél [definite animate subjects], rrú (indefinite human subjects], bém [indefinite animate subjects], kádu [inanimate subjects] - explicitly perfective (also only used with transitives)
ðumú [human subjects], k'ipè [nonhuman subjects] - explicitly perfective (intransitive) 

Adjectival Auxiliaries
Some auxiliaries in Barxáw peculiarly enough distribute more like adjectives than like verbs.
qhúrro -  want to, desire to, hope to
cepá - attempt to, try to cause
azé - resume
c'arr -  refuse to
These are, at times, used like those given above. An interesting quirk is that these can also mark objects (but not obliques), signifying, for instance, that someone wants to be verbed - Séli kelað  qhúrro Joqù = Joqù wants Séli to help him. With azé, marking an object with it signifies returning to a certain object of the activity, whereas the activity itself may have been uninterrupted.
C'arr does never mark non-subjects.

These do not combine with other auxiliaries, and therefore may reasonably be considered auxiliaries in some sense.

Friday, September 19, 2014

Detail #100: Some ideas about counters

Counters are fairly interesting words. If you're not familiar with them, they function a bit like 'heads of' or 'pieces of' work in English NPs along the lines of the following sample:

  • five head of cattle
  • six pieces of metal
  • ten grains of sand

Languages that use such words more often - and where they are semantically more bleached than in the English examples above - sometimes are described as having only mass nouns. Such descriptions may exaggerate what is going on, but I will not get into that right now.

Obviously, counters enable some neat shenanigans with regards to noun-class-like things.
However, I started thinking about other things these words could be used. The prototypical use is basically along the line presented above - whenever the noun phrase has a quantifier, it needs a counter.

A few ideas that probably don't go all that well together:

  • predicative amounts: five CNT house = five house(s), house five CNT = there are five houses
  • plural existential statements: solution CNT = there are solutions. The singular would be done by some construction that is invariant with regards to noun class.
  • distinguishing indefinite determiners - some distinction like some vs. any marked simply by having a counter or not having one.
  • separate counters for specific or estimated amounts. 
  • distinguishing communal possession from separate possession - both with predicative possession and attributive possession. (i.e. 'the family's houses (owned communally)' vs. 'the houses of the family members')
  • individuating plural actors who would normally be assumed to be acting in consort with regards to certain verbs (CNT we go there' - we go there, each by our own path/manner/etc. CNT the family eat = the members of the family all get some nutrition in whatever way, vs. the family eat = the family eats a meal together. 

Wednesday, September 17, 2014

An important point about sign languages

As a little bit of information that some conlangers (as well as journalists and writers in general) may not be aware of:

In general, sign languages are not constructed by committee or by a kind hearing person like you or me to bestow the gift of communication upon the deaf community. The deaf too have the drive to communicate, just like you or me, and they too know what kinds of things can come in handy for communication (get it, handy, drumroll and laughter). This leads to sign languages appearing ad hoc where deaf people happen to congregate, and since some causes of deafness are dominant, there have been communities around the world where a significant minority have been born deaf. Most (ALL!) modern sign languages have developed out of such communal sign languages. Most of those spoken in Europe and the Anglosphere tend to have one particular added stage: when schools for the deaf were first opened, the pupils often originated in a few communities with separate sign languages. In these schools, the languages of the first generations of speakers interacted, affected each other - and one or another won out by whatever reason - an important speaker knowing one of them very well for whatever reason (in many of the communities with dominant genes for deafness, many non-deaf also knew the sign language!) may have tipped the balance, lots of kids from one village may have tipped it, or the status or whatever of one particular specific language may have won.

It is natural linguistic evolution in small communities. Yes, language planning has entered into it - just like it has with regards to Swedish, Icelandic, French and even English. This does not make these languages 'conlangs'.

Attention Conlangers!

An esteemed friend of mine has started a crowdsourced bibliography of grammars. The bibliography can be found here.

If you know of a grammar of some obscure natural language, go and check whether it is in there, and if it isn't, add an entry. If it is public domain and you know an online source for it, even better.

Sunday, September 14, 2014

A Thing about Reflexives and Coordination

How about coordinated subjects with a reflexive object - when only one of the subjects is reflexively referred to? As an example:

John or Eve is going to drive himself to town. (I.e. either John or Eve does the driving, and only in the previous case is it a genuine reflexive.)

Neither these men nor the authorities are going to keep themselves from behaving badly. (I.e. not the authorities doing the misbehaving.)

Is such a half-reflexive permitted in any language?

Saturday, September 13, 2014

Detail #99: A verb/determiner quirk

As a side note, this blog just passed 10 000 pageviews which I feel is not all that bad for a conlanging blog. So hooray for that, I guess.

Let us assume a language with fairly strict word order either along SVO or SOV order (applies, mutatis mutandis, to OVS, OSV, etc). Like many European languages, it also is pretty strict about requiring verbs in clauses. However, certain determiners - especially interrogative ones - serve both as verbs and determiners. This has two effects:

  • it permits odd gaps in sentences, where the "verb-like thing" is in peculiar places and
  • creates some slightly odd situations when there is a verb that normally should be finite verb present as well.
For instance, 'some' also operates as an existential predicate:

 naw maŋra - there is food (some food)

sil əisyl naun maŋran kairril
they us-dat some.acc food.acc give-past-3pl

maŋra vispr-yl-ix mirə
food important is-3sg
(visprylix is formed with a 'derivative dative', which appears in a lot of adjectives.)

maŋra naw kadžylix mipqi
food some good be-participle
(kadsylix too has the derivative dative)

Friday, September 12, 2014

Detail #98B: Onwards with derivative tone

How about if the tone pattern can extend beyond a word, and thus turn an entire phrase into a derived word-like phrase.

láx ŋìmnɛ̀ ťɛn tʝèŋ        → òòoó     →ŋìmnɛ́ t'ɛn tʝén
man carry very quickly                      very quick carrier

Detail #98: A tonal detail

In a language with lexical tone, have some derivative patterns that entirely consist of tones, and which are entirely independent from whatever tone the lexeme previously had.  If the pattern is longer than the relevant word, insert tone-carrying elements at the end (or the onset) of the word.

If the word is too long, affect the last tones by tone sandhi rules (but leave them unchanged except for tone sandhi).

Sunday, September 7, 2014

Demonstrative-quality adjectives in Dairwueh

By adding certain suffixes to adjectives or verbs in Dairwueh, they acquire a quality-wise demonstrative meaning.

Doxs embəl šildrot laxerıŋlo
Need I.dat spear.acc long-thus.acc
I need a spear this long (usually accompanied by visual cues)

Kneel! (like I do) - often, when used with imperatives, either the person uttering the imperative will demonstrate how to, or the action is prompted by culturally significant reasons - such as an emperor or other important person arriving or participation in religious rituals.

It can also have an adverbial function:

Surdunıŋla, kwarša eme vepnətıŋla
burden(v).thus, fight-1sgpast I weak.thus
thus burdened, I fought as weakly as I did/comparably weakly (per some external standard)

It can also have a somewhat intensifying function, although most often preceded by or followed by some kind of thing by which to measure the degree of the quality:

Erbaŋ Kalt krasn.ıŋla - agre.šne əkn ke əkn
is.3sgpres Kalt stupid.thus - understand.not one and one (~one plus one)

"fuck it". After the third person past tense suffix -iŋ, the -ı- part of the -ıŋla suffix tends to be omitted.

Monday, September 1, 2014

Detail #97: Object number congruence

In some language with subject congruence, let also the verb mark for plural objects - and possibly conflate singular object marking with intransitivity.

In addition, make a few culturally specific verbs distinguish dual objects, despite dual marking not existing on nouns. However, the dual marking may also refer to the number of the indirect object - whereas the plural marking never appears due to a plural indirect object.

Verbs with a special dual marking could include 'to give birth', 'to bequeath an inheritance to (one|two|many)', 'to adjudicate between two|many, or just generally adjudicate (with the singular/intransitive marking)', 'to lend one|two|many things to someone, or to lend thing(s) to two people', 'to be the chief of one|two|many military units or civilian villages'.