Iqę́hhǫ the language (noun phrases intro)

[akam chinjir]

Introduction to Iqę́hhǫ

Here's a thread for the language Iqę́hhǫ and related matters.

Iqę́hhǫ is supposed to represent a large family of languages spoken in the areas surrounding the lower reaches of the Akiatu River---ideally it would be the ancestor of that family, but I seem incapable of doing a protolanguage, so who knows.

A smattering of typological superficialities: it's verb-initial and mostly head-marking; agreement and cross-referencing are a bit convoluted, but the overall system can probably be described as ergative; there are two grammatical genders, or one, if you think of neuter (nonfeminine) as a lack of gender features (I will); nouns mostly follow simple modifiers; morphology is moderately complex, though with little real fusion, and a lot will turn on where you think word boundaries should be drawn (and I'm going to try to make judgments about that sort of thing a bit delicate); and so on.

[akam chinjir]

Phonology

(I'm going to put older posts between more tags to make it easier to scroll to the latest post.)

More: show

This'll be fairly bare bones, just enough to get started.

Eventually I should come back and put in lots of examples, but not right now.

1. Inventory

1.1. Consonants

	Labial	Dental	Alveolar	Palatal	Velar	Labiovelar	Uvular	Glottal
Nasals	m		n		(ŋ)	ŋʷ <ŋw>
Plosives	p		t	c	k	kʷ <kw>	q	ʔ
Fricatives		θ <th>	s		(x)	xʷ <xw>		h
Rhotic			ɾ <r>
Glides				j		w

(I've given marginal status to ŋ and x on the grounds that the rule introducing them also merges kw with k and therefore is not purely allophonic; see the bit about labialiation dissimulation below.)

I'll write geminate consonants by doubling the letter; only the first letter of a digraph gets doubled.

1.2. Vowels

	Front	Nonfront
High	i į
Mid	e ę	o ǫ
Low		a ą

The ogoneks mark nasalised vowels.

Vowels can occur long, though this is most often the result of morphophonology. Long vowels are written by doubling.

1.3. Tones

There are marked high and low tones, which I write with acute and grave accents, respectively.

2. Phonotactics

Each syllable must have a single vowel as its nucleus.

Syllables don't require an onset consonant, and vowel-initial words are common. Consecutive vowels are always in hiatus, and the second is always more sonorous than the first, given the hierarchy a > o, e > i. (Nasal spreading ensures that nasality isn't a complication here. oe occurs almost exclusively across morpheme boundaries.)

There are no restrictions on which vowels can occur next to which consonants, except that nasal vowels never occur before coda r.

All consonants can occur in onset position.

Intervocalic clusters are limited to geminates (all nonglides can occur geminate) and sequences of r, h, or ʔ followed by any non-glottal non-glide.

Clusters of a glottal followed by a non-glottal non-glide can also occur word-initially.

All non-glides can occur word-finally.

At most one vowel within a (phonological) word can host tone; which vowel this is is lexically determined. You can think of the tone-attracting vowel as somehow prosodically prominent, and if there's a long vowel, then that's the one that'll host tone, but there's not much independent reason to think of these vowels as stressed.

Consecutive vowels, as well as vowels separated only by glottal consonants, nasals, or glides, must agree in nasality.

3. Processes

3.1. Tone

A phonological word can have at most one tone-bearing vowel. Thus, if a clitic or affix sponsors a tone, that tone will be realised, if at all, on the tone-hosting vowel of its host. If that vowel already has a tone, two outcomes are possible:

• If both tones are low, the result is a single low tone.
• Otherwise, the result is a rising tone; I'll write this with a hacek. The rising tone actually requires two morae to host it. If a coda r is available, that will do; otherwise, the tone-hosting vowel will lengthen (but I will not write this lengthening).

3.2. Consonant cluster resolution

If morphology produces an illegal consonant sequence, the following resolutions are possible:

• If the first consonant is ŋw, kw, or xw, then a preceding e or a becomes o before anything else happens.
• If the first consonant is a plosive, it becomes ʔ; if it's a fricative, it becomes h. Note that this cannot feed nasal spreading (on which see below).
• If the first consonant is a nasal, then it nasalises a preceding vowel and drops; this can feed nasal spreading.

3.3. Vowel sequence resolution

If morphology puts two vowels side-by-side, this gets resolved as follows:

	i	e	o	a
i	i	ie	eo	ea
e	i	e	eo	ea
o	oe	oe	o	oa
a	e	e	o	a

Note that though io and ia are legal sequences, they become eo and ea across morpheme boundaries (maybe there'll be some lexical exceptions to this rule).

3.4. Nasal spreading

Vowel nasality spreads across nasals, glottals, and glides. There's one tricky point: across morpheme boundaries, surface ʔʔ and hh can represent an underlying oral plosive + ʔ or oral fricative + h sequence, and the oral obstruent will block nasal spreading before debuccalising.

3.5. Labialisation dissimulation

When a labialised velar (not including w) follows another labialised velar (including w) within a phonological word, it loses its labialisation.

3.6. Coda

The processes listed above all have the phonological word as their domain. What that exactly means is probably going to be a bit complicated, and it may turn out that there are same bound forms that (for example) resist nasal spreading but conform to labialisation dissimulation.

4. Phonetics

This part is super-sketchy. I mostly want to say something about how the glottal consonants play with everything else.

Here are the main phonological paradigms:

Plosives	Fricatives	Sonorants
ata [ada]	asa [asa]	ama [ama]
atta [atːa]	assa [asːa]	amma [amːa]
ahta [aḁta]	ahsa [aḁsa]	ahma [aḁm̥a]
aʔta [a̰ʔta, a̰t'a]	aʔsa [a̰ʔsa]	aʔma [a̰m̰a]
ta [ta]	sa [sa]	ma [ma]
hta [ʰta, hʰa]	hsa [ssa]	hma [m̥a]
ʔta [t'a]	ʔsa [ʔsa]	ʔma [m̰a]

So preconsonantal h and ʔ are mostly apparent in their affects on neighbouring sounds, devoicing and glottalising, respectively. You'll notice that intervocalic ʔ+plosive gives two possibilities, preglottalisation and ejection; I don't know what if anything conditions the alternation. Similarly with preaspiration vs aspiration. ʔp will usually be ɓ rather than p'.

Relatedly, aʔa is usually just a̰ː, and similarly for other vowels.

Word-final obstruents debuccalise to ʔ or h, as appropriate. Word-final nasals drop after nasalising the preceding vowel; this feeds nasal spreading.

...I think that's enough to get started with?

[/i]

[akam chinjir]

Nouns and simple noun phrases

1. Gender

There are two grammatical genders, feminine and neuter (F and N in glosses).

There's a lot I don't know about how semantic gender works in this society and language. As you'd expect, nouns referring specifically to female humans will usually be feminine, and nouns that can't refer to female humans are normally neuter. But more than that I'm not ready to say, so I'll focus here on narrowly grammatical issues.

The main thing is that neuter is a default; think of Iqę́hhǫ grammatical gender in terms of a privative feature [feminine], which neuter nouns simply lack, not in terms of a binary feature such as [±feminine]. This'll be important when we get to agreement: verbs sometimes do and sometimes do not agree in gender with the subject, but failing to agree in gender is morphologically indistinguishable from agreeing with a neuter noun---in both cases you get no overt marker of agreement.

2. Adjectives

There'll be a class of underived adjectives, maybe a fairly large one, as well probably as productive ways of deriving adjectives. More on that some other day.

Adjectives most often go before the noun; in this position, they do not agree in gender or number:

• ahtǫ coáte big spider(s)
• ǫąwą ícije yellow feather(s)
• mióme émolę happy fish

Prenominal adjectives come in a fairly rigid (and unsurprising) order, for example with size coming before colour: you get ahtǫ ǫąwą émolę big yellow fish rather than ǫąwą ahtǫ émolę yellow big fish, for example.

Adjectives can also occur after the noun, in which case they do agree:

• coáte ahtǫtha big spiders
• ícije ǫąwąthea yellow feathers
• émolę miómetha happy fish

Word order among postnominal adjectives is fairly free. For this and other reasons, I think of these adjectives as reduced relative clauses.

Here are the agreement markers that occur on adjectives:

	S	Pl
F	-(th)i	-thea
N	∅	-tha

(You'll notice that in the examples above, ícije feather is feminine, coáte spider and émolę fish are neuter.)

The singular feminine agreement marker is -thi after i or ʔ or h (and a stem-final h drops); otherwise it's an i that coalesces with a stem-final in the usual way. When there's a long vowel in the stem, it'll usually shorten in the presence of feminine or plural agreement.

Some adjectives have reduced variants that can occur only immediately before the noun, as clitics; in several cases these are suppletive. For example, you could have iki=coáte for big spiders, with iki replacing ahtǫ.

There are some interpretive differences worth mentioning: postnominal adjectives can only have intersective interpretations; reduced clitic adjectives will have a nonintersective interpretation when one is available; otherwise prenominal adjectives can get either sort of interpretations.

But I've got enough to say about the distinction between intersective and nonintersective interpretations that I'll give that its own section.

3. Intersective vs non-intersective modification

A phony priest isn't someone who's a priest and is also phony; a main idea isn't an idea that's main. These are cases of nonintersective modification, cases where the set of all F Gs isn't the intersection of the set of all Fs with the set of all Gs.

It can be more subtle than with phony and main. If you describe something as a large spider, you probably mean that it's large by spider standards---a large spider is still much smaller than a small boar. But if you refer to a spider, which is large, you likely mean that it's large not just as a spider but by more general standards; if I tell you that in the next room there's a spider, which is large, and a boar, which is small, you might well take me to be implying that the spider is bigger than the boar.

You can also get ambiguities: is an Italian professor someone who's a professor and is Italian (intersective), or someone who's a professor of Italian (non-intersective)?

In English, modifiers that occur before the noun can be intersective or non-intersective (which is why you can get ambiguity). When you maneuvre an adjective after the noun, though, it'll always have an intersective sense, and predicate adjectives are also always intersective.

That part is language-specific. In French (assuming I've got my French right), l'ancien régime, with the adjective first, is the former regime (nonintersective); le régime ancien, with the adjective second, is probably the ancient regime (intersective). The French rule is that adjectives before the noun are nonintersective, adjectives after the noun can go either way.

Or in Mandarin you've got nonintersective lǎo péngyǒu 老朋友 old (longtime) friend vs hěnlǎode péngyǒu 很老的朋友 friend who is old: without de 的 you get a nonintersective sense, with it an intersective sense (in the second case the modifier also has to be bisyllabic, I think).

As I mentioned above, the Iqę́hhǫ rule is that adjectives after the noun are intersective; reduced clitic adjectives are nonintersective when that's an option; and other prenominal adjectives can get either sort of interpretations. Thus, for example, an ahtǫ coáte big spider might be big only compared to other spiders, but a coáte ahtǫ must be big by more general standards.

Another example.

A ʔtaké is a member of your band or camp (I'll say bandmate). Iqę́hhǫ bands would generally have maybe two or three dozen members. Relations with one's bandmates aren't rigorously distinguished by relations established by descent or marriage, either lexically or in social norms. Membership in a band is somewhat fluid (details to be worked out), but generally speaking these are people you'd know for a long time and well.

So a ʔtaké will often be a wetíthe ʔtaké old bandmate, with wetíthe old most likely taking on the unsurprising nonintersective sense longtime. You could describe the same person using a reduced adjective as weti=ʔtaké, in which case only the nonintersective interpretation would be available. And if you wanted to characterise a bandmate unambiguously as elderly, you could put the adjective after the noun: ʔtaké wetíthe. (In this case the head noun is neuter and singular, so there's no overt sign of agreement; if you were describing several elderly bandmates, you'd get ʔtaké wetítheatha)

4. Number

Nouns are not inflected for number; to indicate number, use an actual cardinal number, or the plural word iqqe. Like adjectives, these can either precede or follow the head noun, and they agree with the head noun only when postnominal.

When both a number and an adjective occur before a noun, the number is first, as you'd expect. After the noun both orders are possible.

I want there to be a difference of interpretation depending on whether the number is before or after the noun---something akin to the intersective vs nonintersective distinction I drew for adjectives. But so far I haven't come up with a way of doing it that satisfies me.

Here are the first ten cardinal numbers (it's a base-ten system):

	N S	N Pl	F S	F Pl
one	akw	aʔtha	akwi	aʔthea
two	męą	męątha	mę	męąthea
three	htòò	htòtha	htòe	htòthea
four	éhe	éhetha	éhi	éhethea
five	ąjǫ	ąjǫtha	ąję	ąjǫthea
six	mí	mítha	míthi	míthea
seven	cǫ̀ʔ	cǫ̀ʔtha	cǫ̀ʔthi	cǫ̀ʔthea
eight	ʔpo	ʔpotha	ʔpoe	ʔpothea
nine	kéme	kémetha	kémi	kémethea
ten	saar	sartha	sari	sarthea

(When an adjective does not agree, it gets the default neuter singular form. As you might expect, the agreeing feminine singular forms are rare, if they occur at all, with numbers greater than one---I haven't ruled out sentences analogous to the group was three, so I'm including them for completeness.)

The plural word iqqe agrees regularly: iqqe, iqqetha, iqqi, iqqethea. (As with numbers greater than one, you'll see non-default singular agreement on iqqe rarely if ever.)

Both akw one and iqqe some have uses as indefinite articles, implying specificity/referentiality, though only prenominally. They're never obligatory, though. Both can occur with nouns you might be inclined to think of as mass nouns.

5. Definiteness

There's a definite article, a clitic that attaches most often directly to the noun, and which agrees with the noun in gender and number. Here are its basic forms:

	S	Pl
N	=(e)ʔ	=ʔa
F	=ʔi	=ʔea

There's a morphophonological peculiarity: when the definiteness marker is ʔ-initial, the glottal stop metathesises with a stem-final consonant; thus the monkey becomes iʔkeʔti (this being a t-final feminine noun).

There's some fancy business here. The definite article is actually a second-position clitic, its position prosodically determined: it follows the first phonological word in the noun phrase. Unsurprisingly, this will most often be the head noun. But there are complications.

Suppose we wanted to make ahtǫ coáte big spider definite. We couldn't append the article to the noun (*ahtǫ coáte=ʔ), because that would put it after two phonological words. And you can't simply reverse the adjective and noun, to yield coáte=ʔ ahtǫ, because ahtǫ big would get a different interpretation after the noun, intersective rather than possibly nonintersective.

There are three possibilities:

• Put the definite article on the adjective: ahtǫ=ʔ coáte.
• Use a reduced, bound form of the adjective: iki=coáte=ʔ.
• Use the deictic we to host the definite article: we=ʔ ahtǫ coáte

(You get the same possibilities when the noun occurs with a number or with the plural word.)

The difference between these is pragmatic, to the extent that it has any significance at all: the position to the left of the definite article can be used to focus a particular element of the noun phrase.

One last thing. As we'll see, in a transitive clause the verb can agree in gender with its subject. But the noun phrase (or DP) as a whole does not automatically inherit the gender of its head noun, and when it does not, the verb gets default gender agreement. Details can be a bit complicated; but one case in which the noun phrase as a whole does not inherit the gender of the head noun is when the head noun occurs to the right of the definite article. That's to say, for the purposes of transitive-subject agreement, ahtǫ=ʔi iʔket big monkey is neuter: the head noun is feminine, but it's the adjective that hosts the definite article. Meanwhile, iki=iʔkeʔt=i, with the reduced adjective occuring with the head noun before the article, is feminine, and will control feminine agreement when it's the subject of a transitive clause.

[k1234567890y]

not bad so far.

however, I have one minor opinion related to the terminology: I think the "neuter" gender in your language would be called "masculine" if it were other languages, as male people and animals would belong to your "neuter" gender.

but this is just my opinion, you don't need to follow that.

[akam chinjir]

Thanks for the comment!

I'll start my response with a bit more about how I'm imagining grammatical gender interacting with semantic gender in Iqę́hhǫ, initially restricting attention to words for human beings.

Some words will (by Iqę́hhǫ lights) refer specifically to women or girls---including eppe woman and mąmǫ mother for example. These will have a very strong tendency to be of the feminine gender.

But also:

• There will also be words like qohro man or pąmǫ father that (by Iqę́hhǫ lights) refer specifically to men or boys.
• And there'll be words like ęąwę́nǫ drinkspirit, shaman that refer specifically to people who (by Iqę́hhǫ lights) are neither women nor men (nor girls nor boys)---they're in another gender category. (I'm afraid I'm not yet in a position to say anything interesting about this category, or any others the Iqę́hhǫ might recognise, except that this one is very much bound up with ideas about and experiences of the ancestors.)
• Finally, there'll be words that are gender nonspecific, like wąto person.

Words of all these three nonfeminine types will have a very strong tendency to be of the gender I called neuter; but it's only the first of the three types that you'd think of as masculine; and that's a reason for calling this gender neuter.

It's also not right to call this gender nonfeminine, exactly---wąto person isn't feminine, but it also doesn't refer specifically to people who aren't female. This is supposed to be the point of explaining the system in terms of a unary feaure [feminine] rather than a binary feaure [±feminine].

There are some corollaries. One is that neuter will be the elsewhere gender: a very large majority of nouns that aren't semantically feminine will be neuter. (But I so far have very little to say about what nouns will count as semantically feminine, by Iqę́hhǫ lights.) Another is that agreement with a neuter noun will be shown by the lack of an agreement marker. (And yes, lack of a marker, not a null marker, if you care about distinctions like that.)

I hope that helps make sense of my decision here!

[k1234567890y]

ok (:

after all it is your lang so you decide what to do on it.