Sunday, July 31, 2016

Ŋʒädär Person Morphology

Beyond its direct/inverse marker, Ŋʒädär does have some extent of person marking on its verb. There is, however, some extent of optionality to the markers.

Ŋʒädär has polypersonal verb agreement - it marks for both the subject and objects. Objects and subjects are not morphologically distinguished per se, neither in the nominal morphology nor in the verbal agreement system. Thus, the two dimensions of the agreement systems are not full cartesian products along the lines of (persons x numbers) x (persons x numbers). Since we have dealt with reflexives previously, we also can tell that there are no elements along the 'diagonal', with the exception of obviative/obviative
The columns represent the lower ranked element, the rows the higher ranked element. The inverse/direct morpheme tells which of the two is the subject.


1 sg2 sg3 sg prox
/ intr.
3sg obv1pl2 pl3 pl prox3 pl obv
1sg
-sAvU--sA--sqO--zdA--zvUr-tAs*--tAsqO*-
2sg

-U--UqO--dvU-Ur--Us--UsqO-
3sg prox
/ intr


-ε- **-h(I)qO--dA--Ur--s(I)--s(I)qO-
3sg obv


-qO**--dA--UrqO--s(I)qO--OqO-
1pl




-dAvUr--dAs--dAqO-
2pl





-Urs--UrQo-
3pl prox






-s(I)qO-
3pl obv







** the ε symbol signifies the empty string; for 3sg proximative, the form only permits an intransitive parsing; the 3sg obv / 3sg obv both permit an intransitive or transitive parsing. The 3sg. prox and 3sg. obv are also used with intransitive plural subjects.

* the initial -t of the suffix sometimes assimilates, along the following lines.
nt → n
lt → l
rt → r
st → s
qt → q
kt → k
pt → p
ft → t
mt → t
jt → j
ht → t
Voiced stops and voiced fricatives tend to turn into voiceless fricatives:
bt → ft
zt → st
dt → st
gt → xt
  ...

Saturday, July 30, 2016

Interrogatives in Ŋʒädär, pt 1

The basic interrogatives come in two basic forms, corresponding quite closely to 'who' and 'what':
kaɣo, təɣə
the animate form has a plural absolutive as well:
kaɣant
Other than that, the case forms are conflated for plural and singular throughout the pronouns' paradigm:
dati.  kaɣam, təɣəm

lative kaɣlus, təɣlıs
locati. ka
ɣŋa, təɣŋo
ablat. kaɣluno, təɣlınə

gen-c. kaɣas, təɣəs
instru. kaɣak, təɣək
compl. ka
ɣuv, təɣıv
A similar pair of pairs of lexemes with similar morphological forms exist for 'pick one out of many' and 'pick one out of two'-questions. Both pairs have the same animate-inanimate distinction at their core, with the animate given to the left in the next lists.

One out of two:
abs. k'opo, roto
dat. k'opom, rotom
lat. k'oplus, rotus
loc. k'oŋa, roŋa
abl. k'opluno, rotuno
g-c. k'opos, rotos
ins. k'opok, rotok
cmp. k'opuv, rotuv
One (or more) out of several:
sg. abs.
camu, təmıt
pl.  abs
camant
The rest of the inanimate are identical to the what form, whereas the cam- forms are formed analogously to the other animate forms given in the tables above. -ml- turns to -vl- in most dialects, but to -mn- in the easternmost dialect.
Ŋʒädär has two words for 'when',
ɣok'nu, when, (future)
ɣ
ok'oś, when, (past)
these also have demonstrative analogues:
əqnu, then, (future)
əqoś, then, (past)
Unlike the very absolute future/past distinction of the interrogative pair, the demonstrative pair seem to showcase a rather relative future/past distinction.

Thursday, July 28, 2016

Comparison in Ŋʒädär

In Ŋʒädär, the adjective for which two NPs are compared is in the instrumental case:
pöntü-rük ('with rough', 'with coarse', in comparative structures: rougher, coarser)
 However, it is also inflected like a verb:
lesnı-rık-ta-jut '(s)he is faster than you'
If both nouns are third person, the lesser one is marked by a circumposition, Un-[dative]-bI.
To compare non-subjects - something like 'I like her more than him', one would rephrase it as 'I like her, she un_him_bi like_ptcpl_instrumental_(3sg)_direct', or 'me.dat she un_him_bi like_pass.ptcpl_instrumental_(3sg)_direct'.

For even more oblique comparison, such as 'it's better at home than in the forest', one would construct the sentence as
soman* un orvur-(u)m bı maba-rak-s
home (than) forest-dat (than) good-instr-intransitive
For oblique comparisons, there is no way of using the inverse and direct to compare nouns of different rank - the circumposition is necessary.

Tuesday, July 26, 2016

Detail #301: Lexicalized Paths becoming Verbalized

Most readers probably are aware of satellite-framed and verb-framed languages. This idea takes verb-framing to eleven – by introducing way more lexemes.

Let's consider a culture where a variety of terms for different types of paths exist. A significant number of specific, real-world roads, paths, sea lanes, and waterways in general have specific proper nouns.
So, basically, verbs of movement often derive from proper nouns. These have affixes expressing
  • movement towards major location along the path 
  • movement towards personally significant location along the path (home, ritually important place, hunting grounds, etc)
  • movement away from a major location along the path
  • movement close to that path, possibly zig-zagging over it
  • movement up- or downstream with rivers, or uphill/downhill with very steep paths
Whenever movement along such a path is expressed, it is grammatically mandatory to use the path-specific verb. If no path is known, or the utterance refers to movement along paths in general, the verb is derived either from the four cardinal directions or more generic path-types.

The language doesn't let you walk, it lets you be moving along a named path, and maybe with an optional adverb that expresses 'walkingly'.


Monday, July 25, 2016

Detail #300: Suppletive "Half-Gender" Congruence

Consider a language with two genders, ostensibly masculine and feminine. Have another distinction that almost creates a four-way gender system:
{masculine, feminine} x {animate, inanimate}
or
{masculine, feminine} x {human, nonhuman}
Now, let's have some marker that goes on verbs sometimes (maybe, say, only in the present tense, or maybe only in realis, or whatever, the details are not so important). However, we get some verbs having a suppletive form for only one out of the four combinations:

animateinanimate
masculineroot1-root1-
feminineroot1-root2-
Each gender/animateness combination may be the one to get the exceptional form for some given verb. 

Of course, another thing can also happen: suppletive roots for the animate/inanimate distinction, but gender congruence according to masc/fem.


Saturday, July 23, 2016

Sargaĺk Possession

Attributive possession in Sargaĺk has a few small complications:
  • the possessor is in the pegative-genitive case
  • if the possessum is in a locative case or the comitative cases, it remains in those cases
  • if the possessum is in the pegative case, it remains in the pegative case
  • if the possessum is in the nominative case, it will turn into the comitative or the familiar comitative case.
There are two basic ways of expressing "X has Y". The first, and most common with animate nouns has the following structure:
subj.peg1 pronoun.nom2 object.nom2 is
Thus, the owner is the subject, and the possessum is represented by a personal pronoun (agreeing in gender with the possesssum), and a noun phrase, the direct object, that is the possessum itself.
Sometimes, the pronoun agrees with the subject, and most speakers seem to grasp this as meaning the same thing.

This is probably analogous to how the ergative in many languages can be used in constructions along the lines of
noun1.erg noun2 is
for meanings along the lines of noun1 has noun2. The extra pronoun serves to make it ditransitive and thus license the use of the pegative case. 

The other construction uses a dedicated verb, k'ir-. This is common with inanimate nouns, abstract nouns, and with an adjective for object, it expresses some command over a quality - an ability to control or use a quality.
Xivar c'oman k'ir : Xivar has a lot of endurance
Beyond this, k'ir in combination with an infinitive expresses ability:
Osini falməs k'ir-m
Osini read has.fem
Osini knows how to read
Literacy is very unusual in Sargaĺk villages, and so Osini would typically function as the village's record keeper and in an almost semi-diplomatic fashion when interacting with Ćwarmin and later Bryatesle-Dairwueh officials.

Sunday, July 17, 2016

Sargaĺk Participles and their Use

I previously outlined the semantics and the morphology of the Sargaĺk participles. This post sets out to describe their syntax and their pragmatics.

Participles often serve a role comparable to relative subclauses in English. However, the strategy used to express similar ideas usually is one of using different styles of subordinated coordination: when the x was swimming, rather than the swimming x. The three voices of the participles enable some relative-clause-like structure for subjects, objects and recipients of the embedded verb.
Topics tend to prefer the use of separate, subordinate clauses.

An exhaustive list of "general" traits of which one to pick would be:
  • "heavy" embedded VPs are more often rendered as subclauses than by participles
  • "heavy" clauses with an additional subclause that could be rendered by a participle often have participles rather than an additional subclause
  • topics prefer subclauses
  • focus prefers participles
  • certain participles form compound-like lexical structures, and these form an inseparable unit with their head noun. These generally have very full case congruence, and their evidentiality is inferential*. Examples include
    • "xalval ecdo", sweating house or house that sweats, essentially 'sauna'. xalval is not a gerund or anything like that; to some extent, this is a sort of implicit causative, 'house that sweats (you)'.
Beyond the use of participles as attributes in NPs, we have the use of participles with auxiliaries. Participles can be combined with the copula to form statements a bit like "X is verbing Y". The binary copula is only used when emphasizing a positive answer to a yes-no-question. Other than that, the copula-construction is almost always used when the subject is the focus. Thus, a focused subject enforces evidentiality marking in its VP.

Objects and recipients too can, by voice markings on the participle, be marked in this way, but this is somewhat less common.

A handful of auxiliaries require participles:
sanət - be [reputed to/inferred to/seen to/heard to/...] verbThis requires the active form of the present or past participle. (Uninflected). This auxiliary basically emphasizes the content of the evidentiality marker.

xk'arp- - resume

mər- continue

mərmər- continue despite attempts by others to stop one from doing

tomŕ-
cease (participle in the ablative)
Further, a lot of verbs lack some participle forms, e.g. verbs of perception often lack  non-primary perception evidentiality forms, verbs of verbal interaction often lack everything but second-hand forms, etc. For some of these, this is more of a conflation of forms, for others it seems more like an actual gap. Lack of past or present forms, or of passive or recipient forms is also not unusual.

Finally, there exists an adverbial case that only exists for participles. This signifies by doing. Its marker is -(k)o, which also reduces the previous syllable's vowel if morphophonologically possible. This conflates a fair share of forms as well.

Sargaĺk: Adjective Congruence and the Copula

The congruence bit of this post is obsolete.

The Sargaĺk copula is somewhat complicated – it is both morphologically defective, in lacking several forms that most other verbs have (specifically, it exclusively has indicative and a basic irrealis form. It lacks imperatives, and any other modal distinctions are just omitted). It also is morphologically extended, in having several forms that few and even no other verbs have. Some of the forms do not historically speaking derive from verbs, but from pronouns and participles, but are nevertheless syntactically and morphologically verbs by now.

Adjective Congruence


Sargaĺk has some congruence on its adjectives.


masc
fem
sg-ti
-da
pl-er
-so

The above table marks the agreement marker in nominative and accusative noun phrases. The following set of markers appear with the pegative:

masc
fem
sg-ta
-na
pl-sa
-sair
Beyond this, the only "normal" congruence markers are -er, which appears for all animate oblique case-number combinations, and -i, which appears for all inanimate oblique case-number combinations.

The case markers used on nouns do appear on adjectives at times as well - including a zero marker for feminine and masculine nominatives. This invariably happens when the adjective is the head of an NP, i.e. constructions analogous to English '(a|the) ADJ one'. This also serves to intensify the adjective, or to mark topicality of the NP or to attract attention to the adjective. This specifically may happen when demonstratives are involved.

Congruence

There are two copulas: one for clearly binary qualities or memberships of sets, one for qualities with degrees to them. Whether a quality is considered binary or not is very culturally determined - gender is binary, as is being asleep or awake. Being a father is not binary, but being a mother is; being a male is binary, being female is not. This goes with both nouns and adjectives, so this is in a sense another two-way division of the noun/adjective space in addition to gender and animacy.

Colours are generally not binary, except eye colours. Darkness of hair is binary, but light hair colours are considered binary. Hunger vs. satedness is binary, illness is not binary. Deadness and liveness is binary. Etc.

(The order for the verb forms given below is 1 p., 2 p., 3 p. masc, 3p. fem, the upper row being singular, the lower plural)
The two verbs are as follows:
binary:
present:
k'iʒ | k'ip | k'ir | k'iva
k'iko | k'iyo | k'ivo | k'ivo

past perfective:
fem:
sg: ak'o
pl: ak'yo
masc:
sg: ak'ə
pl: ak'e

past imperfective:
k'aʒa | k'apa | k'ara | k'ava
k'avi | k'aya | k'ava | k'ava
non-binary:
present:
əvin | əvi | əvir | əvo
əko | əvyo | əvo | əvo
past:
an | avi | avir | ava
aki | ava | ava | ava
neutral:
future*:
k'əvk (singular)
k'əvka (plural)
 
The perfective-imperfective distinction in the past binary form is unique to the copula. The future is not fully unique, although its formation for the different verbs that have it is not very regular at all.

(1pl has a thing where -i is an old inflection that appears in some verbs in the past tense.) A peculiar thing with the two copulas is that if the complement is a noun, and thus has intrinsic gender, the congruence marker for adjectives will appear as a suffix on the verb, giving forms such as k'iʒda, əviso, etc. With adjectives, there is no such congruence on the verb, but the adjective does show the gender-number congruence. These two verbs are the only verbs to show gender-number congruence with more than one constituent; we will later, however, find verbs that have congruence not with the subject, but with some other constituent – and for these, the congruence morphemes are the same as for the complement congruence here.

Causatives of the binary version imply a more perfective causation, whereas causatives of the non-binary imply increasing something's quality as something or other. For the causatives, the subject congruence is dropped altogether.


Friday, July 15, 2016

Sargaĺk: Participles and Evidentiality

NB: This is partially a draft, and may still be changed.

Sargaĺk has a system of participles. These are used both as adjectival attributes in NPs, and as complements as times. They are also required by certain auxiliaries. There are a fair share of participles, and we need to classify them to speak about them.

First, we have different voice and transitivity concerns: we have passive participles, recipient participles and subject participles. However, the subject forms differ depending on the voice and the typical transitivity of the verb. Beyond this, the participle codes for evidentiality.

Two distinctions need to be explained here: primary and non-primary relate to the sensory organ by which this knowledge has been acquired. Primary is whichever is expected for the verb or the NPs involved - a stench would be known by smelling, a sound by hearing, and most things by seeing. Non-primary then is using any other sense. "Recipient-hand" and "subject-hand" is knowledge acquired second-hand from the recipient or the subject.

As can be seen, there is some overlap between the three transitivities. The overlap is not the same in both of the tense-aspect variations. Congruence markers can cause some morphological changes as well.

Imperfect aspect or non-past tense

typical transitivity:voice:evidentiality:




ditransitiveactivefirst-hand primary-al


first-hand non-primary-saŋ


second-hand -sur


inferential-sapud

recipientfirst-hand primary-tŕ


first-hand non-primary-tŋ́


second-hand-tud


"recipient-hand"-təvŕ


inferential-təpud

passivefirst-hand-sək


second-hand-səŋək


inferential-sənul
transitiveactivefirst-hand primary-al


first-hand non-primary-saŋ/-tŋ́


second-hand-tud


"subject-hand"-təvŕ


inferential-təpud

passivefirst-hand-sək


second-hand-səŋək


inferential-təpud
intransitive
markers as ditr.
recipient





Perfect aspect and past tense
typical transitivity:voice:evidentiality:




ditransitiveactivefirst-hand primary-jir


first-hand non-primary-jir(u)


second-hand -surem


inferential-sepem

recipientfirst-hand primary-trem


first-hand non-primary-teŋm


second-hand-rem


"recipient-hand"-trem


inferential-tremud

passivefirst-hand-səm


second-hand-səmək


inferential-səmem
transitiveactivefirst-hand primary-jir


first-hand non-primary-jir(u)


second-hand-rem


"subject-hand"-trem


inferential-sepem

passivefirst-hand-sər


second-hand-səmək


inferential-təpem
intransitive
markers as transitive
passive






Thursday, July 14, 2016

A Public Service Announcement

I've been writing some longish posts for a while - a "post-mortem analysis" of Tatediem and Barxaw, some historical linguistics of Sargaĺk, Dairwueh and Bryatesle, some ditto for Ŋʒädär and Ćwarmin, some sample texts for Bryatesle, as well as the verb systems of Ŋʒädär and Sargaĺk, as well as some derivative morphology for Ćwarmin.

I might be exaggerating a bit when I call the Tatediem and Barxaw post a post-mortem analysis, since they both do have life in them - both Tatediem and Barxaw need significant changes, though, and such will happen in the upcoming winter.

Anyways, here's a short sample of a sample of Bryatesle:
Tsarmuvex kulpity ɕmargɕity xistvëmxi xarda, ɕixue ɕvaraɕ kama. Tërsi ɕtekan lrexmud kerfeklër ɕxarda. Nədvər, ednën rufal, tatsek. Anzïmub teka du kar iskrar. "Bəcək nədvərśkə xuršupuru", ednën rufalïr tersëk furaven ixutë.
Ćwarmins are insulted by signs of human decay and death, such as worms and rot. Insulting certain rituals (insults) them too. Baptism, first washing, head (among these). Insults often join the two: "Bəcək nədvərśkə xuršupuru" - in your first washing, the worms swim.
A small challenge, for anyone interested: there should be enough online to be able to figure out most of this text's grammatical structure - the vocabulary should be possible to figure out from the translation. A good glossing would be a sufficient enough achievement.

Sunday, July 10, 2016

Detail #299: An Idea for a Morphophonology

So, I have been planning this morphophonological system for Sargaĺk, originally, but have abandoned it because it just didn't really fit with the imaginary history of the language.

The idea is that morphemes do not consist of strings of phonemes (although phonemes are also permitted to occur), but by strings of underspecified phonemes. Thus, you can have a thing like
passive: {+back}l{-occlusive +voiced}
Depending on the phonological context in which this appears, phonemes will be picked to realize these features using some set of rules. The fun appears when multiple morphemes appear together. Also, "weak brackets" would be ones that permit for merging with the previous/next phoneme. A hierarchy of features would determine how the resulting set of features ends up. Let us use {/ and /} for such brackets.

Each underspecified phoneme could of course maybe have a 'strongest' feature that could be written in bolds or as the leftmost feature or whatever, to which special rules apply. Coming back to the passive marker presented previously, we could suffix the following morpheme:
first person: {/ +semivowel}a
Thus we get 
{+back}l{-occlusive +voiced}{/ +semivowel}a → {+back}lwa
What maybe happens there is that the backness present before l velarizes l, which leads to +semivowel coming out as /w/ rather than /j/. 

Of course, at times a combination of features just might not be realizeable, say e.g. {+occlusive/} {/ -occlusive} or {+velar /} {/ -velar }. For such conflicts, a set of rules need to exist. At other times, the interpretation need not be entirely literal.

Sargaĺk: Detransitivization

Much like other languages, Sargaĺk has some amount of detransitivization going on. It also has deditransitivization.

First, the canonical detransitivization procedure in Sargaĺk usually involves adding some morpheme to the verb. There are exceptions, however, where some stem changes occur instead:
tŕp'ə- - to spin something
tĺp'ə- - to spin (reflexive)
kar- - to carry something
kačń- - to hold up, to stand while burdened, to withstand

fakń- to bring something
fačń - to reach
Sargaĺk, like English, has verbs that are ergative and verbs that are accusative. Above, tŕp'ə-/tĺpə- is an example of an ergative verb. Another example, where both a stem change and an inflected form exist, can be seen here:
ŕma- - to raise
ĺma- - to rise
ŕman- - to rise
(c.f. adjectives ĺmi- 'ascending', ŕmi - high)
There are a number of transitivizing operations as well. Most ditransitive verbs get -an-, but some verbs do not get such a marker. These include typically ditransitive verbs such as
ops- - to give
id- - to show
- - to sell
p'rik- - to pass someone something
lonkə- - to tell
k'əda- - to throw someone something
p'əspə-
- to teach someone something
t'ošni- - to confirm something to someone
vəšni- - to negate something to someone
partə- - to stand as someone's representative to someone
ĺvoʒa- - to betray someone to someone
kŕvoʒa - to betray someone to someone
The relation between intransitive, transitive and ditransitive verbs gets slightly complicated at times. Certain ditransitive verbs, when they lose their pegative subject, promote the direct object to subject, some promote the indirect object to subject. The non-promoted argument can be omitted. A few such verbs are these:
opsopil promotes object to subject
ididil promotes object to subject
vŕ → vril promotes object to subject
p'rik → p'rikil- promotes i.o. to subject
lonkə →lonkə (!) promotes object to subject
k'əda → k'ədil promotes object to subject
t'ošni → t'ošni (!) promotes i.o. to subject
vəšni → vəšni (!) promotes i.o.
partə → partil
promotes object
lvoʒa → lvoʒil promotes i.o.
kŕvoʒa → kŕvoʒil promotes d.o.
As can be seen, -il- is the suffix that marks the omission of a pegative subject. Not all verbs can be made to lose their subject by the addition of -il-, however.

Other verbs lose the object whenever the marker -il- is on them. There's generally no detransitivization marking whenever an indirect object is omitted. The detransitivization promotes the i.o. to object if present, so one can have what looks like a d.o. present even after -li- is applied.

Verbs for which -il- omits the subject can have their direct object omitted by a slightly different strategy: the presence of -an-. -an- normally marks the presence of a pegative noun phrase, but when only two NPs (usually in the nominative) are present it signals the omission of a (usually direct) object, and that one NP is also the proper subject. Usually, the other NP is the one that -il- would promote to subject with that verb.

Thursday, July 7, 2016

Bryatesle: Marginal Phonemes and Marginal Allophones

Bryatesle's phonemes basically form the following two tables:

bilabialdental(post)alveolar /
alveolopalatal
palatalvelarglottal
stopp bt̪ d̪t̠ d̠ [c] [ɟ]k g(ʔ)
fricativeɸ βs̪ z̪ɕ (c͜ç)x (ɣ)(h)
laterals
l̠ʲ


trills

r̠ʲ


nasalsmn̠ʲ


semivowels(w)

(j)


Marginal phonemes are present in brackets. In angular brackets, you have c, ɟ, which have been lost as such, but are realized as k g x ɕ r depending on context.
The dental consonants are only distinct from the postalveolar ones when in front of front vowels, but are explicitly written in all positions - morphophonological processes can make a front vowel enable the distinction again.

 The glottal stop appears in hiatuses that coincide with word boundaries. In some urban dialects ʔ replaces most word-final voiceless stops, but this is frowned upon. It also appears in the word paʔɛɕ, mustard seed. The two semivowels appear in one word, wujɨp or wujip, bundle. Orthographically, wujɨp is represented as uβuɟɨp. In some dialects, -βu- is generally pronounced [wu], however, in most of the Bryatesle area this is at most a marginal allophone, expected of people with odd idiolects and speech defects and of children.

[h] appears in complementary distribution with /x/ in clusters with other fricatives, laterals and the trill (so /lx/, /sx/, /zx/, /rx/ -> [lh], [sh], [zh], [rh]) but it appears exceptionally in the word ɛhɨl̠ʲ (fem), exhalation, release of tension, relief. This is probably due to loss of l̠ʲ in the underlying l̠ʲxɨ-, exhale, but all other comparable situations have the -x- appear as [x].

/ɣ/ appears in three interjections and in one verb. The verb is paɣa-, to smoke fish or meat (with the near-minimal pairs ɸaga-, snore, axa-, to place a longish thing at an inclination towards a wall or other relatively tall thing, to be placed at an inclination towards a wall or other relatively tall thing (when used of a longish thing as subject). Orthographically, this appears in several forms: pagxa, paga, paxa, paxβa, paβxa.

/c͜ç/ appears in ɨc͜çu, the onomatopoeic representation of a sneeze. It also appears in the verbal form ɨc͜ça-. Beyond this, it also appears in some baby-speak, and words related to babies including words such as n̪ec͜çe (m), milk, and buc͜çi (f), cradle. Of these, only basically buc͜çi has any widespread use when babies are not part of the situation. The written forms use cɕ or t̠ɕ.

c and ɟ are not as such phonemes, and are thus never heard. However, their realization is strongly conditioned by morphophonological concerns. If one of these has appeared in a position where there never are morphophonological changes, most authors would use the letter corresponding to the expected realization, but some authors that are influenced by old texts may use these instead.
{c} appears in a velar form when close to back vowels, and as /ɕ/ when close to front vowels. Between back and front vowels, the vowel after the consonant determines what form it takes. {c} is realized as the stop /k/ before stops and nasals if preceded by a back vowel, as /x/ before other consonants, and as /x/ after all non-nasal consonants. After a nasal consonant it also appears as /k/. {ɟ} basically has a similar distribution, but with /r̠ʲ/ standing in both for /x/ and /ɕ/, whereas /g/ stands in for /k/.

i    ɨ    u
  ɛ (ɘ)(o)


  ɑ
In the vowel system, we see an even more disconcertingly large relative number of marginal phonemes. /o/ appears in the words 'flu', ɸomsa (written ɸamsa, sometimes ɸumsa) and 'goat kid', βobɨm (written fvubim). ɸamsa is a distinct word signifying a species of fish, and thus we even have a single minimal pair for the word. There's also the interjection [o:], which basically is the Bryatesle equivalent of 'er'. In other contexts, [o] and similar almost never appear. Using such phones as realizations of either ɑ or u is frowned upon. It appears in some dialects that are considered very rural, and also associated with certain disapproved religious movements.

ə appears in a few interjections, such as ələ, which appears after mistaken words and thus before corrections, gə/xə which correspond to ew, and əb! which expresses disbelief and disagreement. It also appears in the verb xəre- (to shave)

The dental consonants are only distinct from the postalveolar ones when in front of front vowels. 

Wednesday, July 6, 2016

Detail #298: Differential Serial Verb Constructions

Serial Verb Constructions are a thing I have not said a lot about this far, and they're a thing I really really should learn more about and come up with ideas about. Now, guitarplayer approached me with an idea over irc, an idea based on a certain set of possibilities that serial verb constructions have.

There are different possibilities with regards to agreement in SVCs. Some languages seem to require them, some languages do not. Limited congruence is also conceivable - say, number but not person, gender but not person, etc. 

This is a place where we could sneak in differential convergence. Having different types and amounts of of convergence between the verbs could encode differences in TAM. A thing I find it fairly probable to encode as well could be the amount of volition the subject had.

Serial verb constructions are a thing I definitely should post more about, since they are pretty unusual by European standards but common elsewhere. However, I should probably read several books on them before any further writing about them.


Shameless Self Promotion II

Besides coming up with imaginary typological features and working on my conlangs, I also compose music. Over the last several years, I've been mainly focusing on xenharmonic music, i.e. music in tuning systems other than twelve equally spaced tones to the octave.

This sometimes has me ending up with pieces of music that I've made. Here's two that I have not yet shared on here:


Like in my previous post along the same lines, these are in eleven tones per octave - i.e. the semitone is about 9% wider than in the usual tuning, a difference that adds up for each interval. This makes the regular chords well-nigh useless, and the regular scales also do not really work out like previously.



This forces the composer to use unfamiliar chord types and unfamiliar scale types and thus also unfamiliar progressions and unfamiliar everythings.

I hope the results are compelling enough. Since these are, in a very real sense, experimental works, feedback is more than welcome.

Sunday, July 3, 2016

Detail #297: Negation and Subordinated Verbs

With some pairs of verbs, where one is subordinated to the other in some way, which one is negated may affect the meaning. Some languages might restrict which one can be negated, and leave the effect of the negation ambiguous, or resort to alternative organizations of the clause for the other meaning. Examples where which of the verbs is negated affects the meaning are easy to construct in English:
I didn't hear that he would arrive [...]
I heard that he wouldn't arrive [...]
Let's play around a bit with negation in general, and imagine consequences for constructions along those lines.

A reasonable idea for negation that exists in many natural languages and conlangs alike is to inflect the verb for negativity. Now we can imagine any number of restrictions here: only have negative forms for verbs that carry TAM marking (and have the subordinate verb be some type of infinitive). However, I'd like to up the ante beyond that: let's say only some verbs have morphological negativity marking (some even by root suppletion), and can be negated per se when infinitives, while other verbs take some negative particle, and cannot be negated when infinitives. Verbs that take particles come in several types, each which takes a different basic negation marker. The markers can also be replaced for some negative modalities, but the number of distinct negation markers in each modality beyond indicative is somewhat reduced.

Now, whenever a verb V2 is embedded in the VP of a verb V1, whichever of these two is negated gets its preferred negation if possible - if V2 is an infinitive and of the type that can be negated as infinitives, it is negated. If not, the negation migrates up one step (and creates some ambiguity), thus creating situations where
I don't know if he can do it
I know he can't do it
might not be distinct. However, alternative phrasings may provide ways out - i.e. changing one of the verbs, or rephrasing the embedded thing as a noun phrase or something.

Whenever both verbs have the same negation particle, the negation also becomes somewhat unclear – the marker might negate either one, or even both of the verbs.

However, indefinite pronouns (and some other determiners) or differential object marking may also help resolve the situation: only the direct arguments of a negated verb appear with negative indefinite determiners or with negative object marking on them (of course, the negative object marking might also be identical to the diff. obj. marking for imperfective verbs or somesuch, thus potentially reducing the usefulness of the d.o.m. for distinguishing polarity).

One could for that reason even have some kind of dummy pronoun, which is not an anaphor at all, but serves to either indicate that the last verb isn't negated, or that the first verb is negated - depending on whether it is like a negative indefinite pronoun or an "unnegative" indefinite pronoun.

Challenge: "Inverse" Adjective or Noun Marking

Challenge-time again! Come up with a grammaticalization path that creates a single morpheme that both turns adjectives into nouns and nouns into adjectives. Restriction: no sound change shenanigans.