Saturday, October 29, 2016

Ŋʒädär: Residing and Wearing

Ŋʒädär conflates 'wearing' and 'residing' in one single lexeme. As an intransitive verb, it takes the location, or the clothing as an instrumental argument:
Isheep skininstrumenalwearintransitive
Isheep skin
I wear sheep skin
With transitive marking, it can mark 'to clothe' or 'to house':

the village houses fifteen people / has fifteen inhabitants

the mother dresses (her) daughter
The origin of this verb is the PŊĆ root *kɛŋüst-, signifying 'protect, cover, hold'.

A few nouns are derived from this verb, e.g.
käŋürti - clothes
käŋüsmö - settlement
käŋele - anything that forms a natural protection against weather
There is a Ćwarmin cognate, keŋəc,  signifying 'fabric, cloth'.

Wednesday, October 26, 2016

A Sargaĺk Conjunction

In Sargaĺk, a conjunction corresponding to 'and' exists. However, normally when expressing multiple NPs that do something together, either the comitative is used or apposition suffices.

Apposition is the most common strategy for listing things, but is not unusual with two coordinated nouns either. Normally, the pitch contour of nouns in coordination-by-apposition becomes very static, with a slight drop at the final syllabic core.

Two coordinated human nouns, often at different levels of some kind of definiteness hierarchy or person hierarchy, will call for using the comitative. Thus, first > second > third def > third indef

Finally, the conjunction u is used to express meanings such as
not even
In combination with negation, u often signifies 'not even'. Oftentimes, but not always, the negation will come directly after u, giving upin or upic, but sometimes u ... pic/pin ... appears.

who even cares
A similar thing happens with the 'irrelevant' mood, u ḿte, which can be parsed as 'who even knows|cares'
just not
In the order "pic/pin u" signifies, generally, 'just not'

also, even, too
placing u phrase-finally expresses a meaning similar to the three English words given above.

Saturday, October 22, 2016

Ćwarmin Voice Markers beyond the Passive

Ćwarmin has some more voices beyond the passive and active, but unlike the active voice, all of them use the passive -aśp morpheme for the third person, and lack person marking for first and second person subjects. The main voices beyond passive and dative-passive are the causative (of intransitives), and the causative (of transitives). There is also a detransitive-habitual aspect-voice combination.

Morphologically, there is some overlap between the two causatives. Almost all causatives of intransitives are entirely unmarked in the present tense, as are a handful of causatives of transitives. A few entirely regular such verbs are:
taun - sit, seat
wecəm - smile, make someone smile
gaxar - laugh, make someone laugh
miler - rise, raise
colan - sink, submerge
Some intransitives that are marked in the present causative include
pəlnəm - jump
polnom - jump
kəjn - run
The few transitives that are not marked in the present causative are
taslon - leave, make someone leave
kolkor - carry, load
kolćojn - carry, send
In all other TAMs but the present indicative tense, the intransitive causative has the marker -ka(n)-/-ke(n)-, and takes either no person suffix (for 1st and 2nd person) or the passive suffix -aśp/-eśp for the third person. A more "neutral" way of describing -aśp/-eśp would be as a non-canonic voice marker, which in the absense of other markers is parsed as a passive.

The causative, further, has the causee in either the accusative or the nominative, often depending on the amount of coercion or force involved (nominative indicating less such).

The use of the causative voice is clearly and easily related to the structure of a fact: someone made someone be or do something, and that is what is being expressed. Asking someone to do something, telling them to do something, suggesting they do something, placing them in a situation where they are forced to or even prone to do something is all considered forms of causation.
A further voice marker is a sort of passive causative, which is formed by -kar(u)-/-kər(i)-. This passive causative signifies a few different meanings:
(being made to do something (subject in nominative))
doing something with little volition involved (subject in genitive)
resigning to doing something (subject in dative)
accidentally doing something (subject in genitive, aspect being punctual)
No matter the case of the subject with these forms, the congruence works similarly to the passive: -aśp/-eśp for third person, no marker for first and second person.

Usage of the passive causative is more pragmatically marked, mainly indicating that something is accidental, non-volitional or even somewhat undesirable. Sometimes there is no causer stated, and this basically communicates that there is no causer but that the whole state described by the verb is very untowards according to the speaker or the subject.

There is a further voice, the applicative. It turns a locative of any type or an instrument into a direct object. For transitive verbs, it also permits turning the object into an instrumental, although we also find speakers that permit retaining the direct object as an accusative direct object (thus having two direct objects). The applicatives morpheme is -tak-/-tək-.

With locatives the kind of locative (that is, direction or location) correlates with the aspect of the verb.
hethe doors(a few)deliver (by wagon)

he delivers all the way to the doors(paucal)
This serves to emphasize a non-patient argument or to make it available for certain syntactical operations.

Friday, October 21, 2016

Detail #315: Noun-specific Quirky Case

Since I am a fan of various forms of irregularities, I find the idea of quirky case that only applies to particular nouns with particular verbs to be kind of interesting. However, let's phrase it vaguely as 'nominally conditioned quirky case' and we can use this vagueness to generate even more ideas.

Quirky case possession, for instance, could have certain nouns (or even noun-noun combinations) require their possessors to be marked for an exceptional case, or vice versa, certain nouns could be marked for an exceptional case when possessors (despite having a genitive that as far as everything else but possession goes has the same distribution as the genitive for other nouns), or the whole thing requires two nouns that together trigger the quirk.

Quirky case could of course be extended to nouns with adpositions, but this would sort of just be an extension of the previous two - no matter whether the adpositions originate as semantically bleached nouns or verbs, a previous stage exists to account for it.

Tuesday, October 18, 2016

Detail #314: "Untransitive" Verbs and an Odd Voice

Let's consider verbs with meanings such as 'to be at', 'to stand at', 'to lie at', etc. These verbs could easily take accusative complements without being perceived as really being transitive (or for that part, in an ergative language might fail to have the subject marked for ergative!).

Now, we can start considering what it'd mean for an accusative complement of a verb not to be an object. This would mean we can't coordinate them, despite identical marking, with other objects:
*John wasat and hated school
 This in itself is kind of interesting. However, we could imagine something really weird going on here: using some kind of circumstantial (or whatever) voice to turn these verbs into actual transitive verbs; then, some voice marker could be used to permit something along the lines of
John was being be-at-ening and hated school
where obviously be-at-ening is a pretend form in English, a nonsense thing I use to present this form. This use of voice would probably have to be called something like explicit active or somesuch.

Monday, October 17, 2016

The Ŋʒädär Vowel Harmony and its Development

Ŋʒädär's vowel system is somewhat more complicated than that of Ćwarmin. The following system, in fact, is the basic layout:
iüɯ <ı>u
eöɤ <ə>o


This originates with a slightly different system, the ancestor system of both Ŋʒädär and Ćwarmin:




The symbols used are somewhat misleading: ɒ is not necessarily fully rounded, nor is ɶ. The distinction of e vs. ɛ, ø vs. œ etc was only maintained in stressed syllables. The vowels of the neutral column were also retracted significantly in words with back rounded vowels.

This originally allophonic retraction became phonological with the loss of a number of vowels in various positions, such as unstressed final short vowels after sibilants.
/'gisu/ ['gɯsu] > /gɯs/ sling
/'gise/ ['gise] > /gis/ feather
/'keza/ ['kezɒ] > /kɤz/ thing
/'keze/ ['keze] > /kez/ straw of grass

ser'teʒi/ [ser'teʒi]> /ser'teʒ/ keep for later
/'sɛrteʒa/ ['sʌrtɤʒɒ] > /'sɤrtɤʒ/
Final vowels have thence reappeared due to loss of final -t, -p and -k after vowels. These have thence reappeared due to simplification of final clusters (-nk, -nt, -mp, -st, -sp, -sk, -rk, -rt, -rp). 

Another position where vowels have disappeared is unstressed initial ijV-, ıjV-, ıwV-, uwV-, etc. Of these, only uwV- and ar- affect vowel harmony.
/a'rik/ [ɒrɯk] > /rɯk/ knee
/rik/ [rik] > /rik/ shard
/u'wir/ [u'wɯr] > /wɯr/ dough
/i'wir/ [i'wir] > /wir/ soft
/u'jent/ [u'jɤnt] > /jɤt/ sleep (stem)
jent [jent] > /jet/ eyelid

Sunday, October 16, 2016

A Python HTML Table Gloss Maker

Since making glosses using the blogger editor by hand is tedious and boring, I suddenly realized I'm doing it often enough anyways to benefit from writing a short script for dealing with it. The code is not particularly beautiful or anything, but I figure it might be useful for some conlangers who want to write glosses in html, and just find doing so by hand tedious. Adding css or whatever to keep it beautiful is up to you. 

Here goes: 
 def Tokenzr(T, symbols):
        for sym in symbols:
                T = T.replace(sym, ' ' + sym)
        return T.split(' ')

def TRow(words):
        tr = " "
        for word in words:
                tr = tr + "" + word + "
        return tr + " "

def DefElems(description, elements):
        print(description + ":\n")
        output = []
        for element in elements:
                temp = raw_input("\n" + element + "\n")
        return TRow(output)

T_tb_glossd = raw_input("Text to be glossed:\n")
tokend_T = Tokenzr(T_tb_glossd, [',', '.', '-', '_'])

gloss = "" + TRow(tokend_T) + DefElems("Grammatical glosses", tokend_T) + DefElems("Word for word translation", tokend_T) + ""
By request, a python3 version, which is better whenever you use non-ascii symbols:
def Tokenzr(T, symbols):
    for sym in symbols:
        T = T.replace(sym, ' ' + sym)
    return T.split(' ')

def TableRow(words):
    tr = " "
    for word in words:
        tr = tr + "" + word + ""
    return tr + " "

def DefineElements(description, elements):
    print(description + ":\n")
    output = list()
    for element in elements:
        temp = input("\n" + element + "\n")
    return TableRow(output)

T_tb_glossd = input("Text to be glossed:\n")
tokend_T = Tokenzr(T_tb_glossd, [',', '.', '-', '_'])
gloss = "" + TableRow(tokend_T) + DefineElements("Grammatical glosses", tokend_T) + DefineElements("Word for word translation", tokend_T) + "

The script prints the result out to the console - to me at least that's more convenient than printing to a file. Making a browser plugin seems like overkill for such a trivial task. Just copy-paste and put in a .py file if you want to run it. The ugly abbreviations are just my own conventions.

From now on, I should never be caught making badly layed-out glosses for this blog, at least.

Detail #313: A Noun Morphology (for once without cases!)

Let's consider a language in which almost all noun stems begin on consonants; the exceptions are few enough that the minds of the speakers basically are able to keep track of them pretty well. All nouns belong to a variety of noun classes, which are (sometimes optionally) marked by a suffix.

The interesting bit is a set of five prefixes, a-, e-, i- and u-. These have different functions depending on the definiteness, the number and the class of the noun. Before -r-, -l-, and -w-, a- appears in the allomorph o-, whereas in the handful of nouns that begin with vowels, u- appears as w-, i- as y-, and the others get an intrusive -l-, also causing a- to appear as o-.

The class of thin, long things have these prefixes mark for orientation:
i-  away from the current reference (sg, pl)
u- towards the current reference (sg, pl)
a- perpendicular to the line of sight of the reference
e- moving (sg, pl) , in disarray (plural)
The class of round and irregular things have
i- small
a- large
u- large (very irregular shape)
e- varying sizes (only plural)
Places have
i- close by
a- forest
e- inhabited
u- pasture
Humans have
i- sibling
u- at least one generation older
e- slave or otherwise non-free
a- higher in social status
Animals have
i- small
e- tame
u- wild
a- large
Some of these may have meanings differing by definiteness as well, e.g. foodstuffs have this:
i- a small amount of
a- sweet, fat
u- sour, bitter
e- medical, poisonous

i- a small piece of
a- sweet, fat
u- sour, bitter
e- of ritual importance

Unlike nominal or adjectival attributes, these cannot form simple predicates. You can of course take a noun with the prefix and use that as a predicate. However, a special verb exists that has the prefix appear twice in it:
  These of course take noun class congruence. For these, however, the noun class congruence is sort of half-way derivational: saying something like 'my brother is far away' would use the place-class congruence marker as well as the human-class marker.

Saturday, October 8, 2016

Ćwarmin Names

Ćwarmin names come in several patterns, depending on regional and family traditions.

One common tradition is using suffixes on 'name stems' to express the significance of a child:
-ot | -ət
'oldest son'. Whether it goes by mother or father differs by region, but most regions go by father. In some families, it is replaced by
-sako | -səke
upon the death of the father.
-otkom | -ətkəm
A younger twin of an oldest son

A female twin of an oldest son
Outside of the core cases (nom, gen, acc, dat), -ot/-ət is omitted, but the case marking follows the definite paradigm as far as possible.

-otol | -ətəl
'oldest living son'. Changes upon the death of a previously oldest living son; in some regions, goes to oldest living daughter once if no sons remain). In regions with the -sako/-səke morpheme, this is also replaced by -sakol/-səkəl upon the passing of the father.

-oxan | -esən
'daughter born after the death of her father'

-anko | -ənke (son)
-asko | -əśke (son)
-olku | -elki (daughters)
'a son or daughter born significantly later than other children of the same mother'
Patronymics likewise have some doublings of the same information present.
The usual patronym for male offspring is name-stem+julor, from julo, son. The patronym for female offspring is [name-stem]+-ćəŋer, from ćəŋel. However, oldest sons get -julot, as do oldest living sons. Female twins of an oldest son gets -ćəŋet. -ćəŋsən is applied to daughters born after the death of their father.

Toponyms often appear in the general ablative in full names.

Ćwarmin naming is not very solidly set in stone, and differs by what is needed; several small, distant communities basically never use the patronymics, because no confusion appears anyway. Communities involved in trade routes, military raids, or even urban living use increasingly complex names:
toponym patronym [optional descriptive term] name
The optional descriptive term can be basically any adjective or profession. Some other nouns appear at times, but are unusual.

Friday, October 7, 2016

Detail #312: Conjugations, Declinations and Exceptional Lexemes

Consider a language along the style of Latin, Russian or Finnish, where nouns and verbs belong to conjugations or declensions. Each conjugation or declension provides somewhat different patterns for inflecting your nouns and verbs. Usually in languages such as these, a noun or verb belongs to one in particular, although dialectal, *-lectal and even historical variation may exist. A lexeme might also be somewhat ambiguous as to what particular such class it belongs to.

However, you can also imagine lexemes that in most normal use belongs to a particular class, but have individual exceptions – a certain case or TAM+person or whatever behaving like the lexeme belonged to a different class.

These are fairly plausible exceptions and can be found in many languages. However. A thing that I am unaware of any careful description of are semantically conditioned exceptions. I.e. a word that belongs to a particular class, except for a specific set of phrases including that word - for instance, sausages belong to declination III, but in a particular compound or with a particular attribute that turns it into a specific type of sausage, the lexeme "sausage" is suddenly inflected as though it belonged to a different conjugation. Referring to sausages of that particular type even without the compound or attribute may also trigger the exceptional inflectional pattern.

A couple of things one could do with this:
  • restrict this quirk to only some specific case or number - say having the nominative plural behave exceptionally for the specific phrase, but no other case or number; or have it go for all the plurals?
  • generalize it as a kind of derivational tool for those lexemes belonging to a particular declension or conjugation
This kind of detail is an easy way of giving one's conlang a kind of 'realistic texture', the kind of 'lived-in' feeling of, say, Finnish, German, Russian, Latin or Georgian.

The reason, btw, why I used the noun "sausage" as an example is that some speakers of Southwest Finnish have that exact distinction, apparently. Makkara has the plural stem makkaroi- in the plural for most sausages, but a few specific types of sausage have makkari- in the plural.

Wednesday, October 5, 2016

Detail #310 pt 2: A Better Explanation of the Idea

Idea #310 might've been slightly unclearly expressed, so here's a further elaboration.

A Coordinate System
The coordinate system above corresponds to the combinations of elements of two sets. These can be anything - we could, for instance, imagine them to be [Persons] and [Tenses, Aspects and Moods].

This need not be a two-dimensional system, it could as well be something like
A Cube; This time, in glorious 90s colours.
One dimension might encode person, another tense, another aspect.

Since we're dealing with a very finite number of combinations - say, this time, that we're dealing with number * case * gender or something like that - we can conveniently enough flatten this cube by mapping the elements of two of the dimensions onto one dimension, returning us to something like the coordinate system above; we need to arrange it so that one dimension is subordinate to the other, though (e.g. in the multi-dimensional representation, each dimension can keep its elements in the same order everywhere: 1, 2, 3, ... always come in that order; however, if you have a subordinate and a superordinate set of dimensions, A3 may come before B1, despite 1 < 3, if the value of A is lower than the value of B). We get this happening:
Some neighbours are preserved as neighbours: any two that are in the same column in this example, will retain their distance in the new presentation; any two that are in the same row will have their distance multiplied by four. If distance isn't interesting, this is no problem, and even if it is, it's not necessarily all that big a deal, so we'll ignore it for now. We should be aware of it, though.

But now we'l get to an interesting thing: in morphology, we have two 'spaces'/'planes'/whatever. One is the plane of possible combinations of morphemes, the other is the plane of possible combinations of meaning.

Canonical agglutination, if such a term can be used, would refer to the following situation, or higher-dimensional analogues of it:

There is a perfect correspondence between combinations of morphemes (the left coordinate system) and combinations of meanings (the right coordinate system). We find in some languages, though, that this is not the case! We can come up with a lot of things that could be going on, and this image with several forms mapped to meanings should illustrate some possibilities:
A system of correspondences that has been distorted in several ways.

Some of the most common things in real-life languages probably are meaning-conflations (several meanings correspond to one combination of morphemes), morpheme-conflations (several morphemes express the same meaning). Direct twists might be somewhat unusual; however, if a twist/cross exists, I find it likely that more than one pair has a similar cross/twist going, and it's of course imaginable that the twist has more than just one pair of elements involved.

Monday, October 3, 2016

Detail #311: Distributive Participles

Distributiveness can be an interesting feature of verbs and nouns alike. Although I have had this interest in going for more verb-centered, less noun-centered systems for a while, I still find huge verbal morphologies somewhat awkward - they can be pretty clever, but I really don't think I have what it takes to pull a nice huge verb system off.

However, finite verbs is not the only place we can throw stuff at and hope that it sticks – infinite verbs can be just as useful places to stick things, and in many languages, different infinitives and participles appear often enough in all kinds of constructions.

So, let's consider a particular distinction with regards to verbs - that of whether a plural subject or object participates in the role as subject or object in a concerted, group-like manner or in a distributed, non-concerted manner.

Consider, for instance, the difference between
the men of the tribe hunted the mammoth down
the men of the tribe drink tea
The first  we can guess implies the men did so together; the latter, each man may be drinking tea regularly at home, or it's conceivable that they gather somewhere and drink tea together. Conversely,
the man trampled the bugs
the man hunted hares
Trampling bugs probably refers to a single event of bug-trampling where they get trampled in one concerted go, whereas hunting hares seems to consist of multiple separate events of shooting at hares.

So, we can create a semantic distinction there, one that slightly also accords with
  • telicity - telic with plural subjects or objects is more likely to be concerted than atelic with plural subjects or objects
  • perfectivity and perfectness (by the same argument)
  • habituality - habitualness can kind of imply concertedness or not with regards to subject - people doing things together regularly; however, with regards to plural objects, habituality is almost never concerted, though a concerted habitual verb with a singular subject is imaginable.
  • 'groupness' of both subjects and objects
Let's imagine now that the number congruence marker on participles for certain verbs can be replaced by another marker that signals both the (semantic) plurality of the subject or object, and whether the action is perceived as a single overarching action or as multiple iterations of an action. Maybe markers exist for both, but some verbs prefer the overarching action-parsing, and some the multiple-iterations parsing, and thus select the other marker whenever the dispreferred meaning is intended.

This might take on a slightly derivative meaning if participles and gerunds are conflated.

Also, I hear the previous post was a bit unclear, so I should probably write an explanation with actual examples.

Saturday, October 1, 2016

Detail #310: Topologies of Morphologies

Often when making agglutinating morphologies, it's easy to end up making a fairly Cartesian-product like morphology. One way of avoiding the "excessive regularity" of this could be not to introduce irregularities in the formation - to in fact keep it exactly that way, but introduce oddities in how parts of the thing stick together.

Consider the number of features your morphology marks for - also include zero marking for any feature as a distinction. Make a table where you project all the dimensions of the morphology space, in the following way:


If a combination of markers of features - say 2bIIIג - really just combines the meanings of features 2, b, ג and III - we're dealing with a topology that corresponds to an entirely flat surface.

We can, however, make this surface more interesting; we can add a twist to it:


I like throwing ugly colour combinations your way; yep, that's pretty much the only conceivable explanation.
Now, one could imagine a similar twist happening both in 1 and 2, and of course it could just as well appear in some way for the hebrew-letters or for I/II/III. Basically, this 'twist' could appear in any dimension, e.g:


This twist is also sort of 'tube-like' - each highlighted element in 2a is offset by one in the {I,II,III}-dimension, if we assume the entire structure "turns around" so the right edge of III is glued on to the left edge of I. This isn't necessary, though - the offset might be more like a braid-shaped twist.

I am not going to try and impose the little topology I've learned from youtube here, but I think this kind of thing might give a sufficient idea about possible ways of turning cartesian products into interesting and creative things whose structure is less trivial.