Morphology and morphophonology

General principles

Stress

Cuhbi permits only one stressed syllable per word. In both nouns and verbs, this is typically the first syllable of the root:

k-i-ʂi-ø > ciˈʂi ‘I am picking it up’

However, any syllable with a long vowel will carry stress on that long vowel. Typically this presents no problems for morphophonology per se, but when a suffix is added which has a long vowel, or which causes the lengthening of an unstressed vowel, the stress can shift:

i-ʂi-ø-nii > iʂiˈnii

If the word already had a long vowel, that vowel is shortened:

i-ʂi-t-b-nii > iʂiifà-nii > iʂifàˈnii ‘it’s him who snatched it up’

Also important to bear in mind is the fact that reduplicated roots carry the stress on the second syllable, not the first:

i-fé-h=usu > iféˈfehusu ‘isn’t looking at it’

Long-short pairs

Note that often, long vowels alternate with short vowels in different morphological forms of a word. The specific contexts which trigger this are listed under specific morphemes below, but the diachronic reason is typically compensatory lengthening following the elision of a consonant. This alternation is relatively straightforward, but note that /e: o:/ have collapsed into /i: u:/, meaning that the long forms of /i e/ and /o u/ are not distinguished and appear always as /i:/ and /u:/ respectively:

o-ʁ-t > uusà ‘who did’

*ʁ

The morphophoneme *ʁ (generally the reflex of a real phoneme, although analogy has levelled some instances and produced new ones) occurs in several words. In final position it manifests as compensatory lengthening, like a normal latent consonant:

gàà[ʁ] ‘bread’

Intervocalically, it manifests as either /j/ or /v/ following a front or back vowel respectively:

gàà[ʁ] > gàye ‘bread’

In a cluster with a plosive, the plosive is fricated and the *ʁ disappears with compensatory lengthening:

gàà[ʁ]-dèhè > gààzèhè ‘bakery’

In a cluster with a fricative, the fricative is devoiced (if it is voiced) and *ʁ disappears with compensatory lengthening:

gàà[ʁ]-qi > gààxi ‘baker’

In a cluster with a nasal, it assimilates entirely and the nasal is denasalised:

gàà[ʁ]-nyà > gànycà ‘bread oven’

The augmentative -:xà originally had an -/ʁ/, due to which any preceding vowel is lengthened:

dàme-:xà > dàmiixà ‘very beautiful’

Vowel deletion

Unstressed short vowels may, if the resulting consonant cluster is permissible by Cuhbi phonotactics, be reduced to zero:

i-ʂi-h-g-r=usu > iʂìkìrusu > iʂìkrusu ‘is not having his things picked up one by one’

Vowel harmony

Historically Risha Cuhbi had progressive backness-based vowel harmony, with three front-back pairs: /e o/ /i u/ /a ɑ/. This harmony was blocked by [j w] (modern reflexes [j v]) but otherwise spread throughout a word. This principle can still be observed in the majority of native Cuhbi words:

dibe - eyebrow

motùn - yam

Most suffixes (except the most recently grammaticalised) have two allomorphs, one with a back vowel and one with a front vowel:

baad-e > baade ‘lake (OBL)’

motùn-e > motùno ‘yam (OBL)’

However, vowel harmony is no longer entirely productive. Numerous nativised borrowings from other languages, as well as some recently grammaticalised affixes, do not comply with its rules consistently:

botsi (< Tsi pòǂɪ) - ‘puma’

Moreover, the /a-ɑ/ distinction is no longer maintained; [ɑ] merged with [o] in unstressed syllables and [a] in stressed syllables. As a result which set of endings is triggered by a stem with /a/ is unpredictable and often lexicalised:

baa[d]-e > bade ‘lake (OBL)’

amàng-e > màngo ‘seed (OBL)’

Usually, foreign borrowings are assigned to the front class irrespective of their actual vowels, but not necessarily:

botsi-e > botse ‘puma (OBL)’

Retroflices

Retroflices have now become essentially separate phonemes but they originated as allophones of stop-liquid clusters. They still regularly appear as the surface realisation of these clusters in numerous morphological contexts:

sébesíí[d]-r > sébesííɖa ‘professional dancer’

A cluster of /t k/ + /r l/ manifests as a retroflex /ʈ/:

Yii[t]-r > yiiʈa ‘jackal’

Voiced equivalents manifest as /ɖ/.

/s x/ in a cluster with /r/ manifest as /ʂ/, /z ɣ/ as /ʐ/.

imbas-ràng > imbaaʂàng ‘writing’

A cluster of /n ŋ/ with /r/ produces /ɳ/.

Nominal inflexion

Nouns in Risha Cuhbi are relatively morphologically simplistic, especially when compared with other dialects of Cuhbi which retain historical case distinctions which Risha Cuhbi has levelled. All nouns have a bare form and a suffixed form; for animate nouns this corresponds to a distinction between nominative and oblique cases, whilst for inanimate nouns both nominative and oblique roles take the suffixed form. Both animate and inanimate nouns use the unmarked form in compound verbs, and both may additionally take a comitative suffix, which is attached to the suffixed form and distinguishes case, resulting in four forms per noun. There are three declension classes distinguished by the suffix which appears in the suffixed form: -a -e or -r.

-a nouns

The suffix -a (which descends from the Old Cuhbi accusative) always appears as -a but triggers some predictable variations in different kinds of stems. Stems ending in a nasal consonant are completely straightforward:

pàng-a > pànga ‘king, chief’

A stem ending in /a/ is essentially invariable in that the suffix absorbs the vowel:

kà-a > kà ‘man’

Stems ending in -Ci, -Ce, -ai, -ei usually realise the vowel as a semivowel /j/ instead and the tone of the syllable is retained:

Idì > Idyà ‘Idi’

pxai > pxaya ‘castrate’

Women’s speech generally transforms a /tj/ or /kj/ sequence into /c/:

Mate > ♀ Maca ‘Mate’ (♂ Matye)

Stems ending in -Co, -Cu, -au, -ou usually realise the vowel as /v/ instead and the tone of the syllable is retained:

ugù > ugvà ‘house’

Attached to a stem with a latent consonant the latent consonant is realised and the long vowel present in the suffixless form disappears:

pàà[d] > pàda ‘boy’

-e/o nouns

The suffix -e (reflex of the Old Cuhbi instrumental) largely appears with inanimate nouns. With a small number of exceptions, it attaches to words ending in non-latent consonants, /e/ or /o/; other words regularly take -r instead.

Non-latent consonants undergo no changes when followed by -e:

xòòny > xòònyo ‘bait’

/e/ and /o/ become semivowels when followed by /e/:

loro > lorvo ‘placenta’

In women’s speech, a /tj/ or /kj/ sequence usually becomes /c/:

tàtètàtè > ♀ tàtètàce ‘locust’

-r nouns

The suffix -r attaches largely to inanimate nouns ending in /i u a/ or a latent consonant. Some nouns ending in nasals also take it.

Following vowels, it is realised as either /r/ or /l/. Generally speaking, it is /r/:

nyotso > nyotsor ‘tear’

If the word already has an /r/ in it, however, /l/ appears as the effect of disharmony.

ngara > ngaral ‘noise’

Following a nasal, the two merge to produce a retroflex nasal:

tsapàng > tsapàɳ ‘bitter’

Following any latent consonant, the latent consonant appears and the long vowel is shortened; r appears as -ru/-ri:

maa[p] > mapri ‘bark’

Latent /t k/ merge with /r/ to produce /ʈ/:

takata[k] > taktaʈi ‘parrot’

Latent /d g/ merge with /r/ to produce /ɖ/:

raa[d] > raɖu ‘leaf’

Reduplication

Preduplication

 

Nouns may be partially preduplicated. The last syllable of the root is reduplicated and prefixed immediately to the word:

amang ‘seed’ > mangamang ‘some seeds’

ʂoho ‘weave’ > hoʂoho ‘pattern’

There is voicing dissimilation in stops: if the initial sound of the root is voiceless, the initial sound of the preduplicated syllable will be voiced, and vice versa:

tápa ‘papyrus’ > batápa ‘some papyrus’

nyadà ‘beans’ > tànyadà ‘some beans’

If the last syllable has a long vowel, the preduplicated syllable takes the stress and the old stressed syllable is shortened:

urààng ‘milk’ > ràànguràng ‘some milk’

Partial reduplication

Partial reduplication may occur in nouns or adjectives. The last syllable of the root is copied and suffixed:

Idì ‘Idi’ > Idìdì ‘little Idi’, ‘our Idi’

botsi ‘puma’ > botsitsi ‘puma cub’

If the last syllable has a long vowel, the reduplicated syllable has a short vowel – stress remains on the original syllable:

Temìì ‘Temi’ > Temììmi ‘little Temi’, ‘our Temi’

In older forms, retroflexes were reduplicated as their non-retroflex equivalents; that is to say, words with clusters of a stop and an /r/ across syllable boundaries ignored the surface realisation of these clusters as retroflexes and reduplicated only the part of the cluster in the final syllable:

boɽu ‘dog’ > boɽuru ‘puppy’

Although this remains the case in many existing forms, there is an ongoing shift towards copying retroflexes and disregarding the ‘underlying’ forms, bringing into question whether these underlying forms still exist in the language’s phonology:

dèɽe ‘clang’ > dèɽeɽe ‘clanging’

Where the final syllable has an onset and a latent consonant, the latent consonant is realised and an epenthetic /a/ or /o/ is inserted between the latent consonant and the onset.

raa[d] ‘leaf’ > radaraa[d] ‘shell’

taa[k] ‘squawk’ > takata[k] ‘parrot’

Full reduplication

Full reduplication typically involves repetition of the whole word. The words retain separate stress, but the first is invariable for case:

Làànge-Làànge, Làànge-Lààngya - Laange and family

Some long words only reduplicate the last two or three syllables. This is somewhat idiosyncratic:

nyutònguuyì-guuyì - the priest and his entourage

Many fully reduplicated forms have been lexicalised, with one portion - usually the first - losing its separate stress:

uràngurààng - dairy products

Verbal inflexion

The Cuhbi verb is by far the most synthetic part of speech, and its morphemes are short and undergo considerable fusion with one another, sometimes obscuring underlying forms. Where examples do not require a verb with different properties, we will use the root ʂi- ‘carry, take’ to demonstrate conjugation. Translations are very rough and glosses and explanations of function are not provided; for this see the syntax section. We have tried to give the simplest possible combination of morphemes to demonstrate a given rule, and to utilise that combination multiple times with only one alteration when a point needs to be demonstrated.

Prefixes

Personal prefixes

The personal prefixes must always co-occur with a voice prefix. They cannot occur on serial verbs (i.e. verbs without a voice prefix) and cannot co-occur with the conjunct prefix. They undergo significant fusion with the voice prefixes.

First person prefix k-

The first person prefix appears as k- before the passive prefix:

k-u-ʂi-ø > kuʂi ‘I am picked up’

It may also appear as k- before the stative prefix:

k-a-ʂi-ø > kaʂi ‘I am holding it’

Many male speakers have an allomorph q- before the stative prefix:

k-a-ʂi-ø > ♂qaʂi ‘I am holding it’

It is palatalised and appears as c- before the active prefix:

k-i-ʂi-ø > ciʂi ‘I am picking it up’

With a vowel-initial stem, the k- prefix fuses with a stative prefix and produces a retroflex ʈ-, merging with the second person:

k-a-eqi-ø > ʈeɡi ‘I’ve burnt’

Second person prefix t-

The second person prefix appears as t- before the passive prefix and the stative prefix:

t-u-ʂi-ø > tuʂi ‘you are picked up’

t-a-ʂi-ø > taʂi ‘you are holding it’

It appears as c- before the active prefix, thus merging in surface form with the first person:

t-i-ʂi-ø > ciʂi ‘you are picking it up’

With a vowel-initial stem, the t- prefix fuses with a stative prefix and produces a retroflex ʈ-:

t-a-eqi-ø > ʈeɡi ‘you’ve burnt’

Third person prefix n-

The third person prefix appears as n- before the passive and stative prefixes:

n-u-ʂi-ø > nuʂi ‘she is picked up’

n-a-ʂi-ø > naʂi ‘she is carrying it’

It appears as ny- before the active prefix, showing the same patalisation as the first and second person prefixes:

n-i-ʂi-ø > nyiʂi ‘she is picking it up’

With a vowel-initial stem, the n- prefix fuses with a stative prefix and produces a retroflex ɳ-:

n-a-egi-ø > ɳegi ‘she’s burnt’

Fourth person prefix y-

The fourth person prefix appears as y- before the active and stative prefixes:

y-i-ʂi-ø > yiʂi ‘she is picking it up’

y-a-ʂi-ø > yaʂi ‘she is carrying it’

Preceding the passive prefix with a non-vocalic stem, it also appears as y-:

y-u-ʂi-ø > yuʂi ‘she is picked up’

Preceding the stative prefix with a vocalic stem, it appears as i-:

i-a-egi-ø > ilegi ‘she’s burnt’

However, with a vocalic stem, preceding the passive prefix it appears as u-, reflecting historical vowel assimilation:

i-u-egi-ø > uvegi ‘she is being burnt’

Voice prefixes

The voice prefixes occupy the second prefix slot and always directly precede the root. They appear in all non-conjunct main verbs. They cannot co-occur with the conjunct prefix.

Active prefix i-

When the active prefix precedes a consonant-initial stem, it is invariably i-, whether preceded by a personal prefix or not:

i-ʂi-ø > iʂi ‘picks it up’

n-i-ʂi-ø > nyiʂi ‘she picks it up’

It triggers palatalisation in the first, second, and third person prefixes (k- t- n-):

k-i-ʂi-ø > ciʂi ‘I am picking it up’

t-i-ʂi-ø > ciʂi ‘you are picking it up’

n-i-ʂi-ø > nyiʂi ‘she picks it up’

If there is no personal prefix present and the stem begins with a vowel, it is realised as y-:

i-egi-ø > yegi ‘she burns’

With a vocalic stem and a personal prefix, it is elided entirely, although the palatalisation remains on a first, second, or third person prefix:

k-i-egi-ø > cegi ‘I am burning’

t-i-egi-ø > cegi ‘you are burning’

n-i-egi-ø > nyegi ‘he is burning’

Passive prefix u-

When the passive prefix precedes a consonant-initial stem, it is invariably u-, whether preceded by a person prefix or not:

u-ʂi-ø > uʂi ‘is picked up’

k-u-ʂi-ø > kuʂi ‘I am picked up’

Preceding a vocalic stem, it is realised as v­-:

u-egi-ø > vegi ‘is burnt’

When appearing between the fourth person prefix and a vocalic stem, the passive prefix remains v- but the fourth person prefix assimilates to u-:

i-u-egi-ø > uvegi ‘he is burnt’

Stative prefix a-

The stative prefix is a- preceding consonantal stems:

a-ʂi-ø > aʂi ‘is carried’

Preceding vocalic stems, it appears as l-:

a-egi-ø > legi ‘is burnt’

Some older speakers produce al- instead of l- in this context:

a-egi-ø > †alegi ‘is burnt’

When it co-occurs with a personal prefix k- or t- and a vocalic stem, the personal prefix becomes a retroflex ʈ- and the stative prefix is realised as a-:

t-a-eqi-ø > ʈeɡi ‘you’ve burnt’

When it co-occurs with a personal prefix n- and a vocalic stem, the personal prefix likewise becomes retroflex (in this case ɳ-) and the stative prefix is realised as a-:

n-a-egi-ø > ɳegi ‘she’s burnt’

When it co-occurs with a personal prefix y- and a vocalic stem, the personal prefix becomes i- and the stative prefix is realised as l-:

i-a-egi-ø > ilegi ‘she’s burnt’

Conjunct prefix o-

The conjunct prefix descends from an older form ho-. It cannot coexist with the stative, active, or passive prefix, or with personal prefixes, and can be viewed as occupying both of the prefix slots simultaneously. As such, it always directly precedes a root:

o-ʂi-ø > oʂi ‘who is picking up’

Preceding the vocalic roots i- ‘do’ i- ‘say’ and a- ‘sit’, as well as the imperfective allomorph of kà ‘go’ e-, it is typically realised as v-:

o-i-ø > vi

It is likewise realised as v- before the colour roots ela- ‘be yellow’ and egi- ‘be black’:

o-egi-ø > vegi ‘which is black’

Root modification

Venitive marking

The lemma form of a transitive verb or verb of motion is in the andative, and has a front vowel. In the case of roots with /i/ or /e/, the venitive is formed through straightforward backing of the vowel to /u/ or /o/ respectively:

n-i-u/ʂi-ø > nyiʂu ‘she carries it towards us’

In roots with /a/ the situation is slightly more complicated because the back counterpart of /a/ underwent a merger with /a/ in stressed syllables and /o/ in unstressed syllables. As a result, the surface form of the stem does not change and retains /a/:

n-i-ɑ/kà-t > nyikàa ‘she came’

However, venitive stems trigger back harmony, and all vowels thereafter in the verb are backed as a result:

n-i-a/kà-t-nii > nyikàntuu ‘it’s her who came’

The final epenthetic -à used to resolve some consonant clusters is replaced by -ò in venitive verbs:

n-i-a/kà-t-b > nyikààfò ‘she really is coming’

Reduplication

Partially reduplicated allomorphs exist for eleven roots. This allomorph compulsorily co-occurs with imperfective marking, although occasionally speakers produce variants without reduplication with no apparent difference in meaning (these may be production errors, or represent an incipient regularisation of the system, or possibly simply be rare variants derived from historic non-reduplicated forms):

i-ra-ø > iɽa ‘looking at me’

The reduplication is not entirely regular and it is easiest not to attempt to produce rules but instead simply list the reduplicated forms:

i- ‘speak’	aya-	ʔe ‘eat’	ʔehe
ra- ‘see’ fé ‘see[IRR]’	ɽa- fefé	fí ‘drink’	fifí
pà- ‘put on’	papà-	kì ‘pull’	kikì
xá- ‘push’	xaxá-	va- ‘lean against’	vava-
ba- ‘sense’	baba-	nà- ‘fall’	ndà-

 

Stress falls on the second syllable of the reduplicated form, suggesting that this was originally preduplication.

Suppletive roots

Kà ‘go’ has a suppletive imperfective formed with the root e-:

n-i-kà-ø > nye ‘he’s going’

Ra ‘see’ has a suppletive irrealis formed with the root fé:

i-fé-h=usu > iféfehusu ‘isn’t looking at it’

Both are otherwise regular and take the aspect-realis suffixes as normal.

Suffixes

Aspect-realis suffixes

Imperfective realis suffix -ø

This suffix is null for the vast majority of verbs and has no surface realisation:

n-i-u/ʂi-ø > nyiʂu ‘she carries it towards us’

Note however that in verbs with a reduplicated stem variant, this suffix triggers reduplication:

i-ʔe-ø > iʔehe ‘is eating’

Perfective realis suffix -t

This morpheme falls into the category of latent consonants. When it is the last suffix in a word, this morpheme typically appears as lengthening of the preceding vowel:

i-ʂi-t > iʂii ‘carried it’

When directly followed by the intensive suffix -b the -t is deleted with compensatory lengthening and the /b/ is devoiced and fricated to /f/:

i-ʂi-t-b > iʂiifà ‘really did carry it’

Directly preceding the intentive suffix -t the perfective suffix is elided with compensatory lengthening:

i-ʂi-t-t > iʂiità ‘tried to pick it up’

When preceding the inchoative suffix -s historically it took the form -ti/-tu (with vowel harmony), producing -tis/-tus:

i-ʂi-t-s > ­†iʂitisà ‘moved to pick it up’

These forms can still occasionally be heard, especially in set phrases, but more commonly the cluster /ts/ is found:

i-ʂi-t-s > ­iʂitsà ‘moved to pick it up’

When it precedes the inanimate applicative suffix -n the two consonants cluster and as is usual for clusters of this sort, the first consonant /t/ nasalises and the nasal denasalises, which produces the appearance of epenthesis (note also the down-shift of tone in the last syllable of the root).

i-ʂi-t-n > ­iʂìntà ‘was picked up in’

When it precedes the animate applicative suffix -r the two resolve to -ɽ with compensatory lengthening. No final -à/-ò is required, but one occasionally appears:

i-ʂi-t-r > iʂiiɽ ‘is picked up by’

When directly followed by the first-person suffix -kàn or -kàa -t is elided and the /k/ is fricated.

i-ʂi-t-kàa > iʂixàa ‘it was me who picked it up’

Note that historically compensatory lengthening occurred here. Some older speakers (who have -kàn consistently) in fact demonstrate the expected compensatory lengthening:

i-ʂi-t-kàn > ­†iʂiixàn ‘it was me who picked it up’

Younger speakers however do not have compensatory lengthening even with -kàn. However, although the long vowel of the personal suffixes (to which -kàn, the only exception, has apparently been assimilated by analogy) means that there is never a long stem vowel, stems with /e o/ are raised to /i u/, reflecting a historical lengthening to /iː uː/ and a later re-shortening to /i u/:

i-be-t-kàa > ibixàa ‘it was me who stood up’

With the second person suffix -tàa the perfective suffix is elided entirely, again with the vowel change in stems with /e o/:

i-ʂi-t-tàa > iʂitàa ‘it was you who picked it up’

i-be-t-tàa > ibitàa ‘it was you who stood up’

With the third person suffix -nii, as is standard with underlying CN clusters, the -t is nasalised and the -n denasalised, producing the appearance of epenthesis:

i-ʂi-t-nii > iʂintii ‘it was him who picked it up’

Note that although since there is no elision here we expect no historic compensatory lengthening and thus no vowel change, there is nonetheless a shift by analogy:

i-be-t-nii > ibintii ‘it was him who stood up’

When directly followed by the fourth-person suffix -yii -t undergoes palatalisation to -c:

i-ʂi-t-yii > iʂicii ‘it was that other who picked it up’

Again, the vowel shift is present in spite of the absence of historic compensatory lengthening:

i-be-t-yii > ibicii ‘it was that other who stood up’

Imperfective irrealis suffix -h

The imperfective irrealis suffix causes a downward shift in tone in the preceding syllable when it appears in the coda of that syllable:

i-ʂi-h=usu > iʂìhusu ‘isn’t picking it up’

Followed by the distributive suffix -g the two suffixes fuse: -g is devoiced to /k/. If other suffixes follow the resulting -k surfaces as -kì/-kù:

i-ʂi-h-g=usu > iʂìkàsu ‘isn’t picking each of them up’

i-ʂi-h-g-kàa=usu > iʂìkìkàasu ‘it’s his which aren’t being picked up one by one’

Followed by the intensive suffix -b the two suffixes fuse: -b is devoiced to /p/. If other suffixes follow this -p surfaces as pì/pù:

i-ʂi-h-b-n=usu > iʂìpìnàsu ‘it’s really not being picked up there’

Followed by the intentive suffix -t the two suffixes fuse: -t is fricated to /s/. If other suffixes follow, this -s surfaces as -sé/só:

i-ʂi-h-t-nii=usu > iʂìsékàasu ‘it wasn’t me who was trying to pick it up’

Followed by the inchoative suffix -s the suffix is elided entirely, leaving only -s, which followed by other suffixes is realised as -sé/só (note homophony with the intentive):

i-ʂi-h-s-nii=usu > iʂìsékàasu ‘it wasn’t me who was moving to pick it up’

Followed by the inanimate applicative -n the suffix is also entirely elided, leaving only -n:

i-ʂi-h-n=usu > iʂìnusu ‘it wasn’t being picked up in’

Followed by the animate applicative -r, the suffix fuses with the applicative and the two appear as surface -l:

n-u-ʂi-h-r=usu > nuʂìlàsu ‘wasn’t having his picked up’

Followed by any of the pronominal suffixes, the suffix is realised as -ʔa/-ʔo:

i-ʂi-h-kàa=usu > iʂìʔakàasu ‘it wasn’t me who was picking it up’

Perfective irrealis suffix -n

The perfective irrealis suffix -n causes a downward shift in tone in the preceding syllable:

i-ʂi-n=usu > iʂìnusu ‘didn’t pick it up’

Followed by the intensive suffix -b the two suffixes fuse with compensatory lengthening and produce a bilabial nasal:

i-ʂi-n-b=usu > iʂìimusu ‘really didn’t pick it up’

Followed by the intentive suffix -t the two suffixes fuse with compensatory lengthening and produce a single dental nasal:

i-ʂi-n-t=usu > iʂìinusu ‘didn’t try and pick it up’

Followed by the animate applicative -r the two suffixes fuse and produce -ɳ:

u-ʂi-n-r=usu > uʂìɳusu ‘didn’t have his picked up’

All of these may optionally be followed by -à/ò:

u-ʂi-n-r=usu > uʂìɳàsu ‘didn’t have his picked up’

Followed by the inanimate applicative suffix -n, the perfective irrealis suffix denasalises to /t/ and they produce a consonant cluster /nt/ which appears with -à/ò finally but preceding other (consonantal) suffixes appears as -n-te/n-to:

i-ʂi-n-n-r=usu > iʂìntèlàsu ‘wasn’t picked up in’

Followed by the inchoative suffix -s historically the form was -nyi/nu, producing -nyis/nus:

i-ʂi-n-s-n=usu > †iʂìnyisnàsu ‘wasn’t beginning to be picked up in’

However, as with the perfective realis, it is now more common to see an analogised -n-s-à/-n-s-ò:

i-ʂi-n-s-n=usu > iʂìnsànusu ‘didn’t begin to be picked up in’

Followed by any of the pronominal suffixes, there are two possibilities. One, previously a marked quality of male speech but now fading under the influence of Long Island Cuhbi and perhaps of the flow of innovation from women’s speech to men’s speech, is to have the suffix appear invariably as -nyi/nu:

i-ʂi-n-kàa=usu > ­†♂ iʂìnyikàasu ‘it wasn’t me who picked it up’

The other is for the consonants to cluster and nasal assimilation to occur (including denasalisation of 3p -ni), producing -ng-kàa -n-tàa -nii/-nuu -n-yii:

i-ʂi-n-kàa=usu > iʂìngkàasu ‘it wasn’t me who picked it up’

Modal suffixes

Intensive suffix -b

The intensive suffix -b appears as -b following the imperfective realis suffix:

i-ʂi-ø-b > iʂibà ‘really is picking it up’

Following the perfective irrealis suffix -n the two suffixes fuse and become -m, which may optionally appear with -à/ò if it is the final suffix:

i-ʂi-n-b=usu > iʂimàsu, iʂimusu ‘really didn’t pick it up’

With other suffixes following this combination appears as -mì/mù:

i-ʂi-n-b-n=usu > iʂimìnusu ‘really wasn’t picked up in’

Following the perfective realis suffix -t the two suffixes fuse with compensatory lengthening and b is devoiced and fricated to -f:

i-ʂi-t-b > iʂiifà ‘really did pick it up’

This appears as -fì/fù if other suffixes follow:

i-ʂi-t-b-n > iʂiifìnà ‘really was picked up in’

Following the imperfective irrealis suffix -h the two suffixes fuse and are realised as -pà/pò finally and pì/pù otherwise:

i-ʂi-h-b=usu > iʂìpàsu ‘really wasn’t being picked up’

Inchoative suffix -s

The inchoative suffix cannot co-occur with imperfective marking. With the perfective suffixes there are two alternate sets of forms. One, the older (and now falling out of use) set has the form -isè/usò, whether the suffix is final or followed by other suffixes:

i-ʂi-t-s > †iʂicisè ‘moved to pick it up’

Note that the venitive variant causes palatalisation of -n and -t to -ny and -c:

i-ʂi-n-s=su > †iʂinyisèsu ‘didn’t move to pick it up’

With the applicative suffix -r no vowel typically appears in final position. Instead the suffixes fuse and appear as -iʂ/uʂ, or -iʂi/uʂu if followed by other suffixes:

u-ʂi-t-s-r > uʂiciʂ ‘had his picked up’

More common is for the suffix to appear as -sè/sò if followed by other suffixes and -sà/sò otherwise, clustering with the consonant of the aspect/realis suffix:

i-ʂi-n-s=su > iʂinsèsu ‘didn’t move to pick it up’

An exception is with a perfective suffix and following -r, where forms in -iʂ/uʂ are much more common.

Intentive suffix -t

The intentive suffix appears as -t following the imperfective realis suffix -ø. With following suffixes it appears as -tè/tò; if it is the final suffix it appears as -tà/tò:

i-ʂi-ø-t > iʂità ‘tried to pick it up’

With preceding imperfective realis suffix and following animate applicative suffix, it fuses with the applicative suffix to become -ɽ:

u-ʂi-ø-t-r > uʂiɽ ‘isn’t picked up for’

For many female speakers -tè is variably realised as -cè:

i-ʂi-ø-t-kàa > ♀iʂicèkàa ‘it was me who tried to pick it up’

Following the imperfective irrealis suffix -h, the two suffixes fuse and -t is realised as -s. With following suffixes it appears as -sé/só; if it is the final suffix, it appears as -sà/sò:

i-ʂi-h-t=usu > iʂìsàsu ‘was not trying to pick it up’

If perfective realis -t precedes, the perfective suffix is elided entirely with compensatory lengthening and the intentive suffix appears as -tè/tò; if it is the final suffix, it appears as -tà/tò:

i-ʂi-t-t > iʂiità ‘tried to pick it up’

Following the perfective irrealis suffix -n, the two suffixes fuse and produce a single dental nasal with compensatory lengthening of the preceding vowel. This appears as -nè/nò before other suffixes; finally it can optionally take -à:

i-ʂi-n-t=usu > iʂìinàsu, iʂìinusu

Distributive suffix -g

The distributive suffix appears as -g after the imperfective realis suffix -ø. With following suffixes it appears as -gì/gù; with no following suffix, it appears as -gà/gò:

i-ʂi-ø-g-n > iʂiɡìn ‘is picking them up one by one’

Following the imperfective irrealis suffix -h, ­the two suffixes fuse and -g is devoiced to -k, which appears as -kì/kù with following suffixes and -kà/kò:

i-ʂi-h-g-n=usu > iʂìkìnusu ‘isn’t picking them up one by one’

Applicative suffixes

Animate -r

The animate applicative suffix typically appears as -r following the imperfective realis suffix -ø but preceding the nominaliser -i/u or the fourth person -yi:

u-ʂi-ø-r-i > uʂiri ‘who is picked up for’

It appears as -rà/-rò following a sequence of the imperfective realis suffix and the distributive suffix -ø-ɡ or the intensive suffix -ø-b:

u-ʂi-ø-g-r > uʂigrà ‘is having things picked up one by one for him’

It also appears as -rà/rò following a sequence of the perfective irrealis suffix and the intensive suffix -n-b (fused to -m):

i-ʂi-n-b-r=usu > uʂìmràsu ‘really isn’t having things picked up one at a time for him’

As a general rule, if it is syllable-final, it is realised as -l:

u-ʂi-t-t-r > uʂiitèl ‘was tried to pick up for’

Note however that a (strictly unneeded) paragogic -à/ò is often added to the end of verbal forms with ­-r as their final suffix, and in this case forms with -là/lò are in free variation with forms with -rà/rò (although the former are more common):

u-ʂi-t-t-r > uʂiitèlà, uʂiitèrà ‘was tried to pick up for’

When -r follows the the imperfective irrealis suffix -h, -h disappears entirely and -r is realised as -l (from a historic [lʰ]):

i-ʂi-h-r=usu > iʂìlàsu ‘wasn’t picked up for’

/r/ regularly fuses with other consonants to produce retroflex consonants (for a discussion of the phonemic status of retroflex consonants, see the phonological section). When -r follows the perfective realis suffix -t the two fuse and produce -ɽ, with compensatory lengthening:

u-ʂi-t-r > uʂiiɽ ‘was picked up for’

Following the perfective irrealis suffix -n, the two suffixes fuse and are realised as -ɳ with compensatory lengthening:

u-ʂi-n-r=usu > uʂìiɳusu ‘wasn’t picked up for’

Following the intentive suffix -s the -r usually fuses with -s to produce -iʂ/uʂ, or -iʂi/uʂu if followed by other suffixes:

u-ʂi-t-s-r > uʂitiʂ ‘[someone] tried to pick it up for him‘

Following the imperfective realis suffix and intentive suffix -ø-t the -r fuses with preceding -t and produces -ɽ:

u-ʂi-ø-t-r > uʂiɽ ‘[someone] is trying to pick it up for him’

Inanimate -n

The inanimate applicative suffix -n appears as -n following the imperfective realis suffix -ø:

i-ʂi-ø-n > iʂìn ‘is being picked up in’

It also appears as -n following the imperfective irrealis -h, which is entirely elided:

i-ʂi-h-n=usu > iʂìnusu ‘is not being picked up in’

With following suffixes it appears as -ni/nu; finally it appears as either -n or -nà/nò:

i-ʂi-h-n-nii=usu > iʂìniniisu ‘it’s not there that it’s being picked up’

When it follows the perfective realis suffix -t the two consonants cluster; -t is nasalised and -n denasalised to t, producing -n-ti/-n-tu preceding other suffixes and -n-tà/-n-tò finally:

i-ʂi-t-n > iʂìntà ‘it was picked up there’

When it follows the perfective irrealis suffix -n, the perfective suffix is elided with compensatory lengthening:

i-ʂi-n-n=usu > iʂìinusu - ‘it was not picked up there’

Before the nominaliser -i in venitive verbs -n is realised as -ny:

i-ʂi-ø-n-i > iʂìnyi ‘who picks it up’

Pronominal suffixes and nominaliser

This slot was relatively recently grammaticalised and as a result has undergone relatively little fusion with other morphemes. The presence of long vowels in the pronominal suffixes, however, cause stress shifts which can lead to wider changes in the word.

First person -kàa/kàn

The first person suffix -kàa triggers a regular stress shift away from the stem and shortening of a stem long vowel:

i-ʂi-t-b-kàa > iʂiifà-kàa > iʂifàˈkàa ‘it’s me who snatched it up’

For older speakers, the older form -kàn is used consistently and does not trigger a stress shift:

i-ʂi-t-b-kàn > †iˈʂiifàkàn ‘it’s me who snatched it up’

However, younger speakers (who typically have -kàa and -kàn in free variation) typically display the stress shift to -kàn and subsequent shortening of vowels in spite of the lack of a long vowel to trigger this process:

i-ʂi-t-b-kàn > iʂifàˈkàn ‘it’s me who snatched it up’

Following the perfective realis suffix -t the suffixes fuse and -k is fricated to -x. Older speakers consistently have compensatory lengthening before -xàn (the expected outcome of a historic [Vtk] sequence being [V:x]):

i-ʂi-t-kàn > †iˈʂiixàn ‘it’s me who picked it up’

Younger speakers, however, typically treat -xàn identically to -xàa (where the long vowel in the suffix has led regularly to a stress shift and shortening of the preceding vowel) and thus do not display compensatory lengthening:

i-ʂi-t-kàn > iʂiˈxàn ‘it’s me who picked it up’

i-ʂi-t-kàa > iʂiˈxàa ‘it’s me who picked it up’

Note that whilst in the form -xàa compensatory lengthening no longer produces a length contrast, the historic process still has some implications: stems with final-syllable /e o/ have their vowels raised to /i u/:

i-be-t-kàa > ibixàa ‘it’s me who stood up’

This is because historically Cuhbi tolerated more than one long vowel in a word and this combination of morphemes produced first *ibeexàa, then following the collapse of [eː] into [iː], *ibiixàa. Finally the shortening of non-stressed long vowels took place, reducing the word to its existing form ibixàa.

Following the perfective irrealis suffix -n, -kàa itself does not change but does trigger nasal assimilation to -ng-kàa:

i-ʂi-n-kàa-su > iʂìngkàasu ‘it wasn’t me who picked it up’

Elsewhere no morphophonological changes affect the form of -kàa:

Second person -tàa

The second-person suffix -tàa triggers a stress shift and shortening of the stem vowel:

i-ʂi-t-b-tàa > iʂiifà-tàa > iʂifàˈtàa ‘it’s you who snatched it up’

When it follows a perfective realis suffix -t the perfective suffix is elided completely, producing -tàa. As with the suffix -kàa this triggers raising of /e o/ in the last syllable of the stem to /i u/:

i-be-t-tàa > ibitàa ‘it’s me who stood up’

Elsewhere it appears invariably as -tàa with no further morphophonemic implications.

Third person -nii/nuu

The third person suffix -nii triggers a stress shift and shortening of the stem vowel:

i-ʂi-t-b-nii > iʂiifà-nii > iʂifàˈnii ‘it’s him who snatched it up’

Following a perfective realis suffix -t the suffixes metathesise (or fuse, depending on analysis) and produce -n-tii:

i-ʂi-t-nii > iʂintii ‘it’s him who picked it up’

Note that in spite of the fact that historically there was no compensatory lengthening here, this suffix nonetheless (presumably by analogy) triggers raising in the last syllable of a root with /e o/:

i-be-t-nii > ibìntii ‘it’s him who stood up’

When -nii follows the perfective irrealis suffix -n the -n is simply deleted. Conversely, although compensatory lengthening is expected here historically, no raising occurs:

i-ʂi-n-nii=usu > iʂìniisu ‘it wasn’t him who picked it up’

i-be-n-nii=usu > ibèniisu ‘it wasn’t him who stood up’

Elsewhere -nii appears consistently as -nii.

Fourth person -yii

The fourth-person suffix -yii triggers a stress shift and shortening of the stem vowel:

i-ʂi-t-b-yii > iʂiifà-yii > iʂifàˈyii ‘it’s him who snatched it up’

Note that the /j/ blocks vowel harmony, meaning that there is no form *-yuu even in venitive verbs:

i-u/ʂi-t-b-nii > iʂuufò-yii > iʂufòˈyii ‘it’s him who snatched it up’

Following perfective realis -t the suffixes fuse to produce -cii. This form occasionally exhibits vowel harmony but usually (presumably by analogy with -yii) does not:

i-u/ʂi-t-yii > iʂuut-yii > iʂucii, iʂucuu ‘it’s him who picked it up’

Likewise, following perfective irrealis -n the suffixes fuse to -nyii /ɳiː/, although the orthography does not indicate this:

i-u/ʂi-n-yii=usu > iʂùn-yii > iʂùnyii ‘it’s him who picked it up’

Nominaliser -i/u

The nominaliser -i/-u absorbs the epenthetic vowels added to many consonantal morphemes (i.e. it ignores paragogic/epenthetic -à/ò, ì/ù, è/ò etc):

i-u/ʂi-t-b-i > iʂuufò-i > iʂuufì ‘who snatched it up’

Preceded directly by either the perfective realis suffix -t or the intentive suffix -t the venitive nominaliser causes palatalisation to -c in women’s speech, producing -ci:

i-ʂi-t-t-i > ♀iʂiici ‘who tried to pick it up’

When the imperfective irrealis suffix -n or the inanimate applicative suffix -n precedes -i directly, the two suffixes fuse to -nyi:

i-ʂi-ø-n-i > iʂìnyi ‘where it was picked up’

The nominaliser has the form -y/-v (for andative and venitive) preceding the oblique suffix -e:

i-ʂi-ø-n-i-e > iʂìnye ‘where it was picked up (OBL)’

The paragogic -à/-ò

A number of contexts have been listed here where -à/-ò’s addition is optional, but sometimes occurs nonetheless. -à/ò began as an epenthetic vowel to break up final consonant clusters; there is an argument to be made that it is fast becoming a secondary, or even primary, marker of the venitive/andative in Risha Cuhbi. Its expansion by analogy into new contexts is reasonably long-established in women’s speech; its stigmatisation by men as a feminine trait may stem the change in men’s speech for a while but à/ò are nonetheless making some headway there, too.

The irregular verb ʁ-

ʁ- is by far the most common verb in Risha Cuhbi. Over half of all compound verbs use it as their main verb, and most new verbal coinages employ it in this function. Unlike other Cuhbi verbs, which are almost entirely regular (except for suppletive forms), ʁ- displays significant deviation from the established verbal patterns. The reasons for this are complex and include the tendency of commonly-utilised paradigms to maintain more irregularities (observable cross-linguistically), the absence of a vowel in historic ʁ-‘s stem (leading to clustering with aspect-realis suffixes and the prevention of their clustering with other suffixes with all sorts of implications for suffixial morphology) and the use of several suppletive forms. We will deal here exclusively with the ways in which ʁ-‘s morphology differs from other verbs.

Prefixes

Prefix morphology remains largely regular. Pronominal prefixes have the usual allomorphs before the different voice prefixes. The voice prefixes themselves, when they co-occur with the root ʁ- and not a suppletive form, are lengthened:

i-ʁ-t­ > iisà ‘did’

o-ʁ-t > uusà ‘who did’

Venitive marking

All of the venitive forms of ʁ- are typically formed by male speakers with the (regular) suppletive root no-:

i-V/ʁ-t > ♂inuu ‘did (in my direction)’

This form has been borrowed wholesale from the dialect of Long Island. The older Risha form, sometimes observable in older men but now typically restricted to female speech, is to instead use forms of the verb kà ‘go’:

i-V/ʁ-t > ♀ ikàa ‘did (in my direction)’

Aspect/realis suffixes

Imperfect realis

For ʁ- forms with no other suffixes compulsorily take final -à, and the root is realised as y- after the active and stative prefixes and v- after the passive and conjunctive, with compensatory lengthening of the preceding vowel:

i-ʁ-ø -> iiyà ‘is doing’

u-ʁ-ø -> uuvà ‘is being done to’

These forms can be followed directly by the nominaliser or the personal suffixes with no further changes beyond the stress shift and shortening of the long syllable:

i-ʁ-ø-kàa -> iyàkàa ‘it’s you who was doing’

i-ʁ-ø-i -> iiyi ‘that does’

Other imperfective realis forms - i.e. those with additional suffixes - are somewhat irregular because they largely retain the expected outcomes of various historical sound changes as opposed to analogising them away. Unlike the forms without further suffixes, these forms exhibit root-reduplication, which produces a root form yiʁ- after the active and stative prefixes and vuʁ- after the passive and conjunctive prefixes. This ʁ- then fuses with following prefixes in relatively predictable fashion. When the intensive suffix -b is added, it absorbs ʁ- and is fricated with compensatory lengthening:

i-ʁ-ø-b -> iyiivà ‘really is doing it’

When the distributive suffix -g is added, it absorbs the ʁ and fricates to -q with compensatory lengthening:

i-ʁ-ø-ɡ -> iyiiqà ‘does each of them’

When the intentive suffix -t is added, it absorbs the ʁ and fricates and voices to -z with compensatory lengthening:

i-ʁ-ø-t -> iyiizà ‘tries to do it’

The inchoative (-s) and intentive are homophonous:

i-ʁ-ø-s -> iyiizà ‘begins to do it’

Directly following the aspect suffix the applicative -r absorbs the ʁ- and becomes -ɽ with compensatory lenɡtheninɡː

u-ʁ-ø-r -> uvuuɽ ‘is having it done for’

The ʁ- is elided with compensatory lengthening preceding applicative -n with a drop in tone:

i-ʁ-­ø-n -> iyìin ‘is having it done in’

Further suffixation is regular:

i-ʁ-ø-b-n -> iyiivìn ‘where he really is doing it’

Perfect realis:

Attached to the verb ʁ- with no other suffixes, -t is realised as -sà/sò with compensatory lengthening:

i-ʁ-t -> iisà ‘did it’

Attached to the verb ʁ- with one of the pronominal suffixes, -t is realised as -si with underlying compensatory lengthening:

i-ʁ-t-kàa -> isikàa ‘it’s me who did it’

With other suffixes, the verb acts as if it had a suppletive stem sà to which a second perfective -t is attached and which is essentially completely regular:

i-ʁ-t-b -> isàafà ‘really did it’

Imperfective irrealis:

The imperfective irrealis appears as -q, with compensatory lengthening of the preceding vowel and following -à if it is the last suffix:

i-ʁ-h=usu -> iiqàsu ‘is not doing’

Historically, the imperfective irrealis with additional suffixes had a complicated set of irregular forms derived largely regularly through sound change. In modern Risha Cuhbi, however, analogy has led to [ːɣà-] being reanalysed as a stem in itself, which is then treated as if it regularly suffixed an -h (with all of the additional fusions with the -h suffix that regular verbs have). That is to say, if we ignore the anomaly of the forms in which the imperfective irrealis is the last suffix, the imperfective irrealis forms of ʁ- can essentially be treated as having a regular suppletive stem [ːɣà-] which takes -h and additional suffixes regularly:

i-ʁ-h-b=usu -> iiqàpàsu ‘really isn’t doing’

i-ʁ-h-n-yii=usu -> iqànyiisu ‘it’s not there it’s being done’

Perfective irrealis

Without any further suffixation, the perfective irrealis appears as /`ːn/:

i-ʁ-n=usu -> ìinusu ‘did not do it’

u-ʁ-n=usu -> ùunusu ‘was not done to’

Occasionally female speakers produce forms with final -à, but at present this is anomalous:

i-ʁ-n=usu -> ♀ìinàsu ‘did not do it’

When these verbal forms occur directly before personal suffixes and the nominaliser these forms are treated as if they were regular (this is probably because of the relatively recent grammaticalisation of this slot in the verbal complex; i.e. after the transformation of historic /ʁn/ into a single consonant), complete with the vowel alteration of conjunctive /o/ to /u/ in spite of a lack of historic justification for it:

i-ʁ-n-tàa=usu -> ìntàasu ‘it’s not you who was doing it’

o-ʁ-n-tàa=usu -> ùntàasu ‘that it’s not you who was doing it’

With other suffixes, the verb is treated as though its stem were [ːne] and an additional -n is affixed in much the same way as the imperfective irrealis, with regular fusion with following suffixes:

i-ʁ-n-b=usu > ìinèmusu ‘really wasn’t doing’