Detail #342: Other Quirky Things besides Case

Quirky case is probably familiar to anyone with an interest in case systems. We could, however, considered similar lexically caused quirks for other things than case!

Quirky Voice

Consider some nouns that trigger exceptional voice marking on some, or even all vers? Maybe they cannot be direct objects, so whenever they are objects, they force some detransitivizing voice and some kind of oblique marking.

Quirky TAM

Some adjectives as complements, or maybe some nouns as verbs or objects, might cause unusual behaviours with regards to tense, aspect or mood. Maybe some noun as subject (or as object) always triggers some irrealis mood regardless of the reality of the VP and of the subject.

Quirky Gender

Some Semitic languages have some numbers mark opposite gender agreement. We could consider, though, verbs or adjectives that mismatch gender. Coming up with a historical reason for these might be interesting.

Quirky Adjective/Adverb Things

One could imagine certain nouns or verbs forcing comparatives or superlatives, or vice versa, some nouns or verbs blocking comparatives and superlatives (but making superlative or comparative meanings mandatory).

Quirky Number

Some adjectives or verbs or adpositions could imaginably also require some noun in some relation to it to mark number exceptionally. Alternatively, some verb could break the number congruence if verbs have that in the language.

Quirky Definiteness

I guess by now the reader should be able to come up with some ideas for this as well, analogously to the ideas given above.

Ćwarmin: The Phonology, pt I

The basic consonant inventory:

bilabial

alveolar

postalveolar

palatal

velar

voiceless

voiced

voiceless

voiced

voiceless

voiced

voiceless

voiced

voiceless

voiced

stops

p

b

t

d

c

k

g

affricates

č

(j)

ć

j

fricatives

f

v

s

z

š

ž

ś

ź

x

ǧ

rhotics

r

glides

w

y

laterals

l

nasals

m

n

ň

ń

ŋ

The voicing of /b/, /d/ and /g/ sometimes borders on stiff voice rather than modal voice, especially word-initially.  In previous transcriptions, I've used <h> for <ǧ>. I prefer  ǧ over ǵ due to ˇ being easier to spot. Sequences of stop + fricative even over word boundaries are rendered as affricates if possible. Otherwise, one part of the sequence is  lost. Generally, voiceless stops win over voiced fricatives, so ts → t, pf → p, but voiced fricatives win over voiceless stops, so dz → z, bv → v. A light lengthening of the resulting segment may occur in some idiolects, giving ...Vd zV.... → ...V z:V, ...Vp fV... → Vp: V...

Since this is the only position where consonant length has any significance for obstruents in Ćwarmin, and the distribution of this is rather the same as the distribution of stop+fricative, length is not considered a distinctive feature for stops.

Historically, /ts/ has also existed in the phonemic inventory, but has merged with /t/ and /s/ depending on position in the word – /s/ favoured in the end and onset, /t/ favoured elsewhere.

ń and ň are very marginal phonemes. Ń and n are only distinguished over morpheme boundaries; ...-ńć... would come out distinct from ....n-ć... This situation sometimes appears in some verbal inflections. ň and n are distinguished in a few words before and after /ž/. Minimal pairs and near minimal pairs are /kažna/ (snow drift), kažňa (accident), uňžo (friend), unžon (to sharpen), ižňət (important), ižnel (reject, leave, abandon). These are not distinguished in the orthography.

/n ń ň/ do not assimilate to ŋ before velars. /m n ŋ/ all three go back to proto-ĆŊD. The only assimilation that has occurred is n, ŋ > m / _labial obstruent.

A cluster cannot contain both palatals and postalveolars in sequence. The leftmost sound's place of articulation wins out in cases where such sounds somehow get stuck together. Postalveolars and palatals also tend to win over alveolars. Rhotics and laterals only ever appear in postalveolar or palatal versions as allophones of /r/ and /l/ in such clusters.

The maximal syllable would be CCVVCC. It is unusual for three or more consonants to occur in a row. The sonority hierarchy is rather strictly observed, except that fricatives, affricates and stops are tied. Rather peculiar clusters do occur even initially, such as xpan (rake), ǧdisən (to sieve), fkarul (badger), but these are not very common. Clusters of fricatives also occur, such as lašxur ('to peel'), "ǧzin yə ǧvin" ('this and that'), vǧum ('to yawn'), vžuk ('a tick'), vźuc ('a thread'). Word initially, labials do not cluster with each other, except mw- and a few bw-.

Stress is almost invariably on the initial syllable, after which every other syllable has secondary stress. However, if the final syllable would carry stress and there is a following word, the final syllable's stress is weakened. The language is roughly speaking syllable-timed, with stressed syllables being slightly longer. Some styles of poetic recital, however, seem to intentionally go for something more like stress-timed rhythms.

Sargaĺk Vocabulary: Colors, Clothing and Certain Status Symbols

The Sargaĺk vocabulary for colours is not very extensive, but some terminology exists. The nouns for these colours - e.g. '(the colour) red' are syntactically not quite nouns, and thus often require some kind of auxiliary noun toform well-formed NPs. These nouns only have two cases, the absolutive and pegative. The form given with extra indentation is the adjective form, which does have a full set of congruence marking.

stal red, includes most pink, some purple, some brown and orange

stax

kəma white, includes some pink and gray.

kən

buxu yellow

buxĺ

ŋoca - blue, includes some brown and gray.

ŋot

ŕt'a - green

əŕt'

sokca - black, overlaps partially with red and blue

sokcń

These are the basic color terms. To express, say, 'the colour green', you would use the noun sibik (fem) and the colour name in the pegative:

sibik stalta, sibik kəmta, sibik buxta, sibik ŋocta, sibik ŕt'əta, sibik sokta

Sibik is a normal noun and can be inflected in any case.

As for colours more generally, the Sargaĺk do not have access to many coloured fabrics or paints. Green and red paint do exist, but are very expensive.

White clothing is a status symbol - it's hard to keep clean, and obtaining really white cloth is not trivial either. The colour ŋoca, ~blue, includes many seal fur colours.

Detail #341: A Relativization Strategy

Indo-European languages seem to have a liking for using relative pronouns. Some languages prefer relativizing particles, resumptive pronouns or participle-like constructions. Let's consider, instead, an approach where a certain verb serves as relativizer. We will look at some of the possible complications as well.

For the sake of simplicity, let's assume we're dealing with an SVO language, where pro-drop is permitted. Much like English, certain contexts require verb-initial statements, but these contexts are more common than in English, including, for instance, almost every modality except indicative in main clauses.

The relative verbs are a small set of verbs, maybe two or three at most. What dimension they are distinguished along is not all that necessary, but we could go through a few possible ones:

active vs. passive vs. oblique
active animate vs. all other possibilities
({imperfective, perfective}*{active,passive}) + {circumstantial}

In this language, there is person congruence, and this inserts one place for introducing some quirks. Let's consider the second paradigm given above: active animate vs. all other possibilities.

a being a-anim.3sg infinitive [obj ...] ....
a being which is infinitiving [object ...]  ...

a being allotherpossibilities.1sg infinitive
a being that I am infinitiving

a being allotherpossibilities.1sg there(verbal!)
a being I am located at

 However,

a thing allotherpossibilities.3sg infinitive [...] ...
a thing which is infinitiving

a thing allotherpossibilities.1sg infinitive [...] ....
a thing that I am infinitiving

a thing allotherpossibilities.3sg infinitive a noun [...] ...
a thing that is infiniting a noun

a thing allotherpossibilities.3sg a noun infinitive_?
a thing that a noun is infiniting

_? signifies some other infinitive form, maybe a participle?

If there is no noun that the verb can refer back to, and the verb is inflected for first or second person, this basically means "I, who ..." or "you, who ...". Analogous third person constructions favour having an explicit pronoun - but it can be either exo- or endoheaded: "he a-anim.3sg infinitive" OR "a-anim.3sg he infinitive": he who infinites.

Using the infinitive feels slightly predictable, but so would using special relative moods for each verb.

Case Prefixes: Dagurib and Solgır

Dagurib was the first branch in the ĆŊ family to diverge from the rest. Through its history, however, it has been in extensive contact with the Ŋʒädär branch, and due to its relatively small number of speakers - for most of the time, only about 5% of the ĆŊ population - it generally does not provide much data with regards to early ĆŊ or even ĆŊD linguistics. However, many Dagurib languages have conserved a feature that is also shared by Solgır (a Ŋʒädär language), and which also has left traces throughout the family. These are the case prefixes.

The case prefixes are always single segments - in Solgır a-, n-, r-, e-, ä-, o-, q-, k-, f-, in Banar (a Dagurib language) m-, l-, z-, i-, a-, -ä-, u-, k-, k'-, p-, in Dagurib m-, l-, s-, i-, a-, ä-, u-, ü-, k'-, p'-. Some words just take them in some cases, and what cases they take them with differs from noun to noun. Sometimes, a noun whose stem begins in any of these is reanalyzed, and the segment is lost in all forms but a few.

A system of this kind seems unlikely to develop by itself twice in separate branches, and would therefore rather seem likely to be a reduced retention of a previous system. This is further supported by the fact that in other Ŋʒädär, Dagurib and even Ćwarmin languages, there are words where cognates have random-looking losses of initial elements. (Or is that random additions of initial elements?) Further down, we'll see some examples of cognate-sets where potential prefixes could explain some of the alternations between forms.

The Solgır and Dagurib prefixes do not seem to mark anything – the distribution is lexically determined. Essentially, for some nouns, some cases for that noun take some prefix. An extreme example would be the Solgır noun ermi, rope.

	sing	plur
abs	ermi	ermeyi
dat	fermen	ermeyin
abl	ermelin	kermeylin
loc	kermene	kermeyine
lat	ermeli	fermeyli
gen	nermeni	nermeynit
instr	ermerik	nermeyrik

The particular prefixes do not belong to any particular case and can appear with different cases for different nouns. Mostly though, any noun with these prefixes will only showcase one particular prefix - tho' in some Dagurib languages, a certain free variation between two prefixes or between a prefix and none is well attested.

The historical origin for these prefixes is shrouded in mystery - they probably have communicated something in proto-ĆŊD, but the lack even of hints as to what that might have been prevents reconstruction of it.

Nouns that hint of the presence of such a system in proto-ĆŊ can be found. Some of these also have cognate examples even in Dagurib. Such examples will be marked with a + in the Ćwarmin cell or Ŋʒädär cell.

Here are some examples within the Ćwarmin branch:

Ćwarmin		Rasmjinj		Ətimin		Astami
+ruanas	hair	wanos	hair	ronas	a hair	uonoh	fur
arać	skin	karos	leather	karaš	skin	karaś	skin
+nitis	thing	njidjis	thing	iedəs	thing	itih	thing
səkin	tally mark	ädʒin	property mark	seəčən	personal 'flag' on tents	səki	property mark
+els	oath	felz	oath	feləs	oath	hel	oath
maruw	liver	aru	liver	-	-	maru	liver
+ənve	chin	pänne	cheek	peənyə	chin	-	-
+maso	grass	amozo	grass	mazo	grass	amoh	grass
əced	bridle	ced	bridle	ceəd	bridle	əcet	halter
deǧe	prayer	deǧe	prayer	ədeje	prayer	ado
kinim	braid	ičinjin	braid	čienən	braid	ikini	woven material

The Dagurib cognates are:

ruanas - (r)ouno
nitis - (l)iti
els - (k')eili
ənve - (m)einyi
maso - (m)eizyi (due to random vowel harmony reassignment)

Within the Ŋʒädär branch these are a few examples:

Ŋʒädär		Törǧdär		Suwdan
+soman	home	oma	home	sömä*	home
iqe(-k)	man	miqe(-q)	man	liqe	man
ragaz	bumblebee	aragaz	bumblebee	arga	wasp
+anək	animal	ranık	livestock	anə	calf
+ästil	pelt	stıl	fur	istil	fur

*Süwdän has had an unusually large share of seemingly random vowel harmony changes; in ĆŊ, the vowel-harmony conservation ratio for 1000 years is something like 98%; for Süwdän, however, it seems the percentage is closer to 90%.

The Dagurib cognates are

soman - (p')amai- (house)
anək - (m)aŋu- (reindeer)
iqe(-k) - (i)q- (member of the tribe)
ästil - (ü)t(:)i- (clothing)

The situation seems to be that the prefix has been generalized to all the forms or none of the forms in the languages that have lost the prefixes, and rather randomly at that.

More examples can be found between the branches, and the Dagurib family itself is particularly rich with these, and especially with traces in those languages that only more recently have lost the prefixes as a grammatical feature.

Detail #340: A Type of Morphology

aA type of morphology I don't think I've seen anywhere but that has a certain similarity to Semitic morphologies (but only on some level, conceptually) is the following:

Many, but not necessarily all, consonants belong to various chains of gradations. Since this isn't going to be a fully constructed idea, I'll just present some possibilities inspired by Finnish:

k: > k > g > w, y, ɣ
ŋk > ŋg > ŋw, ŋy, ŋɣ
p: > p > b > w
w > w
y > y
ɣ > ε (empty string)

The various gradation patterns come in three basic forms: initial, middle and final. Prefixes and suffixes also introduce a new set of patterns: the exfinal and exinitial gradations, i.e. patterns for how a formerly final or formerly initial consonant behaves when forced into a new environment. These patterns are very similar to middle patterns, but differ in some subtle ways.

The w > w and y > y  examples illustrate changes that just don't happen: a root where such a change would be expected remains invariant with regards to that consonant.

A special situation may occur in monosyllabic words if 'penultimate' consonants rather than medial consonants are the actual class that this operates on. However, consonants that are both initial and penultimate may behave in unique ways that partially pattern with medial, partially with initial consonants. (The complement way around could also be conceivable: final consonants in monosyllables behaving in a way that merges final and medial traits; having both occur seems to stretch credulity though, and this would simply mean that in monosyllables, both ends of the word behave wonkily.)

Sometimes, the surrounding vowels may affect how a consonant behaves, giving us many possible outcomes:

g > w (before back, close vowels), y (before front vowels), ɣ (before open back vowels)

These patterns may differ depending on position, due to various historical changes.

Now, a root is a string of the form

(C)V((C)CV)(C)

 Thus yaktub, ka, tak:, bastu or aŋɣa are all permissible. The morphology then operates in two ways: affixes and gradations. We may have, for instance, rules that derive the following inflected forms (on the right) from the roots tak and tak:. Notice that these roots also have homonyms appearing among the inflectional forms of each other:

tak > dak
tak: > tak
tak: > tag
tak > tag
tak > taɣ
tak: > dak:

It is conceivable that root-k behaves different from k-derived-from-k: one can then imagine that the roots tak: and tak differ down the line:

tak: >	tak >	tax
	tak >	tag >	taɣ

Of course, I am not just talking about k vs k: here, any pair whose lines have shared elements could imaginably behave differently down the line of changes, due to whatever underlying historical changes hide behind the analogies that created this morphology.

Medial clusters can muddy the picture a bit, mostly with a variety of assimilations or dissimilations:

yaktub > yagrub (< *yagdub)
yaktub > yaktuw
yaktub > yaxsuw

pastu > bastu
pastu > wastu
pastu > baru (< *bardu < *bazdu < *basdu)

Morphological forms could then be described as

C_aVC_bVC_d

where a, b and d are numbers indicating how far down the chain of gradation the sound is to be shifted; a number that probably never is larger than 3; this gives, for three-consonant roots 27 possible forms; with affixes added, this permits for quite a sufficient amount of morphology. Some roots will have fewer on account of having consonants from which there is nowhere to go: they either vanish at the first stage or get stuck at some point just mutating into themselves but this is not unusual for real-life languages either.

Ŋʒädär: Becoming, Turning, Starting, Increasing, Changing

Becoming something, turning into something, starting out doing something, increasingly being something and changing are all somewhat similar notions. So, there's probably very little surprise that Ŋʒädär approaches these in similar ways. However, it also introduces an interesting distinction between different ways things change or begin or turn.

Certain intransitive verb suffixes in Ŋʒädär take a special third person morpheme, even though the usual third person finite present verb takes no such suffix. This suffix is -n.

For "turning adjectival" or "beginning to verb", similar structures appear:

[adjectival stem] [-lUA-] [TAM and intransitive person congruence markers]

[verbal stem] [-lUA-] [TAM and intransitive person congruence markers]

caban bör(ü)l-ät sarctu-lua-bra-n
road spring-dat.pl rough-get-HAB-3sg_intr_pres
the road gets rough in spring

The -lUA morpheme has a Ćwarmin cognate, -lgU-, which only appears in a few lexical items. In both the Ćwarmin cognates and the Ŋʒädär case, -lUA only signifies changes over time, not changes over space. Thus, if the terrain gets rougher to the west, a different morpheme is used:

[stem] [-(A)rgA] [TAM + intr. person]

caban sıvlı-vımə sarctu-rga-n
road west-through rough-get
the road gets rough towards the west

For more abstract "spaces" over which something changes, both are used:

k'or-gəvi dənt'ı-rga-s-t
salt-on-account-of thirsty-get-1sg_intr-pp*

k'or-gəvi dənt'ı-lua-s-t

"I got thirsty due to (the) salt"

The  acquisition of mass nouns can also use both of these. Most dialects prefix the noun oblique stem by GI- and suffix -LUA-:

gı-təŋə-lua-n
GI-dust-acquire-3sg
it gathers dust

The GI- prefix has some lenition before i: ji- vs. gı-. 

Beginning to do something usually takes -lUA on the verbal stem. If one begins to do something due to geographical location, -rgA can also be used, but is less common since in some sense verbs tend to be distributed temporally.

For becoming [a noun] two approaches are in common use. The first method distinguishes volitionality of the subject. The second signals the definiteness of the complement. The first uses the verb to be with the -lua suffix: ihlüä...
Volition marking is optional, so the lack of such a marker does not imply lack of volition, it only permits for it. The marker is the presence of the reflexive pronoun:

Arbas kamma-ɣuv ihlüä-n
Arbas chieftain-compl become-3sg
Arbas becomes (a) chieftain

Arbas kamma-ɣuv ŋul-ʒuv ihlüä-n
Arbas kamma-ɣuv ŋul ihlüä-3sg
Arbas chieftain-compl self become
Arbas (has applied himself and therefore) becomes (a) chieftain

The reflexive pronoun seems to appear in both the complement case and in the absolutive case. The complement case probably is a result of some kind of dislike for having two absolutive arguments with an intransitive verb and then going for the other case that is present in the clause. This varies from speaker to speaker and community to community.

The other method used ŋul- as the root for the verb:

Arbas kamma-ɣuv ŋul-(l)ua-n
Arbas chieftain-compl self-begin-3sg
Arbas becomes the chieftain

The Ŋʒädär constructions for these are unique to the Ŋʒädär branch of the Ćwarmin-Ŋʒädär-Dagurib languages. The Dagurib branch retains a less synthetic set of constructions with some cognates to the Ŋʒädär -lUA and -ArgA morphemes. The Ćwarmin branch only has a very small number of cognates, and is using quite different constructions.

Ŋʒädär: Possessive Suffixes

The statements here hold for most of the Ŋʒädär branch, and also widely for the Dagurib branch of Ŋʒädär-Ćwarmin languages.

Pronouns as possessors are generally not present as words unto themselves, but as suffixes of the

	absolutive possessum				dative possessum		oblique possessum
	singular proximative	plural proximative	obviative		proximative	obviative	proximative	obviative
1 sg	-sA	-sAr	-s		-As	-sOt	-As- or -sA-	-sIs-
1 pl	-dA	-dAr	-dAx		-Ad	-Od-	-Ad- or -dA-	-dIs-
2 sg	-Un	-nUr	-nUx		-Un	-On-	-Un- or -nU-	-nIs-
2 pl	-Ur	-rUr	-rUx		-Un	-On-	-Ur- or -rU-	-nIs-
3 prox	-sI	-sIr	-sIx		-Is	-IsI-	-sI- or -Is-	-Is-
3 obv	-	-	-(O)q(O)		-	-(O)q(O)-	-	-(I)qIs-

The location of these suffixes vary; the absolutive suffixes go word-finally, and replace the case marker completely. This conflates the singular and plural obviative distinction for the possessum. The dative suffix goes finally for proximative datives, but before the case suffix for obviative ones. The number distinction is lost with dative possessums. The obviative dative suffixes require the plural dative marker, even if the possessum is singular - this same pattern appears with the obviative marker for the dative.

For the obliques, the possessive suffix goes before the case suffixes, and replaces number and obviative markers.

Second person singular and plural are conflated in some positions.

Proximative nouns cannot have obviative possessors. Third person non-personal pronominal possessors (i.e. indefinites and such) may cause the possessum to take the 3rd person obviative marker, but this seems highly optional, possibly emphasizing the type of pronoun of the possessor.

Uses for these include, beyond possessors, object suffixes on some participles, subject suffixes on others. The word 'and', on also can take the absolutive singular proximative forms (or the obviative -q(O), to signify 'and (pronoun)', or '(pronoun) too' and similar things. These forms further can take regular case suffixes, getting forms like onsam 'and to me, to me too', ondaŋa 'and at you, at you too'. Finally, some adjectives can take these for complements, including p'ürkör (similar to X), so p'ürkörüx = similar to you(pl), korqəl (related to), korqəlas-[case congruence]. Different adjectives may take different suffixes here: p'ürkör takes absolute obviative or plural proximative in numeric congruence with the noun that is similar to someone; p'ürkör does not take congruence when it has a possessive suffix. Korqəl always takes case congruence with the head noun, and therefore follows the full complexity of the table above. The third type can be exemplified by varın-, "appealing (to X)" (always dat. prox. poss. forms).

Thus, we have three classes: I (p'ürkör-like adjectives), II (korqəl-like adjectives) and III (varın-like adjectives).

Some examples:

I

kostan- (unappealing to)
mitkis- (insufficient for)

vinei- (too small for)

irib- (angry at)

II

gərəs- (loyal (to X))

sork'o- ( thankful (to X))
nüre- (infatuated (with X))

sajan- (married (to X), only used for the wife

III

ökäm- (married (to X), only used for the husband

sacıd- (insufficient for)

akom- (too much (for X))

t'ebä- (dear (to X))

Detail #339: Simple Gender Systems

wals.info confirms that pretty much all "small" gender-like systems that don't have a masculine/feminine distinction use an animate/inanimate distinction. A simple way of differentiating your conlang from the usual fare presents itself here: a two-gender system that is neither fem/masc or anim/inan. Let's establish two end-points- all humans belong to gender 1, all mass nouns to gender 2. However, this is not a count/mass-distinction either.

Nouns that are neither human nor mass are arbitrarily distributed between the two. Nouns that are both - crowds, tribes, etc - can behave as members of either class.