Gbe

Sandro CAPO CHICHI

1. Introduction

Gbe is the name given to a cluster of languages spoken in an area stretching from
southeastern Ghana to the west to southwestern Nigeria to the east. It is bordered
by Yoruba languages to the east and to the north-east; by Akan, Guang, Ga-
Dangme languages to the West; and GTML to the west and north-west. The
several dozens of Gbe languages are spoken by about eight million speakers. The
use of Gbe, which means ‘language’ in all the languages of the cluster to name it,
was adopted in the Fourteenth West African Languages Congress at Cotonou,
Bénin, in 1980. It has been mostly accepted in the specialized literature since then.
The Gbe language group has been the object of several works of classification,
both internal and external. The issue of Gbe’s external classification will be dealt
with in another work (Capo Chichi & Kambon, this volume). After having
introduced the reader to these major attempts of Gbe’s internal classification, we
will provide a reconstruction of the form and of the function of three Proto Gbe
nominal affixes. We will then introduce the reader to Capo (1991)’s previous
reconstruction of Proto Gbe’s phonology. We will then justify our own
modifications to it. This work will end with a list of 100 reconstructed Proto Gbe
words.

2. Gbe’s internal classification

The two major attempts to classify Gbe internally have been made by Capo (1991),
on phonological grounds and Kluge (2011), mainly on lexical and grammatical
grounds. According to the former, it can be clustered into five main groups,
namely Ewe / Vhe, Gen, Aja, Phla-Phera and Fon (1).
Ewe / Vhe Gen Aja Phla- Fon
Phera
Adángbe Kpándo Agɔ́i Dogbó Alada Agbóme
Agu Kpelen Anéxɔ Hwe Ayizo Arohun
Kpesi Gliji Sikpí Gbesi Kpasɛ
Peci Stádó Kotafon Gun
/ Tado
Aveno Tɔwun Saxwe / Maxí
Tsáphɛ
Awlan Vɛ́ Se Wéme
Be Vhlin Tofin

[Texte]
 Dayin Vo Toli
Fodome Wací Xwela /
Phelá
Gbín Wancé Xwla /
Phla
Hó
(1) Classification of Gbe languages (according to Capo 1991)

According to the latter, who relied on two different wordlists and a phrase-list,
Gbe can be divided into three clusters, namely Western Gbe (Ewe and Gen),
Central Gbe (Aja), and Eastern Gbe (Fon and Phla Phera). Although a few varieties
have been either classified as part of one or another cluster, these divergences did
not impact the regularity of our reconstructions for Proto Gbe found in part 5.

Western Gbe Central Gbe (Aja) Eastern Gbe

WLC1 WLC2 WLC1 WLC2 WLC1 WLC2
Adan Adan Dogbo Dogbo Ajra Ajra E. P-
Ewe
Phera
Awlan Awlan Hwe- Hwe- Movol Toli
Aplaho Aplaho o
ué ué
Aveno Aveno Hwe- Hwe- Setno Tofin
Azovè Azovè
Be Be Hwe- Hwe- Tofin Movol
Gboto Gboto o
Togo Togo Hwe- Hwe- Toli Seto
Tohoun Tohoun
Vo Vo Sikpi Sikpi Easten Easter Fon
Gen
Xwla n Xwla
Waci Waci Gun Gun
Wance Wance Alada Alada
Wundi Wundi Gbeko Gbeko
n n
Agoi/G Agoi/G Arohu Arohu
liji liji n n
Anexo Anexo Fon Fon
Gen Gen Agbo Agbo
me me
Kpesi Kpesi Ci Ci
Kpelen Kpelen Maxi Maxi
Cluste
r-3
Gbin Gbin Weme Weme
Ho Ho Kpase Kpase
Vhlin Vhlin Ayizo Ayizo Weste
rn
Phla-
Phera
Gbokp Gbokp
a a
Kotafo Kotafo
n n
Gbesi Gbesi
Weste Weste
rn rn

[Texte]
 Xwla Xwla
Xwela Xwela
Saxwe Saxwe
Se Se
Daxe Daxe

(2) Classification of Gbe languages (according to Kluge 2000)

The innovations in Kluge (2000)’s classification somehow vindicate some historical

traditions of Gbe speaking people that were left contradicted by Capo (1991)’s
classification. For example, Fon people’s claim of their Alada / Ayizo origin that was
left unconfirmed by their languages’ respective classifications as Fon and Phla-Phera
by Capo has been counterbalanced by Kluge’s reclassification of Alada / Ayizo as Fon
; in the same vein, Gen’s tradition of a Fante-Ga origin of their ancestors having
learned their current language from their Waci neighbors has been confirmed by
Kluge’s classification of Waci within Gen, while Capo did classify Waci as Ewe and as
external to Gen. As Capo had concluded and in accordance to Kluge’s mention of the
close relationship between Waci and the rest of Ewe though, one could conclude that
Waci is indeed Ewe, and that Gen is also Ewe especially since (Proto) Gen language is
not a direct offshoot of Gen, but the result of contacts, mainly between Waci and
other Kwa languages. Similarly, one could conclude that Fon is an offshoot of Phla-
Phera, rather than a direct offshoot of Proto Gbe. Here, social anthropology, an extra-
linguistic discipline, helps us to confirm a purely linguistic classification.

3. Some aspects of Proto Gbe morphology

3.1. Nominal prefixes

One of the most underrated aspects of Gbe nominal morphology is the presence of a
à- ‘prefix’ in all Gbe dialects, along with another prefix. In most Gbe languages, this
second prefix has one or several optional realizations. However, in a restricted
number of languages such as Agbome (Fon), it does not realize phonetically at all (3).

Gloss Agbome Saxwe Hwe Agoi Awlan

(Fon) (Phla- (Aja) (Gen) (Ewe)
Phera)
‘night’ zá̰ ozá̰ ezá̰ ezá̰ ezá̰
‘rope’ ka̰ okà̰ ekà ekà eká
‘water’ sḭ ɛsı̰̀ esì esì etsì
(3) Examples of the realization of Gbe’s second prefix in five Gbe languages

[Texte]
 Typically, this second prefix is mandatory before a monosyllabic stem, but
optional before a disyllabic radical, as one can see in (4).

Glose Gen Gun Maxi

‘magic ēbōgbé ~ ōbōgbé ~ ōbōgbé ~
utterance’ bōgbé bōgbé bōgbé
‘man’ ēgbɛ̄ tɔ́ ~ gbɛ̄ tɔ́ ōgbɛ̄ tɔ́ ~ gbɛ̄ tɔ́ ēgbɛ̄ tɔ́ ~ gbɛ̄ tɔ́
(4) Examples of the optionality of Gbe’s second prefix when it precedes disyllabic
nouns in three Gbe languages.

The absence of phonological regularity between the various phonetic realizations

of the second prefixes excludes, in the present state of our knowledge, the
reconstruction of a single melody for this prefix in Proto Gbe. The à-prefix,
however, is regularly present in all Gbe languages and can thus be reconstructed
for Proto Gbe’s morphology. It can be reconstructed for several Proto Gbe
nominal items as well:

Gloss Proto Fon Phla- Aja Gen Ewe

Gbe Phera
‘tooth’ *à- à-ɖú à-ɖú à-ɖú à-ɖú à-ɖú
ɖú
‘egg’ *à-zı̰̀ à-zı̰̀ à-zı̰̀ à-zı̰̀, à-zı̰̀ à- à-zı̰̀
à-zı̰̀ ʒı̰̀
‘dog’ *à-vṵ́ à-vṵ́ à-vṵ́ à-vṵ́ à-vṵ́ à-vṵ́
(5) Reconstruction of some Proto Gbe nouns marked by the à- prefix

Both prefixes have been deemed ‘non-productive’ from a synchronic standpoint

(cf. Bôle-Richard 1983 for Gen, Brousseau & Lefebvre (2001) for Fon, Morley 2010
for Aja). However, in previous works (Capo Chichi 2011, Capo Chichi 2014), we
have proposed grammatical functions for both affixes in the modern Agbome
language, which belongs to the Fon branch. Names marked by both à- and the
other affix (Ø- in Agbome), display the syntactical characteristics of referential
nouns as defined by Grimshaw (1990). They can be pluralized (9,10), used with a
demonstrative determiner (7,8), with an article. In Grimshaw (1990), this class of
nouns is considered as different from the class of complex event nouns. The
former cannot be used along with temporal modifiers (11) and do not have an
argument structure (except for relational nouns).

Referential nouns Process (complex event) nouns

Do not have argument or aspectual Describe complex events: have
structure; apparent arguments are argument structure (assign theta
called ‘participants’. roles) and aspectual structure.

[Texte]
 Participants are not obligatorily Arguments are obligatorily
present in syntactic structure. present in syntactic structure.
May be deverbal (=‘derived’), but Always deverbal (derived from a
need not be (=‘simple’). verb stem).
Meanings: Meanings:
• Concrete objects Must be in the singular.
• Collective nouns
? not deverbal: ‘picture’
? deverbal: ‘invention’
• Simple events: ‘trip’, ‘mistake’,
‘lecture’

Many (including result and simple Agent may be expressed

event obliquely.
nouns) may be in the singular or
plural
Oblique agent expression is a Agent may be expressed
modifier (in obliquely.
English) or simply not possible (in
Polish)
An implicit argument cannot control An implicit argument can control
an an infinitive
infinitive in a purpose clause. in a purpose clause
Not compatible with adverbs of Compatible with adverbs of
completion or duration; compatible completion,
with frequency adverbs only in the duration, and frequency.
plural.

(6) Distinction between Referential and complex event nouns (adapted from
Rappaport 2001)

Similar characteristics are displayed in names marked by these other prefixes in

the other Gbe languages. It can be seen in the Towun language, which belongs to
the Ewe branch.

Agbome ā- hwà̰ élɔ́ kpɛ̰̀ mḭ̀

(Fon)
Towun ā- ʋà jɛ vɛ ̀
M
(Ewe)
Gloss RFR- War DEM to be 1SG
heavy /
[Texte]
 painful
Translation « This war is painful to me »
(7)

Towun e- ʧì jɛ vìví nà̰ Wò

(Ewe)
Agbome Ø- sı̰̀ élɔ́ vìví nṵ̀ yě
(Fon)
Gloss RFR- Water DEM to be To 3PL
pleasant
/ sweet
Translation « This water is sweet to them »(8)

Towun ā- ʋà jɛ vɛvɛ M
(Ewe)
Agbome ā- hwà̰ ɔ kpɛ̰̀ mḭ̀
(Fon)
Gloss RFR- war DEF to be 1SG
heavy /
painful
Translation « The war is painful to me »
(9)

Agbome à- klá(a)kú l nɔ̰̄ ɖì Ø- fɔ̰̄ l̰ ḭ̀

(Fon) é
Gloss RF Hippopota P HA
loo RF Lazine
R- mus L kB R- ss
lik
e
Translati "The hippopotamuses are lazy (lit. Look like laziness)" (10)
on

[Texte]
 Towun e- j̰ḭ wo zɔ̰ dɔdɔdɔ
(Ewe)
Gloss RFR- bovine PL walk slowly
Translation “Bovines walk slowly”

(11)

Agbo à- - xɔ̀ ɖō ʧɛ́ʤū vìví nṵ̄ mḭ̀

me kɔ́ kp
(Fon) ó
Gloss RFR Cr fo 1 minute (to be) to 1SG
- y r pleasant
of
th
e
co
ck

*The cry of the cock in one minute has been pleasant to me (12)

A relational noun, unlike other referential nouns, has an argument structure.

Referential nouns in English include ‘sister (of)’, or ‘mother (of)’. In Gbe, as one can
see in (13), the word for ‘mother’, for example, displays the same characteristics of
argument structure and of syntactic properties shared by other referential nouns :

Agbome kɔsi nɔ̰̀ ɔ̰̀ j̰ɔ̰́ ɖɛ̄ kpɛ̀

Gloss Kossi mother DEF To pretty
be
good

[Texte]
 Kossi’s mother is pretty (13)

Towun doʤi -si ɛ ʤɛ tugbe

(Ewe)
Gloss Dodzi wife DEF fall pretty
Dodzi’s wife is pretty (14)

Agbome nɔ̰̀ ɔ̰̀ j̰ɔ̰́ ɖɛ̄ kpɛ̀

(Fon)
Gloss Mère DEF être jolie
(15) bien
*The mother is pretty

Towun asi ɛ ʤɛ tugbe

(Ewe)
Gloss wife DEF fall pretty
*The spouse is pretty (16)

According to Grimshaw (1990), referential nouns can be divided into relational

nouns, result nouns, which refer to the result of an action, and simple event
nouns. The latter describe processes. However they possess syntactic properties
different from that of complex event nouns, but similar to that of other referential
nouns.

Further studies based on Grimshaw’s seminal work (Melloni 2007, Bisetto &
Melloni 2007) have showed that the semantic categories of result nouns derived
from verbs can be predicted from the semantic categories of the verbs. Hence,
creation verbs, when derived into result nouns yield the result of the action
expressed by the verb. However, non-creation verbs derived into result nouns
yield referential nouns with other semantic values (mean, target, entity in state,
etc.). These conclusions found for Romance languages seem to be confirmed for
Gbe as seen for Agbome (17) and Anɛxɔ (18) for example:

[Texte]
 Verb Agbome Creation Derived Noun
verb ? noun Semantic
category
‘To make ɖɔ̆ Yes àɖɔ́ ‘urine’ Result
(something)
flow’
‘To reject (a vĭ Yes àví ‘tears, Result
liquid)’ weeping’
‘to suck at’ nɔ̰̆ No ànɔ̰́ ‘nipple’ Target
(17) Deverbal nouns beginning with à- in Agbome (Fon)

Verb Anɛxɔ Creation Derived Noun

(Gen) verb ? noun Semantic
category
‘to suck nɔ̰́ No ànɔ́ Target
at’
‘To sting’ té No àté ‘insect Result
sting’
‘To voice, ʤì Yes aʤì ‘child, Result
to give progeny’
birth’
(18) Deverbal nouns beginning with à- in Anɛxɔ

The unifying feature of à- derived deverbal nouns in Proto Gbe seems to have been
the unpredictability of the exact meaning of the word. However, the derived nouns
would belong to a predictable semantic category in regard to the category of the
verbs they are derived from. This categorization is obviously different from that of
deverbal nouns marked by the second prefix.

Verb Agbome Derived Noun

(Fon) noun Semantic
category
‘to split’ klá̰ klá̰ ‘the Event
splitting’
‘To kpɛ̰́ kpɛ̰́ Event
cough’ ‘cough
(n.)’
‘To bring kplé kplé Event
together’ ‘bringing
together’
(19) Deverbal nouns beginning with the second prefix in Agbome

[Texte]
 Agbome Kɔ̄kpɔ̰́ hù aʤá xɔ́sú xwè kā̰ wēkōɛ̄ nɛ̰̀ ɖíè
(Fon)
Gloss Kokpon kill Adja king year Four ago

Verb Anɛxɔ Derived Noun

(Gen) noun Semantic
category
‘to die’ kú ekú Event
‘death’
‘To steal’ fḭ efḭ ‘theft’ Event
‘To zṵ ezṵ ‘insult’Event
insult’
(20) Deverbal nouns beginning with the second prefix in Anɛxɔ

Deverbal nouns marked by the second prefix often belong to the category of simple
event nouns. Their exact meaning is transparent and largely predictable in respect to
the meaning of the verbs they are derived from. What, would then be the distinction
between à- and the second prefix? In respect to the nominalization of verbs, they
would form referential nouns with a respectively loose and close link to the meaning
of the verb. The second would mostly produce simple event nouns. At this point of
the research, the choice of either prefix for non-deverbal referential nouns remains
undefined. One could possibly expect, as the distinction between both prefixes
suggests, that à- marked nouns would have a more restricted meaning from that of
nouns marked with the second, but this suggestion would definitely require further
investigation.

According to Grimshaw (1990)’s terminology, nouns are syntactically divided into the
referential nouns we have been dealing with, and complex event nouns. The latter
category can be distinguished from the former by its syntactic characteristics. Nouns
belonging to this category cannot be pluralized, used with a possessive determiner,
but can be used along with temporal modifiers and do inherit the argument structure
of the verb they are derived from. In all modern Gbe languages we are
knowledgeable about, a reduplication affix precedes the stem of the derived verb to
form nouns conform to these syntactic criteria. The realization of this reduplication
prefix in modern Gbe languages varies from a partial reduplication of the verbal stem
to its total reduplication. The languages illustrated below are Towun (Ewe) and
Agbome (Fon).

[Texte]
 Towun Agɔkoli wù ʋɛ gbɛtɔ́ wo alaʄaɛnɛ̰ adɛ
(Ewe)
Gloss Kokpon kill Ewe people PL Four ago
hundred
hundred
"Kokpon killed the Aja king four years ago" (20)
Agokoli killed several Ewe people 400 years ago (21)

Towun (Agɔkoli ʄɛ) eʋɛ ví wo wuwu alaʄaɛnɛ̰ adɛ

(Ewe)

Gloss (Aja By) Ewé People PL CE~kill Four ago

hundred

Agbome aʤá xɔ̄sú hū- (Kɔ̄kpɔ̰̄ sḭ̄ ) ʤɔ̀ xwè kā̰w

The murder of Ewe people (by Agokoli) 400 years ago (22)

[Texte]
 (Fon) hù
Gloss Aja king CE~kill Year by happen year
Fou
hun
"The murder of the Aja King (by Kokpon) happened four hundred years ago" (23)

In conclusion, one can confidently claim that Proto Gbe nouns were syntactically
divided into two types of nouns. Those two names would be the referential nouns
and the complex event nouns. The latter were marked with a reduplication prefix,
while the former were marked with either à- or with the second prefix. The second
prefix, when used to create deverbal nouns, would usually have yielded simple event
nouns. Those derived nouns would have had a meaning very close to that of the verb
they were derived from. On the other hand, when marked with à-, deverbal nouns
would usually have been result or simple nouns with an unpredictable meaning vis à
vis the verb they are derived from. The use of those prefixes marking referential
nouns is not productive in Gbe languages nowadays and there does not seem to be
much evidence to prove it has been in the past. However, the regularity of the use of
both prefixes to form nominalizations of verbs they are semantically closely and
loosely related to shows their significance in Proto Gbe morphology. It strongly
suggests, in our opinion, that their functions, while perhaps not productive, have
played a significant part in Proto-Gbe nominal morphology.

3.2. Nominal suffixes

Proto Gbe probably made use of a number of suffixes whose form and function can
be reconstructed on the basis of modern Gbe languages’ morphology.

-tɔ́ ‘agentive, provenance suffix’

Languages from Western and Eastern Gbe branches use a suffix -tɔ́ to express the
agent of a process expressed by a verb or of a noun phrase:

͂
Gen èze wɔ̀ tɔ́
Gloss Pot make AG

[Texte]
 ‘Potter’ (23)

Agbome atı̰́ kpà tɔ́

(Fon)
Gloss wood carve AG
‘carpenter’ (24)

Standard Ewe agbe tɔ́

Gloss Life AG
‘Human being’ (25)

Anɛxɔ (e)plà tɔ́

(Gen)
Gloss Locality of Grand-Popo, PR
Benin
‘Inhabitant of Grand-Popo’ (26)

Agbome xɔ̀gbónù tɔ́

(Fon)
Gloss Locality of Porto Novo, Benin PR
‘Inhabitant of Porto Novo’ (27)

Anɛxɔ

(Gen) ezé wɔ tɔ́

Gloss Pot make AG

‘Potter’ (28)

Agbome (Fon) atı̰́ kpa tɔ́

Gloss wood carve AG

‘carpenter’ (29)

Standard Ewe agbe tɔ́

[Texte]
Gloss life AG

‘Human being’ (30)

-nɔ̀, attributive suffix

Languages from Western, Central and Eastern Gbe nowadays use a -nɔ suffix after
noun phrases to express a person characterized by the noun phrase or who is the
possessor of the noun.

Standard Ewe vodú nɔ̰

Gloss Deity AT
‘Chief priest of a deity’ (31)

Agbome (Fon) tó kú nɔ̰

Gloss Ear Die AT
‘deaf person’ (32)

Aja egɛ̰ nɔ̰

Gloss beard AT
‘Possessor of a long beard’ (33)

-ví, diminutive suffix

Gbe languages from all its branches make productive use of a suffix –ví to form the
diminutive noun of the noun they are suffixed to.

Standard Ewe atí ví

Gloss Tree DIM
‘shrub’ (34)

Agbome (Fon) xó ví
Gloss Word DIM

‘Useless speech’ (35)

Aja afɔ̀ Ví
Gloss Foot DIM
‘big toe’ (36)

*tı̰́, *kṵ́ as noun classifiers ?

[Texte]
A number of modern Gbe nouns seem to end with reflexes of suffixes *-tı̰́,,*kṵ. While
the nouns ending with reflexes of -tı̰́, seem to denote somehow long-shaped things,
nouns ending with -kṵ́ seem to denote round-like things.

Agbome (Fon) *awɔ̰̀ awɔ̰nṵ̀ awɔ̰lı̰́ awɔ̰tı̰́

Gloss Muzzle’ ‘gesture of ‘nose’
disdain done
with the nose’
(37)

Gen *ŋu ŋukṵ́ ŋumɛ̰

Gloss ‘Eye’ ‘face’
(39)

Aja gá gaʧí
Gloss ‘Iron’ ‘spoon’
(38)

Standard Ewe vó vokú

Gloss ‘crotch ‘Testicles’
(39)

Awlan Ewe afɔ afɔtı̰́

Gloss ‘foot, leg, hind leg’ ‘leg, especially a
long one’
(40)
ta Movolo, Toli, Tofin, takṵ́
All non-Phla-Phera Phela, Tofin, Seto
Gbe languages,
Gloss ‘head’ Gloss ‘head’
(41)

Both suffixes resemble the Proto Gbe nouns *atı̰́ ‘tree’, *kṵ́ , seed and their reflexes
both phonetically and semantically. However, the absence of semantic references to
‘tree’ and ‘seed’ in these words seem to indicate they may have been the result of a
grammaticalization process from those words. The exact morphological and
syntactical function in Proto Gbe of these assumed suffixes remains unclear as those
affixes do not seem to be productive anymore in most modern Gbe languages.

-I, as a multifunctional floating suffix

Thanks to their productive use in some Gbe languages and their remnants in some
others, one can reconstruct at least two different –i suffixes for Proto Gbe. The first

[Texte]
one, which would not have been marked by a high tone, seems to have been
associated with non-compositional morphological creation.
͂ ͂ ͂ ͂
Agbome ahlı̰ ha̰ i ahlı̰ hɛ̰
(Fon)
Gloss Peacock NN-CMPS ‘dragonfly’

Agbome agbá̰ i agbɛ̰́

(Fon)
Gloss Plate, lid NN-CMPS ‘terracotta
container’
(42) Examples of –i suffix yielding non-compositional meaning in Agbome (Fon)

The second one was possibly marked by a high tone and characteristic of
compositional morphological creation.

Aja, Ewe kɔ ɖó Í kɔɖwí

Gloss Neck Hold CMPS ‘cushion’

Waci (Ewe) dé kú Í dekwí

Gloss Palm seed CMPS ‘palm nut’’

Agbome glɔ́ í glwé

(Fon)
Gloss Bottom CMPS ‘below’

(43) Examples of the –í suffix yielding compositional meaning in Gbe languages

The important functional diversity of both –I suffixes (change of grammatical

category, creation of instrument nouns, noun composition, etc.) could perhaps be
explained by its status as an epenthetic strategy in modern Gbe languages. Since
their function would not have been covered by proper morphemes, it would have
been covered by one default epenthetic strategy, that would have been –I, in order to
fill a template.

4. Proto Gbe Phonology

4.1. Nasality in Proto Gbe

[Texte]
As most of its daughter languages, Proto Gbe has previously been analyzed as a
language with no phonemic nasal consonants, but with nasal vowels.

t tʰ tˢ k kp
b [m] d dʰ dᶻ ɖ[n] g gb
f s x xʷ
v z l̰ j [j̰] w[w̰] ʁ ʁʷ Hʷ

i e ɛ a ɔ o u
ḭ ḛ ɛ̰ a̰ ɔ̰ o̰ ṵ

(44) Proto Gbe phonological inventory according to Capo (1991)

We however claim as we have argued for Agbome (Capo Chichi 2011), that there
were no phonemic nasal vowels in Proto Gbe either. The nasality trait found on the
so-called nasal vowels would be a supra-segmental trait spreading from the right
edge of the morpheme to the end of its domain to its left. The nasality trait can
realize as a trait found on nasalized vowels. But it also can realize as a syllabic nasal
consonant alternating on slow speech between [j̰ḭ] and [w̰ṵ]. The latter alternations
seem to have been the optimal realization of nasality. They however would only have
been realizable before an occlusive or an affricate consonant inside a disyllabic root.
The trait would have realized as the former in the other contexts. One argument for
this claim in Fon and Phla-Phera languages would be the formation of complex event
nouns with a partial reduplication prefix. In order to form those nouns, it copies the
initial consonant of the radical, but not its lexical vowel. It replaces its lexical vowel by
an epenthetic vowel like [i] or [u], yet copying the nasality trait of the stem on the
epenthetic vowel, as one can see in (45,46,47). In Gbe languages, there are roughly
two classes of consonants in regard to the spreading of nasality. The class A, that
would be nasalizable and allow the spreading of nasality from the right edge of the
word to its left, the class B, that would not.

b I b a̰
b i m a̰
b ḭ m a̰
m ḭ m a̰

[Texte]
(45) Scenario explaining the origin of nasality in the complex event noun mḭma̰ ‘to
share’ in Agbome (Fon) through a long range nasalization process from the right edge
straight to the left of the word

K I k a̰
k I k a̰
k ḭ k a̰

(46) Scenario justifying the impossibility of explaining the origin of nasality as a long
range nasalization process from the right edge straight to the left of the word

(47) Scenario explaining the origin of nasality in Gbe as a trait independent from
vowels through the example of the reduplication of verbs

In the case of the reduplication of a verbal stem whose initial consonant is nasalizable

k I N k a N
k I N k a̰
k ḭ K a̰
(45), one could argue that the nasalization of the epenthetic vowel would have been
due to the spreading of nasality from the vowel to the consonant and then to the
preceding epenthetic vowel. However, in the case of verbal stems beginning with
non-nasalizable consonants (46, 47), this scenario must be excluded. Indeed, the
nasalization process of the epenthetic vowel would have blocked by the non-
nasalizable initial consonant of the verbal stem. There would only be one remaining
scenario : in Fon languages, the partial copying of the stem would involve the initial
consonant and the trait of nasality, but not the vowel. It would imply that nasality is
independent from vowels as well as from consonants.

*i[ḭ] *e[ḛ,ɛ̰ ] *ɛ[ɛ̰ ] *a[a̰] *ɔ[ɔ̰] *o *u[ṵ]

[o̰,ɔ̰]

(48) The Proto Gbe phonological vowel set, according to our analysis

In Central and Western Gbe languages, there are no such partial reduplication
processes used to form complex event nouns. However, those languages use
different types of reduplication for this purpose. Since those processes are clearly

[Texte]
related to those of Eastern Gbe, there is actually no reason to believe that their
processes of formation are different from them and that nasality is also copied
separately from vowels.

Another argument would be the complementary distribution between the nasality

trait found on vowels and syllables j̰ı̰̀ and w̰ṵ̀ , realized in rapid speech as ŋ̀ or as
homorganic syllabic nasal consonants (49).

j̰ḭ and w̰ṵ Vocalic trait

found on
vowels
Found at the No Yes
right edge of
the noun
Found No No
adjacent to the
à- nominal
prefix
Found before Yes No
affricate and
stops in
nominal stems
Found before No Yes
sonorant
consonants
and fricatives
in nominal
stems
(49) Complementary distribution between ̰ jḭ~w̰ṵ~ŋ and the nasality trait found on
vowels

Research on linguistic universals (cf. Scheer 2004:707-744) has shown the tendency of
nasal consonants to be maintained before stops and affricates. They also experience
difficulties in being maintained before sonorants, fricatives and at the end of the
word. This tendency seems to be reflected by the behavior of j̰ḭ and w̰ṵ in Gbe which
does not realize in a context unfavorable to nasal consonants, and the vocalic trait
found in these other contexts.

From a morphological standpoint however, both (supra-)segments seem to share the

same restriction of not being allowed in the vicinity of the à- prefix.

At this point of our research, the question of the function of nasality in Gbe remains
unanswered. We have proposed (Capo Chichi 2011) that its nature would to have

[Texte]
been epenthetic in Fon. Indeed, in other areas of morphology, nasality in alternation
with partial or total reduplication is used to form different parts of speech such as
adverbs. Just like epenthetic vowels such as [i] and [u], nasality alternates between jḭ
and w̰ṵ according to absence or presence of labiality in the surrounding context. Like
other epenthetic strategies, they also are marked by a low tone. Some examples also
show clearer evidence of nasality as an epenthetic strategy to fill empty CV slots
requiring not to be filled.

Through inter-dialectal Gbe comparison, one can notice several alternations between
CuCV/ CiCV or CV1CV1 sequences and w̰ṵ̀ CV / j̰ı̰̀CV sequences.

Proto Gbe Agbome Anɛxɔ Awlan (Ewe)

(Fon) (Gen)
‘name’ *j̰ḭkɔ́ j̰ḭ(kɔ́) ŋkɔ́ ŋkɔ́~ j̰ḭkɔ́
’eye’ *nṵkṵ́ nṵkṵ́ ŋkṵ́ ŋkú
‘leaf ’ *ama+kpa ama̰(kpá) ama̰ aŋgbá~amakpá~amagbá
‘back’ *mḛgbé nɛ̰ gbé ngbé~ mɛ̰ gbé~
ŋṵgbé ngbé
‘male’ *nutsu sṵnṵ ŋusu, ŋutsu
musu
nsu
(50) Alternations between CuCV/ CiCV or CV1CV1 sequences and w̰ṵ̀CV / ̰jı̰̀ CV
sequences among modern Gbe languages

Here, through their status of ‘deletable’ vowels, u, i and a vowel preceding an

identical one in a single morpheme are deleted. Gbe’s phonology cannot license
consonants preceding other consonants (except in the case of consonant cluster with
increasing sonority). Hence, the consonant from the first syllable drops as well and
the whole CV sequence is replaced by either jì or w̰ṵ̀ as one would expect from an
epenthetic strategy used to fill empty CV slots. One could argue, that since all the
consonants preceding the deletable vowels in the examples in (50) are nasalized, one
could assume that they would not have been deleted and would have inherited the
low tone of the deleted vowel. They would thus have formed a syllabic nasal
consonant, later to be reinterpreted as an alternation of jı̰̀ or w̰ṵ̀. However, some
other Gbe alternations show more clearly that jı̰̀ and w̰ṵ̀ have indeed had an
epenthetic function.

a t á k ṵ́ + g w é
a t á w̰ ṵ́ g w é
(51) The alternation between the forms atakṵ́gwé~ataw̰ṵ́gwé in Agbome (Fon)

[Texte]
Here, /k/ is neither a nasalizable consonant nor a consonant allowed to bear a tone.
The evolution of kṵ́ into w̰ṵ́ can only be explained by its whole deletion followed by
the apparition of the epenthetic syllable [w̰ṵ].

Fon, Agbome, ɖ o kp ó
Kpase, Gun
Maxi w̰ ṵ kp ó
(52) The alternation between the forms ɖokpó~ŋkpó in Maxi and other Fon languages.

Here, the nasality of the first syllable in Maxi can also be explained by the
disappearance of */ɖo/ replaced by a [w̰ṵ] syllable used to fill its position it after its
deletion.

In conclusion, one could say that nasality in Proto Gbe was a trait independent from
both consonants and vowels. It was also likely to have been epenthetic in nature. Its
function, however, remains unclear at this point of our research.

4.2. Phonological changes from Proto Gbe to modern Gbe

Below are the evolutions of Proto Gbe Phonemes in modern Gbe languages. Are
not mentioned in those tables shifts caused by combination with other
morphemes.

Proto Gbe Ewe Gen Aja Phla-Phera Fon

*b (*m b (m before b (m before b (m before b (m before b (m before
before nasalized nasalized nasalized nasalized nasalized
nasalized vowel) vowel) vowel) vowel) vowel)
vowel)
*gb ([g] gb ([g] gb ([g] gb ([g] gb ([g] gb ([g]
before *[u]) before *[u]) before *[u]) before *[u]) before *[u]) before *[u])

*kp ([k] kp ([k] kp ([k] kp ([k] kp ([k] kp ([k]

before [u]) before [u]) before [u]) before [u]) before [u]) before [u])
*g (ʤ g (ʤ before g (ʤ before g (ʤ before g (ʤ before g (ʤ before
before non non non non non non
nasalized nasalized nasalized nasalized nasalized nasalized
[i]) [i]) [i]) [i]) [i]) [i])
*k (ʧ before k (ʧ before k (ʧ before k (ʧ before k (ʧ before k (ʧ before
[non [non [non [non [non [non
nasalized nasalized nasalized nasalized nasalized nasalized
[i]) [i]) [i]) [i]) [i]) [i])

*t ([*ʧ t ([ʧ] before t ([ʧ] before t ([ʧ] before t ([ʧ] before t ([ʧ before

[Texte]
 before no non non non non non-
nasalized nasalized nasalized nasalized i nasalized nasalized i])
[i]) [i]) [i]) and [u]) [i])

*d ([ʤ] d ([ʤ, dᶻ] ([ʤ] before

([ʤ] before ([ʤ] before ([ʤ] before
before non before non non non non non
nasalized [i] nasalized [i] nasalized [i]
nasalized [i] nasalized [i] nasalized [i]
and [u])
*ɖ (l as a ɖ (l as a ɖ (l as a ɖ (l as a ɖ (l as a ɖ (l as a
second second second second second second
consonant consonant consonant consonant consonant consonant
in rapid in rapid in rapid in rapid in rapid in rapid
speech) speech) speech) speech) speech) speech)
*tʰ t t t s s
*dᶻ d d d z z
*tʰ t t t s s
*dʰ d d d z z

(53) Phonological changes from Proto Gbe stops into modern Gbe

Proto Gbe Ewe Gen Aja Phla-Phera Fon

*v v v v v v
*f f f f f f
*x x x x x x
*h h h h h h
xʷ (*x xʷ (x before xʷ (x before xʷ (x before xʷ (x before xʷ (x before
before a, o, a, o, ɔ or u) a, o, ɔ or u) a, o, ɔ or u) a, o, ɔ or u) a, o, ɔ or u)
ɔ or u)
*hʷ (*h hʷ (h before hʷ (h before hʷ (h before hʷ (h before hʷ (h before
before a, o, a, o, ɔ or u) a, o, ɔ or u) a, o, ɔ or u) a, o, ɔ or u) a, o, ɔ or u)
ɔ or u)

*s s s ([s] or [ʃ] s (ʃ before s s

before i) [i] and [u])
*z z z z (ʒ before z z
[i])
(54) Phonological changes from Proto Gbe fricatives into modern Gbe languages

Proto Gbe Ewe Gen Aja Phla-Phera Fon

*e [ḛ, ɛ̰ ] merged into merged into merged e [ɛ̰ ] e [ɛ̰ ]
before a one single one single into one
nasalized phoneme in phoneme e single
vowel] each [ɛ̰ ] phoneme e

[Texte]
 *ɛ ([ɛ̰ ] modern [ɛ̰ ] ɛ [ɛ̰ ] ɛ [ɛ̰ ]
before a language:
nasalized ə [ə̰], ɛ[ɛ̰ ], or
vowel) e [ḛ]
*a ([a̰] a [a̰] a [a̰] a [a̰] a [a̰] a [a̰]
before a
nasalized
vowel)
*o ([o̰, ɔ̰] o [o̰] o [ɔ̰] o [ɔ̰] o [ɔ̰] o [ɔ̰]
before a
nasalized
vowel)
*i ([ḭ] before i [ḭ] i [ḭ] i [ḭ] i [ḭ] i [ḭ]
a nasalized
vowel)
*u ([ṵ] u [ṵ] u [ṵ] u [ṵ] u [ṵ] u [ṵ]
before a
nasalized
vowel)
*ɔ ([ɔ̰] ɔ [ɔ̰] ɔ [ɔ̰] ɔ [ɔ̰] ɔ [ɔ̰] ɔ [ɔ̰]
before a
nasalized
vowel)
(55) Phonological changes from Proto Gbe vowels into modern Gbe languages

Proto Gbe Ewe Gen Aja Phla-Phera Fon

*w ([w̰] w ([w̰] w ([w̰] w ([w̰] w ([w̰] w ([w̰]
before a before a before a before a before a before a
nasalized nasalized nasalized nasalized nasalized nasalized
vowel) vowel) vowel) vowel) vowel) vowel)
*j ([j̰] before j ([j̰] before j ([j̰] before j ([j̰] before j ([j̰] before j ([j̰] before
a nasalized a nasalized a nasalized a nasalized a nasalized a nasalized
consonant) consonant) consonant) consonant) consonant) consonant)
*l ([l̰ ] before l ([l̰ ] before l ([l̰ ] before l ([l̰ ] before l ([l̰ ] before l ([l̰ ] before
a nasalized a nasalized a nasalized a nasalized a nasalized a nasalized
vowel) vowel) vowel) vowel) vowel) vowel)
*ɖ ([n] ɖ ([n] ɖ ([n] ɖ ([n] ɖ ([n] ɖ ([n]
before a before a before a before a before a before a
nasalized nasalized nasalized nasalized nasalized nasalized
vowel) vowel) vowel) vowel) vowel) vowel)

(56) Phonological changes from Proto Gbe sonorants into modern Gbe languages

[Texte]
 4.3. Some Proto Gbe’s innovations from Pre Gbe

We call Pre Gbe the hypothetical synchrony preceding Proto Gbe. Although
evidence from it cannot be directly found from the reflexes of its daughter
languages, it can be inferred from linguistic universals, or from contact with other
languages. This is notably the case of the following consonants reconstructed for
Proto Gbe by Capo (1991). In our opinion, they can be traced back to the
following earlier phonemes:

**ɗ+**j *dᶻ ;

**ƭ+**j *tˢ

**tˢ *s

In our opinion and in the present state of our knowledge, the Proto Gbe consonant
system reconstructed by Capo (1991) does not require much revision. One necessary
revision would be the origin of the segments he reconstructed as *tˢ and *dᶻ.
According to Capo’s Proto Gbe’s phonological rules of evolution into its daughter
languages, one cannot explain some alternations such as in (59) :

Gloss Awlan (Ewe) Awlan (Ewe) Agbome (Fon)

‘Skin’ agbadᶻe agbalɛ̰ ‘leather’ agbaza
‘Crab’ agatˢa agala̰ agasa

(57) Alternations between affricates and nasalized liquids in Awlan (Ewe)

As one can see in (57), the reflex of the affricate before a vowel alternates with l̰ V̰,
sometimes with a clearly different meaning. These initial phonemes would have
evolved into an affricate plosive and a nasalized liquid. To explain these
evolutions, one can only argue that the initial forms for these reflexes would have
been to have been a nasalizable consonant with the potential of becoming an
affricate. This ‘affrication’ would likely have taken place before a palatal context.
The ideal candidate would be a non-explosive coronal consonant such as *ɗ or *ƭ,
that would realize as l̰ V̰ before a nasality trait and as ʧ or ʤ before a palatal
segment, such as /j/. Both segments, nasality and /jV/ in their respective contexts
would have had a morphological function, which would need further research to
be properly defined.

*ts *s

[Texte]
 The Gbe theonym sakpatá~sakpatɛ́ can be reconstructed with confidence for
Proto Gbe as *sakpatá. Both from a linguistic and extra-linguistic standpoint, it
seems clear that the theonym *sakpata is related to the Proto-Yoruba theonym
*t̪ à̰ kpà̰ ná̰ or to a later evolution such as *ʧà̰ kpà̰ ná̰ . Since the Yoruba languages
where Proto Yoruba *t̪ has evolved into a /s/ are marginal (Ohiri Aniche 1995)
have probably underwent this shift as the result of a very recent development, and
that the sakpatá form is widespread in Gbe, one can assume that the borrowing
occurred between *ʧà̰ kpà̰ ná̰ and the Proto Gbe (or Pre Gbe) form. However, as
exemplified by other borrowings from Proto Gbe to (Proto) Yoruba such as the
theonym *olit̪ a ‘Orisha’ that likely evolved into Proto Gbe *litsa, if Gbe had
borrowed both theonyms at the same time, it would have had the forms *litsa and
*tsakpata, not *litsa and *sakpata. This would suggest that the borrowing of
*tsakpata/*t̪ à̰ kpà̰ ná̰ Gbe had taken place before the *olit̪ a / litsa borrowing at a
time where Proto Gbe’s *s was an affricate we reconstruct as *ts. The existence of
such an earlier realization for Proto Gbe’s *s seems to be vindicated by the
existence of this initial affricate in Gbe in the hapax <tsakpata> found in
Westermann (1954), in Gbe herited words in Cuba such as ʧakwata or in Yoruba
toponyms such as ʧebelu rendered by Gbe speakers in a form now pronounced
[savalu] ‘city of Savalou, Republic of Bénin’.

*ƭ *t

The *tsakpata/*t̪ à̰ kpà̰ ná̰ borrowing also raises the issue of the status of the
equivalent of Proto Gbe’s *t in Pre Gbe. If, at the time of the borrowing, Pre Gbe
equivalent of *t was a non-nasalizable consonant, it could not have been nasalized
in Pre Gbe or evolved into */l̰ / *[n] in Pre or Proto Yoruba. The most likely
situation would have been that the predecessor of Proto Gbe *t had the
characteristics of a liquid consonant, that of a voiceless consonant, that would be
likely to have been *ƭ. The predecessor of *Proto Gbe’s *t similarity with a liquid
consonant seems to be confirmed by a few alternations between modern Gbe
languages such as Ewe kóló, Aja and Gen kótó ‘vagina’.

4.3. Syllabic structure

Capo (1991) claims the existence of phonological consonant clusters in Proto Gbe.
Hence, sequences phonetically alternating between CLV and CVLV are
systematically analyzed as /CLV/ by the author. Similarly, sequences alternating
between CJV and CiJV are analyzed as the realizations of /CJV/ phonological

[Texte]
 sequences. However, evidence from Western and Eastern languages shows it is
likely not to have been the case. First of all, several CLV seem to be derived from
CV verbs would have been suffixed a morpheme of the LV form, perhaps, among
other functions, in order to form verbs with an imaged meaning.

Gloss CV CLV
‘to spend the night’ dɔ̆ drɔ̆
/ ‘to wake’
‘to be discording, fà̰ flă̰
to sing wrong, to
chat, to sing (for
birds)’ / ‘To speak
vaguely of
something’
‘to walk, to fly’ / ‘to zɔ̰̀ zrɔ̰̆
jump’

(58) Examples illustrating the presence of a suffix –LV to form verbs from other verbs
in Agbome (Fon)

Gloss CV CLV
‘to disappear, be bu blu
lost, lose’ / ‘to
confuse, to be
confused’
‘to hide’/ to be ble
deceive, to be
untruthful
‘stand up, rise, fɔ́ flɔ́
arise, to rise, erect/
‘to reach a certain
height or size’
(59) Examples illustrating the presence of a suffix –LV to form verbs from other verbs
in Awlan (Ewe)

Moreover, in several modern Gbe languages, words cognate with words with a CLV
structure in the rest of Gbe are realized as [CRV] if C is coronal. However, if the
sequence realizes as disyllabic phonetically, CVLV never realizes as *CVRV. If the CLV
form was the underlying form for these sequences, one could hardly understand why
the R would have such a rare and isolated realization. Moreover, in Gbe languages,
since the epenthetic vowel can either be /i/ [u, i] or a vowel identical to the

[Texte]
adjacent vowel in the same morpheme, one could wonder why the appearing vowel
is always the vowel identical to the adjacent vowel to the stem and never /i/.

In modern Gbe languages, there is further evidence of an underlying CVLV structure

for some items other than verbs. In those cases, nouns with this CLV structure have a
Hence, in Ewe, the number ‘seven’ is formed through the composition of the numbers
‘six’ and ‘one’.

a n d e + ɖ e
a n d e l e
a n d r e
(60) Illustration of the formation of the numeral ‘seven’ through the composition of
‘six’ and ‘one’ in Awlan (Ewe)

ɖ ɛ + l e
d ɛ l e
d e l e
d r e
(61) Examples illustrating the formation of the phrase ‘the others’ through the
combination of ‘other’ and ‘PL’

4.4. Some aspects of Proto Gbe tonology

Several studies on Gbe tonology have showed the existence of two phonological
tones, high and non-high, inherently present on lexical vowels. These tones often
combined with other tones, notably tones derived from preceding consonants. Based
on the high tones found on vowels after voiced stops and sonorants as well as the
high tones found on vowels following unvoiced consonants in a few Gbe languages,
Gbeto (1995) reconstructed those three consonant derived tones along with lexical
tones for Proto Gbe. Combined together, they would have led to the creation of
several complex tones, such as falling, rising, both sometimes alternating with a
middle tone.

5. Proto Gbe wordlist

Glossing rules

• √ : root
• RFR- or ??-: Unknown noun prefix
• RFR1- : à- prefix
• RFR2- or ?- : second noun prefix
• -? : segment of undefined meaning

[Texte]
 1. ‘tooth’ /RFR1- /*a-ɖú/ [*aɖú]
√tooth/
2. ‘eat’ √to eat /*ɖu/ [*ɖu]
3. ‘suck’ /√suck (nipple)/ /*ɖɔ́N/ [*nɔ̰]
/CEN?-√to /*RED-gbɔ̰/ [*gbigbɔ̰]
breath?/
4. ‘breast’ /RFR1-(suck ( /*a-ɖɔ́́N/ [*anɔ̰́]
nipple))/
5. ‘Die’ /√die/ /*kú/ [*kú]
6. ‘Death’ /RFR2-die/ / *?kú/ [*kú]
7. ‘life’ /RFR-life?/ / *??gbɛ/ [*??gbɛ]
8. ‘Sun’ /RFR2-Sun?/ /*?hʷe/ [*?hʷe]
9. ‘moon’ /RFR2-moon- /*?hʷele(tı̰́)/ [*?hʷle(tı̰́)]
(long-
shaped)/
10. ‘Star’ /RFR2-moon- /*?hʷele(tı̰́)vi/ [*?hʷle(tı̰́)vi]
(long-
shaped)-DIM/
11. ‘head’ /RFR2- /*?ta/ [*?ta]
√head/
12. ‘water’ /RFR2- /*?tˢḭN/ ‘water’ [*?tˢḭ]
√water, milk/
13. ‘thigh’ /RFR1-to /*a-tʰa/ [*atʰa]
crawl/
14. ‘crawl’ /√to crawl/ /*tʰa/ [*tʰa]
15. Belly /RFR-√belly?/ /*??dɔ/ [*??dɔ]
/RFR2-√belly/ /*?xʷo?/~ [*?xo]
/?xʷa?/
16. Foot /RFR1-√leg, /*a-fɔ/ ‘foot, [*afɔ]
foot, hindleg/ leg, hindleg’
17. Hand RFR1-√hand, /*a-lɔ/ ~ /*a-sí/ [*alɔ] ~ [*asi]
handle?/~
/RFR-√hand,
finger?/
18. Body /RFR-√body, /*?-gbaɗa-ja/> [*?gbadᶻa]
skin+?/ /*?-gbadᶻa/
‘Body, skin’
19. Skin [RFR2-body /?-wuN-tó-I/? [*?ntwé]
(face?, /?-ɖuN-tó-I/?
head?)-?I
20. Mouth /RFR2- /?-ɖúN/ [*?nṵ́]

[Texte]
 √mouth,
opening?/
21. Nose /RFR-√nose- /??-wɔ̀N-tiN/ [*?wɔ̰tḭ]
long shaped/
22. Eyes [RFR2-body /*?-ɖuN-kúN/ [*?nṵkṵ́]
(face?, /*?-wuN-kúN/ [*?ŋ̀kṵ́ ]
head?)-
round-
shapedI
23. Ear /RFR2-√ear/ /*-tó/ [*?tó]
24. Tongue /RFR- /*?-ɖɛ́/ [*?ɖɛ́]
√tongue?/
25. Neck /RFR2-√neck/ /*?-kɔ́/ [*?kɔ́]
26. Sky /RFR-√above, /* ?-djí/ [*?ʤí]
sky/
27. Human /RFR-√Life- /*??-gbɛ-tɔ́/ [*??gbɛtɔ́]
AG/~ /*??-bɛN/ [*??mɛ̰ ]
/RFR-
√person/
28. Male /RFR1-√male, /*a-tˢú/ [atˢú]
husband?/
29. Female /RFR1- /*a-sì/ [*asì]
√female, wife,
priest/
30. Man RFR-(?)- /**?-ɖúN-tˢu/ [*?ntˢu]
√male, man /**?-wúN-tˢu/
/**?-ɖúNtˢu/
/**?-wúNtˢu/

31. Woman /RFR2- /*-jɔ́(-)ɖuN/ [*?j̰ɔ́nṵ]

√woman(-)?/

32. One /NUM-√one-?/ /?-ɖe-ká/ [?ɖeká]

/NUM-√one/ /?-ɖo(-)kpó/ [?ɖokpó́]
33. Two /NUM-√two/ (*be? ) [*?ve, ?*we]
/*?ve,*?we/
34. Three /NUM- /?-tɔN/ [*?tɔ̰]
√three/
35. Four /NUM-√four/ /?-ɖɛN/ [*?nɛ̰]
36. Five /NUM-√five/ /a-tɔ̀N/ [*atɔ̰́]
37. Six /NUM-√six/ /a-dʰɛN/ [*andʰɛ]

[Texte]
 38. Flower /RFR1- /*?-sɛ́/ [*?sɛ́]
√flower/
39. Seed /RFR-√seed?/ /?-kúN/ [*?kṵ́ ]
40. Meat, animal /RFR-√meat, /*?-laN/ [*?l̰ a̰]
animal’/
41. Bird /RFR2-√bird, /?-xɛ/ [*?xɛ]
flying animal/
42. Snake /RFR2- /*?-daN/ [*?da̰]
√snake/
43. Dog /RFR1- /*a-vúN/ [*avṵ́ ]
√dog?/
44. Tree /RFR1-√long /*a-tíN/ [*atı̰́]
shaped
object/
45. Lake /RFR2-√lake/ /*?-to/ [*?to]
46. Rope /RFR2-√rope/ /*?-kaN/ [*?ka̰]
47. Blood /RFR2- /*?-hʷuN/ [*?hṵ]
√blood/
48. Bone /RFR2- /*?-xʷú/ [*?xú]
√bone/
49. Egg /RFR1-√egg? /a-zíN/ [*azı̰́]
/
50. Tail /RFR2-√tail/ /a-sí/ [*así]
51. Feather /RFR2- /*?-fuN/ [*?fṵ]
√feather/
52. Hair /RFR2-√hair /*?-ɖa/ [*?ɖa]
of the head/
53. Father /RFR-√father/ /*?-tɔ́/ [*?tɔ́]
54. Mother /RFR- /*?-ɖɔN/ [*?nɔ̰]
√mother/
55. Child / RFR2- /*?-ví/ [*?ví]
√child/
56. To Cry /VRB + tears/ /?-aví/ *Verb + [*aví]
57. Tears /RFR-√tears/ /*a-ví/ [*aví]
58. Leaf /RFR1-√leaf/ /*abaN/ [*ama̰]
59. Oil /RFR1-√oil, /*abiN/ [*amḭ]
grease, fat/
60. Excrements /RFR2- /*?bíN/ [*?mı̰́]
√excrements/
61. Urine /RFR1-(CEN)- /*a(RED)ɖɔ́/ [*a(ɖi)ɖɔ́]
to urinate/

62. Nail /RFR2-√nail/ /?fɛN/ [*?fɛ̰]

[Texte]
 63. Large /√large-ADV/ /gbaʤa-RED/ [*gbaʤaa]
64. High /to be high/ /*gaN/ [*ga̰]
65. Thick /to be thick/ /*tilí/ [*tlí]
66. The /DEF/ /*ɔ~a/ [*(l)ɔ~(l)a]
67. I /1SG/ /*N/ [*w̰ṵ~j̰ḭ]
68. You /2SG/ /*wo?/ [*wo?]
/*o?/ [*o?]
69. He /3SG/ /*é/ [*é]
70. We /1PL/ /*?bíN/ [*mı̰́]
71. You /2PL/ /*?biN/ [*mḭ]
72. They /3PL/ /*wó?/ [*wó?]
/*ó?/

73. Breath /to breath/ /*gbɔN/ [*gbɔ̰]

74. Spit /to spit/ /*tuN/ [*tṵ]
75. To cough /to cough/ /*kpɛN/ [*kpɛ̰]
/*kpeN/ [*kpḛ]

76. Cough (n.) /RFR-to /*??kpɛ̰ / [*??-kpɛ̰]

cough/ /*??kpḛ/ [*??-kpḛ]
77. Name /RFR2- /*?jiN(-)kó/ [*?j̰ḭkó]
√name/
/RFR2-
√name- ?/
78. House /RFR-√house/ /*??-xʷe/ [*??-xʷe]
79. Year /RFR2-√year, /*?-xʷé/ [*?xʷé]
festival/
80. Day /RFR2-√day/ /*?-gbe/ [*?gbe]
81. White /√to be /*wé/ [*wé]
white/
82. Black /√very black- /*kiɖi-RED/ [*ʧriʧri]
ADV/
83. Dry /to be(come) /*xʷú/ [*xú]
dry/
84. Bovine /RFR2-√big /*?jiN/ [*j̰ḭ]
terrestrial
animal/
85. Listen /to listen/ /*se/ [*se]
86. Puke /to puke/ /*tʰru/ [*tʰru]
87. (Be) cold /to be cold/ /*fa/ [*fa]
88. Hot /√fire-ADJ/ /*dᶻo-RED/ [*dᶻodᶻo]
89. Stone /RFR2- /*?kpɛ́N/ [*?kpɛ̰́]

[Texte]
 √stone/ [*?kpḛ́]

90. Fire /RFR2-√fire/ /*?dᶻo/ [*?dᶻo]

91. Night /RFR2- /?záN/ [*?zá̰ ]
√night/
92. Milk /breast- /??ɖɔ́N+?tˢiN/ [*??nɔ̰tˢḭ]
water/
93. Noun plural /-PL/ /*-wó/? /*-ó/? [*-wó?]

94. Wing /RFR1-√wing/ /*a-wa/ [*awa]

95. Heavy /to be heavy/ /*kpɛN/ [*kpɛ̰]
96. Drink /to drink/ /*ɖuN/ [*nṵ]
/to drink+?/ /*ɖɔN+I?/ [*nɔ̰i~*nwɛ̰?]

97. Knee /RFR2-√knee/ (**kʷali>)/*?koli/ [*koli]

/*?kpoli/
98. to laugh /to laugh/ /*ko/ [*ko]
99. God /NEG-to /*ma̰-hʷu/ [*ma̰hʷu]
surpass/
100. Urinate /√to make /ɖɔ/ [*ɖɔ]
something
flow/

Bibliography

• Abaglo, P. & Archangeli D. (1989). Language-particular underspecification:

Gengbe /e/ and Yoruba /i/, Linguistic inquiry 20.
• Bisetto A. & Melloni Ch.(2007), "Result Nominals: a Lexical-Semantic
Investigation".In On-line proceedings of the Fith Mediterranean Morphology
Meeting (MMM5) Fréjus 15-18 September 2005, Booij G., B. Fradin, E.
Guevara, S. Scalise & A. Ralli (eds).
•

Bôle-Richard R. (1983), Systématique phonologique et grammaticale d'un parler ewe:

le gen-mina du Sud-Togo et Sud-Bénin, Paris : Éditions l'Harmattan.
• Brousseau A.M. & Lefebvre C. (2002), A Grammar of Fongbe. Berlin : Walter
de Gruyter.
• Capo, H.B.C. (1990) : Systèmes numériques et hétérogénéité ethnique des
communautés de parlers gbe
• Capo, H.B.C. (1991), A Comparative phonology of Gbe, 14. Berlin/New York:
Foris Publications & Garome, Bénin: Labo Gbe (Int).
• Capo Chichi S. (2011), La nasalité en fon à l'interface morphologie /
phonologie, Mémoire de Master 2, Université Paris VII-Paris Diderot.

[Texte]
 • Capo Chichi S. (2012), On the origin of a Yoruba deity and its consequences
for the History of the Bight of Benin, Communication made at the Séminaire
des Doctorants of the LLACAN 2011-2012, Paris, 7/12/2012
• Capo Chichi S. (2014), Essai d’explication sur la fonction de
quelques affixes nominaux en fon, Communication made at
the Journée des Doctorants du LLF, Paris, 26/06/2014
• Clements, G.N., Osu S. (2003), Ikwere nasal harmony in typological
perspective, in Patrick Sauzet & Anne Zribi-Hertz (eds.), Typologie des
langues d'Afrique et universaux de la grammaire. Vol. II. Paris: L'Harmattan.
• Stewart, J.M. (1995). Implosives, homorganic nasals and nasalized vowels in
Volta-Congo. In Emenanjo, E. N. & O.-m. Ndimele (eds), Issues in African
languages and linguistics: essays in honour of Kay Williamson, Aba
(Nigeria) : National Institute for Nigerian Languages.
• Scheer T. (2004), A lateral theory of phonology. Vol 1: What is CVCV, and
why should it be ? Mouton de Gruyter, Berlin G. N. & Rialland A. (2008),
Africa as a phonological area, in Bernd Heine & Derek Nurse (eds), A
Linguistic Geography of Africa. Cambridge: Cambridge University Press.
• Gbeto F. (1995), Quelques aspects comparatifs et diachroniques de la
tonologie du Gbe, Afrika und Übersee 78, pp. 73-100.
• Gbeto F. (1997), Le Maxi du Centre-Bénin et du Centre-Togo, Cologne,
Rudige Köppe Verlag.
• Gbeto F. (2000), Les emprunts linguistiques d’origine européenne en Fon
(Nouveau Kwa, Gbe , Bénin), Cologne, Rudige Köppe Verlag.
• Gordon R., Grimes B., (eds.), 2005, Ethnologue: Languages of the world,
Dallas (Texas) :SIL International.
• Grimshaw J. (1990), Argument Structure,. Cambridge, Mass.: MIT Press.
• Kluge A. (2000), The Gbe language varieties of West Africa: a quantitative
analysis of lexical and grammatical features, Submitted in partial fulfillment of
the degree of Master of Arts in Language and Communication Research,
School of English, Communication and Philosophy, University of Wales,
College of Cardiff.
• Melloni Ch. (2006), Logical Polysemy in Word Formation: E and R Suffixes,
Lingue & Linguaggio 2.
• Morley, E.A. (2010), A grammar of Ajagbe , München : LINCOM Europa.
• Rappaport G.C. (2001), The Geometry of the Polish Nominal Phrase:
Problems, Progress, and Prospects, in Banski P. and Przepiórkowski A.
(eds),Generative Linguistics in Poland: Syntax and Morphosyntax, Warsaw:
Polish Academy of Sciences, pp.173-89.
• Rassinoux J. & Segurola B. (2000), Dictionnaire fon-français. Madrid :
Sociedad de Misiones Africanas.
• Tchitchi, T.Y. (1981), Eléments de lexique thématique aja-français, [Cotonou,
Benin] : Ets Z. Akinocho.
• Tchitchi, T. Y. (1984), Systématique de l'Ajagbe, Thèse de Doctorat de troisième
cycle, Etudes africaines, Université Paris 3.
• Westermann, D.H. (1907), Grammatik der Ewe-Sprache, Berlin : E. Vohsen.
• Westermann, D.H. (1954). Wörterbuch der Ewe-Sprache. Berlin : Akademie-
Verlag.

[Texte]
 • Westermann D.H. (1973), Evefiala or Ewe-English dictionary : Gbesela Yeye or
English-Ewe-dictionary, Nendeln : Kraus reprint.

[Texte]