1

Chapter ##

FROZEN PLURACTIONALS

hapter ## describes three verbal pluractional formation processes in Bole—one
productive process of initial CV- reduplication, available for essentially all verbs,
and two active but non-productive processes, viz. infixing -Ki- within a verb root

and geminating the second consonant of a root. The latter two active processes apply
only to a limited number of CVC- verb roots. Lexical evidence suggests that the latter
two processes at one time may have been more productive than they now are and may
have applied to a larger range of root types than they now do. This evidence is in the
form of “frozen pluractionals”, for which unmodified roots are no longer in use and
which do not convey obvious pluractional meaning (see the introductory section of
Chapter ## for a description of canonical pluractional meanings). This chapter presents
several patterns of frozen pluractionals. The concluding section presents a
comparative/historical perspective in which we argue that all Bole geminate and infixed
pluractionals have their source in an infixed root consonant.

1. Frozen Pluractionals with a Geminate Consonant

1.1. CVGV verbs (G = geminate consonant). A feature of Bole that strikes anyone
first approaching the language is the large number of verbs of the form CVGV. In our
currently available list of nearly 700 verbs, 117 are of this type, excluding those that have
an active relationship to roots without a geminated C

2

and those not found in the Fika

dialect. Of the total 117, 107 are in Class A2 (the morphological correspondent to Class
A1 non-pluractionals) and 10 are in Class B2 (the morphological correspondent to Class
B non-pluractionals). What is particularly striking is the proportion of CVGV verbs in
the verbal lexicon. The total number of verbs with two distinct root consonants (C

2

=

singleton or geminate) is 302. Thus, over one third of all biconsonantal roots have a
geminated C

2

. Additionally, there are 38 verbs of the type CVGV that have a transparent

connection to existing CVCV roots. Below are some examples of Bole CVGV roots.
Where available, we have included cognates from the Yaya dialect of Ngamo (Ng) and
Karekare (Ka) to show that Bole has not inherited geminate consonants in these verbs
from an ancestral language.

1

The forms are cited in the completive, marked by a suffix

that comes from earlier *-kò. In some cases the verbs belong to different classes (A1, A2,
B) in the different languages.

1

Cognate items with the same meanings are more readily available between Bole and Ngamo than between

Bole and Karekare because Ngamo is linguistically closer to Bole than is Karekare. In the Yaya dialect of
Ngamo, Class A2 verbs in the completive end in -ô, sometimes pronounced -o’ò < *-u-kò. Class B verbs
end in -â, sometimes pronounced -a’à < *-a-kò.

C

2

Class A2 verbs

boækku-wo¥yi

‘burn, roast’

Ng baæka]

daæppu-wo¥yi

‘gather, collect’

Ng da¥fo]

fioæppu-wo¥yi

‘follow’

Ng fia¥fo] Ka fia¥fu-kaæu

goæjju-wo¥yi

‘buy’

Ng kaæja]

goællu-woæ

‘for dawn to break’

Ng gol-kò Ka gàl-kâu ‘spend the day’

kaæppu-wo¥yi

‘plant, sow’

Ng kap-kò

kuænnu-wo¥yi

‘thank’

(Ngamo kuænno] bor. from Bole)

2

mbìttu-wo¥yi

‘extinguish (fire)’

Ng mbit-kò Ka mbìtu-kàu

noæssu-woæ

‘spend time, rest’

Ng nas-kò

oæppu-wo¥yi

‘dig’

Ng op-kò Ka aæfa\-kaæu

oæssu-wo¥yi

‘grind’

Ng es-kò Ka &yaæsu-kaæu

poæ&&u-woæ

‘dry up’

Ng ho&i-koæ

poæyyu-woæ

‘break, shatter’

Ng hoi-kò

roækku-wo¥yi

‘drive away’

Ng ràkâ Ka ràku-kàu

suærru-wo¥yi

‘fry’

Ng sur-kò Ka sìr-ka]u

tuæbbu-wo¥yi

‘push”

Ng tu¥bo]

Class B2 verbs

baætta\-wo¥yi

‘trick, scare’

bìtta\-wo¥yi

‘expand an opening’

kaæbba\-woæ

‘swell’

Ng kàbâ (also Bole kaæwa\- ‘be sated’)

maæssa\-woæ

‘remain, be left’

sìfifia\-wo¥yi

‘wipe, wipe off’

Ng sa&a-koæ ¶safi-koæ¶ ‘rub’

The distinction between frozen pluractionals and “active” geminate pluractionals is

blurry. Among active pluractionals we have included some verbs where there is a clear
form and meaning relationship between an underived verb and a verb with geminate C

2

,

but where one or both verbs have shifted in meaning such that the geminated form does
not bear a canonical pluractional relationship to the simple form. Consider the following
examples.

Class A1 verb

Class A2 verb

ga&yu-woæ

‘feel sudden pain’

gaæ&y&yu-wo¥yi

‘chop’

go∫-wo¥yi

‘sip’

goæ∫∫u-wo¥yi

‘consume liquid with hand’

moyu-wo¥yi

‘observe, look’

moæyyu-wo¥yi

‘wait for’

pulu-woæ

‘boil; be angry’

puællu-wo¥yi

‘abuse’

washu-wo¥yi

‘nip off’

waæccu-wo¥yi

‘trim (e.g. hair)’

2

Bole has had considerable influence on Ngamo. The fact that Ngamo has few words with geminate

consonants and does not form pluractionals with gemination makes it certain that this word is borrowed.
Most Yobe State languages have borrowed the Hausa root gode to mean ‘thank’.

3

Class B verb

Class B2 verb

bìfia\-wo¥yi

‘untie’

bìfifia\-wo¥yi

‘untangle’

kuæla\-wo¥yi

‘widen pot mouth’

kuælla\-wo¥yi

‘scrape out’

zuæla\-woæ

‘protrude’

zuælla\-wo¥yi

‘glare at’ (protrude eyes?)

In some of these verb pairs, it is difficult to see what is “pluractional” in meaning

about the geminated form. Thus, while the relationship of goæ∫∫u-wo¥yi ‘consume liquid
with hands’ to go∫u-wo¥yi ‘sip’ could be understood as “taking repeated sips”, it is not so
obvious how puællu-wo¥yi ‘abuse’ could be understood as “repeated boilings” or “repeated
angerings” from pulu-wò.

As for frozen CVGV pluractionals with no extant CVCV counterpart, most have no

obvious element of plural action. Although some verbs like kaæppu-wo¥yi ‘plant’ or
oæssu--wo¥yi ‘grind’ involve repetitive action and thus could credibly have their origin in
pluractional derived forms, no such natural relationship to plural action holds for fioæppu-
wo¥yi ‘follow’, suærru-wo¥yi ‘fry’, kaæbba\-woæ ‘swell’, or, in fact, most of the 117 CVGV
verbs that we have identified. Bole has numerous morphological environments that
create geminate consonants (see for example, ##, ##, ##). It seems that Bole has a
Sprachgefühl that welcomes gemination. Once Bole developed gemination as a method
of pluractional formation, some of these geminate pluractionals became the primary verbs
and the underived bases were lost, obscuring the origin in pluractionality. The door was
then opened to geminating C

2

of certain verbs just because it “sounded good”, without the

requirement that plural action be a necessary semantic corollary.

3

1.2. CVGVCV verbs (G = geminate consonant). In modern Bole, the only active
geminate pluractionals are based on class!A1 and class!B verbs, i.e. verbs with root
structure C

V

¨C-. Likewise, most underived verbs that have the same form as geminate

pluractionals (CVGV) have just two root consonants. There are, however, a few verbs
with three root consonants where C

2

is geminate. Since both comparative Chadic and

internal Bole evidence point away from inheritance of underived geminate consonants,
these CVGVCV roots suggest that geminate pluractional formation may have at one time
applied to a larger set of verbs than just those of classes A1 and B. The list below is
exhaustive for currently available materials.

bìccìru-wo¥yi

‘expand s.th. circular, e.g. fan’

fiuækkuæru-wo¥yi

‘knead, form into balls’

= fiuækku-wo¥yi

gìccìyu-woæ

‘be crooked, be tilted’

koællìmi (noun)

‘drumming’

cf. kòlmò ‘praise singing’

kuæsshìmu-wo¥yi

‘wad up’

3

A few Bole CVGV verbs may come from *C

1

VC

2

C

3

V, with assimilation of C

2

to C

3

, e.g. Bole

shàttu- ‘uproot’—cf. Ngamo shapto ‘extract, pull out’. However, nearly all Bole CVGV verbs with
identified cognates in other languages correpond to CVC- roots in those languages. Moreover, the most
common way for Bole to resolve undesirable consonant sequences is epenthesis, e.g. Bole yaæ∫ìru- ‘scatter’,
Ngamo ya∫ro.

4

laækkìleæ (noun)

“jacks”

(a game involving flipping and

catching a stone)

lìkkìtu-woæ

‘descend en masse, e.g. flock of birds’

loækkìfiu-woæ

‘become entangled’

= loækku-wo¥yi

muækkuæfiu-wo¥yi

‘stuff the mouth with food’

muæmmusu-wo¥yi

‘rub back and forth between palms’

= muærmuæsu-wo¥yi

taækkìlu-woæ

‘become flustered’

toæggìfiu-wo¥yi

‘stitch’

yaækkìnu-wo¥yi

‘recognize’

zuæ&y&yìru-wo¥yi

‘purse the lips (to show disdain)’

Most of these verbs involve inherent plural action of some sort—repetitive action

and/or action by several individuals—reinforcing the notion that they are frozen
pluractionals. A couple have the form of verbal nouns but the source verb, if there ever
was one, is no longer in use.

Two of the verbs (‘knead’ and ‘become entangled’) are related to verbs of the CVGV

type. Bole, like other West Chadic languages, has a number of “remnant verbal affixes”
(##)—non-productive suffixes of uncertain function. For the two verbs in question, it is
hard to say whether the affixes -r- and -fi- have been attached to a CVGV stem or the
medial consonant of the affixed verb was geminated. Another verb pair not included in
the table is goæjju-wo¥yi ‘buy’ vs. goæjjìtu-wo¥yi ‘sell’. In this case, we know that a semi-
productive causativizing affix -t- (##) has been added to a CVGV stem. In the case of the
verb meaning ‘rub back and forth between palms’ it is hard to be sure whether this is a
CVGVCV verb or a verb with an infixed copy of the initial CV, described in §3.

2. Frozen Pluractionals with an Infix -Ki-

Chapter ##, §2.1 describes pluractional verb formation by infixation of a syllable

-gi- in the Fika dialect, e.g. &yoærangoæ ‘they stopped’, pluractional &yoægìrangoæ ‘they
stopped one by one’. That section also mentions that the Gadaka dialect productively
forms pluractionals of Class A1 and B verbs by infixing -ki-/-ku-/-k- (quality of vowel or
absence of vowel determined by context). Bole has numerous verbs with an internal
syllable -Ki- (K!= a velar stop, “i”!= [i], [u], or in a few cases, Ø) which formally
resemble infixed pluractionals and many of which, like the geminated verbs described in
§1.2, involve inherent plural action. The lists below are exhaustive for currently
available materials, omitting pronunciation variants. Those on the left are formally
identical to active infixed pluractionals, i.e. the -Ki- syllable is sandwiched in a frame
CV__CV. Those on the right have an additional consonant abutting the -Ki- syllable.
We have grouped the latter verbs according to the consonant that precedes the
-Ki- syllable.

CVKiCV

CVCKiCV

baægìdu-wo¥yi

‘pound grain 2

nd

time’

baængìlu-woæ

‘become confused’

buækuælu-wo¥yi

‘roll along’

deæænkìfiu-wo¥yi

‘stuff into’

joægìlu-wo¥yi

‘stir a thin substance’

dìnkìfiu-woæ

‘be crowded’

5

juæguælu-wo¥yi

‘churn milk’

fiaængìlu-wo¥yi

‘cling to, adhere to’

leækìtu-wo¥yi

‘disperse; spread (news, etc.)’

fiaængìlu-woæ

‘put on airs’

moækìtu-wo¥yi

‘get wet; moisten’

fiìngìmu-woæ

‘submerge’

njeækì∫u-woæ

‘become sloppy’

laænkìnu-wo¥yi

‘rinse’

pìktun-go

4

‘come to (after fainting)’

rìnkì∫u-woæ

‘be well dressed’

ruæguæzu-wo¥yi

‘raze, destroy’

shaænkìlu-wo¥yi

‘rinse’

sìkìtu-wo¥yi

‘reduce in amount’

suænkuænu-wo¥yi

‘sniff at, smell’

taækuæmnu-wo¥yi

5

‘chew’

zuænkuænu-wo¥yi

‘singe’

toækìlu-wo¥yi

‘stir with small motions’

baærgìlu-woæ

‘dislocate, become sprained’

yaækìlu-wo¥yi

‘lack’

bìrkìtu-wo¥yi

‘confuse, become confused’

zoægìnu-wo¥yi

‘toss, flip’

buærgìlu-wo¥yi

‘stir with swizzle stick’

zuækuælu-wo¥yi

‘stir thin subtance’

∫oærkìtu-wo¥yi

‘locate a gap or soft spot’

maærkìfiu-wo¥yi

‘wring out, twist’

za¥gìru-woæ

‘shout’

pìrkìtu-wo¥yi

‘make hole by prodding’

baæskìnu-wo¥yi

‘curse’

muæskuæfiu-woæ

‘feel nauseous’

paæskìfiu-wo¥yi

‘translate’

wìlkìfiu-wo¥yi

‘pick apart, pick open’

At our current state of knowledge, we cannot be sure that all of these are frozen

pluractionals with a -Ki- infix rather than a root that happened to have a velar stop as one
of its consonants. For example, corresponding to buærgìlu- ‘stir with swizzle stick’ is
Gadaka buærgu-, which may or may not be a borrowing from Hausa burga¥. The verb
fiaængìlu- ‘put on airs’ may be related to fiìngu-woæ ‘have a concave back’ (referring to
the posture of someone who puts on airs?). Note also that the final consonants of the
verbs in the lists above comprise a restricted set of the Bole consonant inventory, esp. l
(14 cases!), n, t, and fi. These are among the most frequent “remnant affix” consonants
(##) and hence could be additions to a CV(C)KV- root rather than being the original final
consonant of a root containing an infixed -Ki-. Still, there is evidence that at least some
of these do derive from infixation. The verb maærkìfiu- ‘wring out’ coexists with
maærfiu- in the same meaning. The verb muæskuæfiu- ‘feel nauseous’ has a Gadaka dialect
counterpart muæsfiu-. The verb wìlkìfiu- ‘pick apart, pick into’ may be related to yùllu-
‘investigate’. The verbs sùnkùnu- ‘sniff at’ and tàkùmnu- ‘chew’ look to be cognate
with Hausa sansaæna\ and tamna¥ (= tauna¥ in “Standard” Hausa) respectively.

Note in the right hand column that the range of consonants that precede putative infix

-Ki- is restricted to the alveolar sonorants (n, r, l) and s. In the left hand column, all the

4

The stem pikt- is used only with the ventive extension (##), parallel to Hausa fa®fafio\, with the same

meaning. Furthermore, the Hausa verb is pluractional in form with no transparently related simple stem in
current use. The typological parallel of verbs in certain meanings having the form of frozen pluractionals
holds between Bole and Hausa (and undoubtedly for other Chadic languages) for a number of verbs. Note
in the lists here, for example, Bole suænkuænu-, Hausa sansaæna\ ‘sniff at’ (with a cognate root) and Bole
zuænkuænu-, Hausa babbaæke\ ‘singe’, the latter cognate with Bole bòkku- ‘burn, roast’, which has the form
of a frozen CVGV pluractional.

5

This verb has an additional final stem consonant, probably a remnant affix (##).

6

verbs except za¥gìru- ‘shout’ have short vowels in the intial syllable. We return to the
significance of these skewings in section 4.

3. Verbs of the form C

1

VC

1

iC

2

V

Several verbs have what appears to be an infixed -Ci- syllable, where C = repetition

of the first stem consonant rather than K. The lists below are exhaustive for currently
available data. In addition to verbs, there are a few nouns that looka as if they could be
verbal nouns.

C

1

V

|C

1

iC

2

V

C

1

VC

2

C

1

iC

3

V

bÏbìyu-wo¥yi

‘swing around, whirl’

buæmbuælu-wo¥yi

‘roll on the ground’

gÏgìyu-wo¥yi

‘tie rope tightly around’

muærmuæsu-wo¥yi

‘rub between palms’

ja¥jìnu-wo¥yi

‘groom a horse’

ngìrngìru-wo¥yi

‘cling to’

ka¥kìlu-wo¥yi

‘suck particles from teeth’

ko¥kìlu-wo¥yi

‘gnaw’

mu¥muæsu-wo¥yi

‘roll thing around’

ngu¥nguæru-wo¥yi

‘mumble, murmur’

pÏpìlu-wo¥yi

‘nibble at’

pu¥puæru-wo¥yi

‘shake dirt loose from roots’

rÏrìmu-wo¥yi

‘become quite, abate’

sa¥sìlu-wo¥yi

‘touch lightly’

so¥suwu-wo¥yi

‘touch lightly with hand’

su¥suæwu-wo¥yi

‘whisper’

te¥tìyu-wo¥yi

‘try hard’

la¥luæmoæ

‘dried guinea corn leaves’

pu\puro

‘lumps of flour in atti’

ro¥rìyoæ

‘trembling, shivering’

Like CVGVCV verbs described in §2.1 and CV(C)KiCV verbs described in §3, most

of the verbs with infixed -Ci- involve inherent plural action of some kind, and though we
have not yet found relationships between any of these verbs and CVCV stems, some of
them have meanings that other languages express with verbs that are pluractional in
form—cf.ko¥kìlu- ‘gnaw’ with Hausa gwêgwìya\ in the same meaning.

4. Historical Perspective on Bole Pluractional Formation

Frozen pluractional morphological types are in near complementary distribution

corresponding to the structure of the hypothetical bases from which the pluractionals
were derived:

6

6

In labializing environments, a few of the frozen pluractionals have -u- in the infixed syllable rather than

the -i- in the schemata below. We list only patterns for verbs with two or more original base consonants.
There apparently are no frozen pluractionals derived from monoconsonantal bases.

7

• *C

1

V

¨C

2

V

> C

1

VG

2

V (§1.1), C

1

V

¨KiC

2

V (§2)—see below for further comment

• *C

1

V

|C

2

V

> C

1

V

|C

1

iC

2

V (§3)

• *C

1

VC

2

C

3

V > C

1

V®

2

KiC

3

V (® = n, r, l, s) (§2), C

1

VG

2

iC

3

V (G ≠ n, r, l, s) (§2)

Gimba (2000:161) pointed out that only verbs with a sonorant or fi as the second root

consonant have active infixed pluractionals (##). Most of the verbs listed in §2 likewise
conform to this restriction, the only exceptions being verbs where the final consonant is t,
one of the most common remnant affixes as well as being a semi-productive transitizing
affix (##), and the verb ruæguæzu- ‘raze, destroy’, where the -g- may be an original root
consonant rather than an infix (cf. Hausa rugu®guæza\ ‘smash; crumble’, a frozen
pluractional from a base *rug(u)za). The complementarity in frozen pluractionals would
cover all verbs with two or more base root consonants if the C

1

VG

2

V pattern applied only

to C

1

V

¨C

2

V roots where G

2

is a consonant

OTHER THAN

a sonorant or fi. This is not the

case, but in §1.1, we provided evidence that the C

1

VG

2

V pattern has taken on a life of its

own, and Bole has extended this pattern lexically to many verbs where plural action is not
a factor.

Comparative evidence points to the path which Bole has followed to arrive at this set

of pluractional patterns from a single source. In the Chadic languages of Yobe State
other than Bole, the prevailing pattern for pluractional verb formation is

REDUPLICATIVE

INFIXATION

. This is true of both the West Chadic B languages, Bade and Ngizim, and the

West Chadic A languages, Karekare and Ngamo. We present data from Karekare and
Ngamo (Yaya dialect), the languages most closely related to Bole within Yobe State.
Boxes enclose the reduplicative infixes.

KAREKARE

7

Verb class

Base form

Pluractional

A1

fiaæfu\-si-kaæu

fiaæ faæfu\-si-kaæu

‘he spoke’

aæsu\-si-kaæu

aæ saæsu\-si-kaæu

‘he picked up’

B

baæsa\-si-kaæu

baæ saæsa\-si-kaæu

‘he shot’

A2

fiìbaætu\-si-kaæu

fiìbaæ taætu\-si-kaæu

‘he sold’

NGAMO

8

A1

nì dugo-k koæsu

nì duædgo-k koæsaæsaæ

‘I killed a mouse/mice’

at-na\"

aæ kti-na\"

‘he bit me’

B

nì baæsa-k ra¥yi

nì baæ bsa-k ra¥yaæku

‘I shot a bird/birds’

A2

tu¥bo-k aæmba

tuætu¥bo-k bìya

‘he sent a boy/peopleÆ

nì aædgo-k zuægoæno

nì aæ&aædgo-k zuægoæno

‘I scratched my body’

7

In the Karekare forms, -si- is the mark of the totality extension when no object is expressed; -kàu is the

completive suffix. A common feature of Yobe State languages is to use the completive plus the totality
extension when translating Hausa sentences in the completive.

8

In the Ngamo forms, -k is the mark of the totality extension before nominal objects. See the previous

footnote for use of the totality extension.

8

Karekare derives all pluractionals with two or more consonants by infixing

-Ca- where C = the final stem consonant. Ngamo forms pluractionals of C

1

V

¨C

2

V roots

(i.e. verbs of Class A1 and B) by infixing a copy of C

1

after the first CV. If the verb has a

Ø onset, the verb infixes -k- after the first CV. For longer verbs, i.e. those of Class A2,
Ngamo prefixes a copy of the first CV-, with V being short for all verbs, including those
with a long vowel in the first syllable of the base. For bases wih Ø onset, glottal stop is
the default onset separating the initial syllable of the base from the reduplicated syllable.

Using this evidence, we hypothesize that all the frozen pluractionals in Bole as well

as all active CVGV and CVKiCV pluractionals (##) derive from a pattern
[XC

R

Y]

pluractional

, i.e. infixation of a root consonant. The -i- that shows up in pluractionals

other than the CVGV type is epenthetic. Ngamo is far more tolerant than Bole in terms
of permissible consonant sequences. Bole has resolved consonant hiatus in two ways, viz.
epenthesis of -i- or complete assimilation, resulting in a geminate. Epenthesis is always
the choice when the syllable preceding the infixed consonant is/was heavy (C

1

V

|C

1

iC

2

V,

C

1

V®

2

KiC

3

V, C

1

VG

2

iC

3

V patterns). Current evidence also suggests that epenthesis was

the choice if infixation would have resulted in a sequence [obstruent]-[sonorant],

9

a

sequence also disallowed in some otherwise “sequence tolerant” languages, such as
Ngizim. Other consonant sequences have resulted in geminates. We reiterate the
observation that at some point, Bole began extending geminate formation as a process in
its own right without requiring a source form a C

x

C

y

sequence. This tendency in Bole has

obscured the reconstructed source for geminates.

This account has ducked two questions: why does Bole infix -Ki- rather than -C

R

-,

and which base consonant is it that was infixed? Ngamo suggests an answer by the fact
that it infixes -k- as a default for verbs with onsetless first syllables. If this was a feature
of proto-Bole-Ngamo, Bole has simply extended the default pattern to all infixed
syllables. Another possibility, not mutually exclusive with this one, is that -K- originally

WAS

one of the root consonants. In a verb like gowu- ‘beat’, the still active infixed

pluractional is gògìwu-. The syllable -gi- is potentially ambiguous as to source—is it
reduplication of the initial g- or is it an infix of invariable form? If Bole speakers
interpreted it in the latter way, it would automatically be the pluractional marker for any
verb, replacing reduplicative infixation. Both these sources for -Ki- infixation fail to
answer the question as to why the infix is -k- rather than, say, -s-. We will not attempt to
answer this question, since any answer would be idle and unproductive speculation.

The second question raised above is which base consonant was reduplicated.

Karekare reduplicates the last consonant, a type of reduplication that could account for
most of the Bole forms, e.g.

*pata-

> pa-t(V)-ta-

> patta-

‘go out’

*&yoru-

> ’yo-r(V)-ru- > ’yogiru-

‘stop’

*pas(i)fiu- > pas-fi(i)-fiu- > paskifiu-

‘translate’

An advantage of this source is that it would get the common CVGV pattern directly,

i.e. the geminate consonant would be, literally, a doubling of the last consonant. It would

9

Epenthesis also applies before fi. This is not a sonorant, of course, but implosive/laryngealized

consonants pattern with sonorants in a number of ways in Chadic languages and elsewhere.

9

be hard to understand, however, why words like ’yogiru- ‘stop repeatedly’ (from ’yoru-)
would have an infix rather than a geminate *’yorru-. It seems more likely that the
infixed consonant was C

1

, as in Ngamo. Not only is Ngamo a closer cousin of Bole than

is Karekare, but also Bole still has a remnant of C

1

infixation in verbs like rÏrìmu- ‘shake

dirt loose’ < *rêmu-. The path for deriving the modern geminate and -Ki- pluractionals
from C

1

infixation would thus be as follows:

*pata-

> pa-p(V)-ta-

> patta-

‘go out’

*&yoru-

> ’yo-’y(V)-ru- > ’yogiru-

‘stop’

*pas(i)fiu- > pas-pi-fiu-

> paskifiu-

‘translate’

As described in Chapter ##, the productive pattern of pluractional formation for all

classes of verbs in modern Bole is initial CV- reduplication, e.g. baæsa\- ‡

baæbaæsa\- ‘shoot’. This pattern has its counterpart in Ngamo Class A2 pluractionals. The
simplest explanation for why modern Bole has initial CV- reduplication for

ALL

verb

classes is that Bole has extended a pattern that was at one time more restricted. The
motivation for such an extension would be simplification of the complex system
pluractional formation that had developed where some verbs formed pluractionals by
gemination, some by -K- infixation, and some by C

1

infixation. This suggestion raises

the question, however, of where initial CV- reduplication came from in the first place.
This is a restricted pattern at best in West Chadic (Hausa being one of the few languages
where it is dominant). We put the answer to this question aside as a subject for future
research.

10

Verbs of the following forms?

jakkutu (is there a verb?)

p loppinu
b lobbinu
b' lob'b'inu
t lottinu
d loddinu
d' lod'd'inu
z/j lozzinu, lojjinu

We have examples with s, 'y, k, g, but more would be good.