1. Introduction
The Eυe language, which is the focus of this study, is spoken in Togo, Benin and Ghana. The Eυe language has a written form known as Standard Eυe (SE) based on three main variants; Aŋlɔ, Tɔŋu and Eʋeme in Ghana. Duthie (1996) indicates that the Eυe language exhibits distinctive varieties even in very close communities. These dialect differences occur in all aspects of the linguistic features of the language: lexis, syntax and phonology. Though the variations seem imperceptible to non-native speakers, they are noted across borders, and also at community levels (Duthie, 1996; Aziaku, 2016).
Trudgill (2002) outlines factors that may trigger spatial diffusion of the various linguistic features in a language as social class and geographical mobility, and also the strength of network ties. The variations in Eυe language that account for its enormous dialects principally emanate from the spread of social and geographical factors (see Aziaku, 2016). It is notable that the three main Eυe dialect groups in Ghana vary in phonology, syntax, semantics and lexis.
Atakpa (1997) and Ansre (2000) identify fourteen Eυe vowels; seven oral and seven nasal phonemes. Ansre (2000: p. 31) mentions that “an eighth pair is attested” as illustrated as follows:
There are consonantal as well as vowel variations among the various speakers in some contexts. For example, the velar stops, /g, k/ and the alveolar fricatives, /s, z/ as well as /i, e, u, ε, etc./ differ among Aŋlɔ and Tɔŋu (see Ansre, 2000 and Aziaku, 2016). The variations can be seen in words like tegli, tekli, tekle “patridge”, zakple, sakple “chamber pot”, aɖuba, aɖiba, “pawpaw”, asike, asikɛ, “tail”, and akpatogui,akpatogoe “salted tilapia”, etc. Structurally, the dialect of the Northern Volta is slightly different from that of the southern part (Ansre, 2000; Ameka, 2008; Aziaku, 2016). This difference is mostly seen in the use of the habitual marker -na, the definite articles la/-a and the progressive markers, “le…m”, etc. The differences are represented as follows:
![]()
Semantically, dialects largely employ different lexical items, and in some cases similar ones with different denotations. A typical example is aɖa in Eυeme and fofoŋ “crystal cane” or “chewing cane” in Aŋlɔ (Ansre, 2000
2. Objective
The paper examines the habitual marker -na in Tɔŋu dialect of Eυe. More specifically, to establish that -na has different forms in the dialect, and that its form extends beyond a prolongation of the last vowels of lexical verbs in Eυe.
3. Methodology
This research is a descriptive analysis of the realisation of the habitual marker -na in Tɔŋu. The data were taken from 500 participants from predominantly Tɔŋu speaking communities such as Mafi Kumase, Mafi Adidome, Sogakope, Dabala, Volo and Mepe. The speech sample was recorded using a Hnady 4 Next audio recorder in various social gatherings such as markets, churches, funerals including homes. The researchers spent at least one week in each community to interact with the people during the data collection period. During this period, we sought their consent through their chiefs and elders.
4. Forms of the Eʋe Habitual Marker -na
We describe the various forms of the habitual marker in the Standard Eυe (SE), in this section, to lay a foundation for the discussion. Comrie (1976) explains that habitual markers are used to describe non-incidental situations which are characteristic of extended period of time. According to Osam (2008: p. 80), “Events marked by the habitual have the feature of being a habit or customary”. The -na in Eυe equally expresses habitual or customary action. As a morpheme, -na is normally enclicised to intransitive verbs. In Standard Eυe, it has a truncated form, -a, whichis obligatorily used in transitive constructions (Banini, 1963; Atakpa, 1993; Amegashie 2004; Aziaku, 2012). According to Ameka (2008: p. 137), “A morphological defining feature of verbs in Eʋe is that they can take a habitual suffix -na/-a [and] a prescriptive rule about the alternation between -na and -a is that the full form -na occurs if there is no complement following the verb and -a is used if there is a complement…”. It is also noted that -na exists as -nε in the SE, which is normally triggered by the use of a pronominal object in a habitual construction (Aziaku, 2012). For example, -na becomes nae(na + 3rdperson singular pronoun e), that is -nε.
Westermann (1930), Banini (1963), Atakpa (1993), Amegashie (2004) and Aziaku (2012) explicate the grammatical rules that govern the use of -na and its counterparts –a and -nε. These scholars variously confirm that in transitive constructions, a truncated form of -na which is -a used while the full -na is employed in intransitive constructions. The third category, -nε manifests only when a pronominal object “e” is enclicised to the -na. For example, if we have “Kofi kpɔa Ama”, and the object “Ama” is changed to the pronoun “e”, then -na must be used in its full form even though the verb “kpɔ” bears a complement. In this case, we will have “Kofi kpɔna e”. Finally, it happens that the -a of the -na that supposedly should join the pronoun “e” rather coalesces to become ε. Thus Kofi kpɔna e becomes Kofi kpɔnε.
Examples 1, 2 and 3 illustrate the customary use of the -na/-a/-nε.
Additionally, -na can be enclicised to lexical verbs to mark progression. This occurrence is plausible with progressive verbs such as “yi” and “gbɔ”. For example, the use of -na with progressive verbs such as “yi” and “gbɔ” in the constructions “Egbɔna (he/she is coming) and “Eyina” (he/she is going) show progression. Ameka (2008: p. 138) interprets it as “motion in progress now”. Aziaku (2012: p. 4) points out that this particular function of -na has not received much attention from scholars because they restricted their researches to “the phrase le …m” as though it was the only form that might indicate progressive action in Eυe. Interestingly, this usage of -na can also be interpreted as a customary action. That is, “Kofi gbɔna” (Kofi is coming/Kofi always comes) or “Kofi yina” (Kofi is going/Kofi always goes). The point here is that -na with the progressive verbs “yi” and “gbɔ” does not only mark progression, but also expresses habitual actions. It is important we note that the verbs used in this form are usually yiandgbɔ and what follows them is always an adverb. This is confirmed by Ameka (2008) where he stated that there is cross-linguistically an affinity between the progressive and the habitual.
Ameka employs examples 4 and 5 to illustrate this function of the habitual marker -na. The word nugbe “journey” in example 4 below is an adverb of place, and the verb particle -ɖáin example 5 also indicates distance.
5. Typology of the Habitual Marker -na in Tɔŋu
The discussion in this section illustrates the typology of -na in Tɔŋu. In the discussion, we juxtapose the forms of -na in the Standard Eυe (SE) with the Tɔŋu variants. The study has shown that -na manifests itself in different forms in Tɔŋu dialect depending on the kind of vowel that ends a lexical verb in Eʋe. This means that its form is determined by the final sound of the verb it is attached to or follows. Let’s look at an example below.
In this example, we have “Eɖaa makani” (Eɖa+a makani), which coincidentally presents the same structure of the SE. In the first place, “-na” has been truncated to “a”, making “ɖa” plus “-a” to become “ɖaa”. Another observation being made here is that the vowels of “ɖa” and -na, have the same height. Similar examples are: “Ebaa ame” (S/He cheats always), “Edzraa nu” (S/He sells always), etc. Notice that in Tɔŋu it does not matter whether the verb is transitive or intransitive, the form remains the same. Let’s take a look at the following intransitive constructions as examples, “Evaa” (S/He comes always), “Ekpaa le ga eve me” (S/He closes always at 2 o’clock), etc., which in SE would have been “Evana”, “Ekpana le ga eve me”. In example 7, however, we have:
Here we have “Eʄu + wɔ”. This in the SE would have been “Eʄua du” (Eʄu + a + du). We can see several things happening here. In the first place, the -na is similarly truncated to -a and n is deleted. But instead of lengthening a, we see a different consonant w inserted between the last sound u of the verb “ʄu” and the habitual marker -a. The “-a” subsequently changed to “ɔ”. This time, it is no longer lengthening of -a as we saw in example 6, but there is a transformation in the -a. That is, the “-a” of “-na” which is unrounded has changed to rounded “ɔ” after w. The change of -a to -ɔ can be said to be the influence of the environment of the labial-velar /w/. It is thus obvious that the presence of /w/ has learned some labial feature to “-a”, changing it to collocate with /w/. Thus, there is a deletion, insertion, labialisation, raising and rounding: n is deleted, w-is inserted, a is raised and rounded to become /ɔ/. “Eʄuna” therefore becomes “Eʄuwɔ”. Similar examples of this form are: “Eluwɔ ɖa”, (S/He shaves hair), “Ekuwɔ tsi” (S/He fetches water), etc. The -a is however maintained if the following word begins with a vowel. For instance, “Eɖuwaati” (S/He chews stick), “Eɖuwaakplẽ” (S/He eats akplẽ), etc.
In example 8, the -na changes to /e/ a different vowel altogether. We have for example:
In “Edziye nake”, we have “Edzi” plus “ye” plus nake, which in the SE would have been “Edzea nake”. In the example 8, “-na” is truncated to -a, “n” is deleted while “y” is inserted. The “-a” subsequently changed to -e. The /e/ of “dze” also changed to /i/. We can say that the change of -a to -e and of /e/ to /i/ are due to the presence of the palatal /j/. That is, since /j/ has a high feature, it has influenced the /a/ (a low vowel) and /e/ (a mid-high vowel) to be raised. Thus, there is a simultaneous raising of the preceding and following vowels to the height of /j/. The -na becomes -ye; /n/ becomes “y”, “a” becomes “e” while /e/ of “dze” becomes /i/. Thus, we have “dzi” instead of “dze”. Note that “i” + “y” are both palatals. It is possible to say that both the /a/ and /e/ have been somehow palatilised. We have other examples such as: Emiye nu (He/she roasts thing), Eυliye nu, (He/she litigates), Efiye fefe, (He/she plays), etc. But note “Eυliy(e)a anyigba” (He/she litigates over land), “Esiy(e) a ame” (He/she fears a person). In these examples, the -a remains because of the influence of the following /a/ of “anyigba” and of “ame”; thus we have “Eυlia anyigba” (He/she litigates over land), “Esiaa ame” (He/she fears a person). Another example to talk about is:
Example 9 is rather interesting. We have “Ewɔ” plus “ɔ” plus “dɔ”, “He works”. In SE, this would have been “Ewɔadɔ”. The -na is similarly truncated to -a, while “n” is deleted. The “-a” is subsequently raised and pushed back to collocate with “ɔ” of the verb “wɔ”. In this example, we can say that “-a” is raised, backed and rounded to collocate with the vowel /ɔ/ and /w/ which are naturally rounded. Other examples in Tɔŋu are: “Elɔɔ nu” (He/she collects things), “Efɔɔ nu” (He/she finds things), “Ekɔɔ nu” (He/she lifts things), etc. Another instance of the use of -na in Tɔŋu is illustrated below.
In example 10, the “-a” similarly changes to “ɔ” just like in example 8. However, the final /o/ of the verb “ʄo” is either maintained or changed to “u”. Thus, in “Eʄoa bɔlu”, we will have “Eʄowɔ bɔlu” or “Eʄuwɔ” bɔlu. Some other examples in the dialect are: “Etowɔ fufu/etuwɔ fufu” (He/she pounds fufu), “Ehuwɔ te” (he/she uproots yam), etc.
In a situation where -na combines with a pronominal object as discussed earlier, it changes to -nε. This is illustrated as follows:
The use of -na in its full form as exemplified in examples 11a and 12a are unacceptable in SE, except the truncated form “-a” in examples 11b and 12b. Nonetheless, the full form -na becomes necessary when the object is pronominal as in examples 11c and 12c respectively. The objects “Ama” and “de” are represented by the pronoun “e”. In the SE the habitual marker -a and the pronoun “e” coalesced into ε in examples 11d and 12d. In Tɔŋu, however, the truncated -a first assimilates to “-e” before the pronoun “e” in example 11e. Example 12e is however unacceptable in Tɔŋu. In 12f, -a assimilates to -e before the pronoun “e” and subsequently coalesced with the final sound /a/ of the verb to become -ε. Note that in both cases there is some sort of diminitisation.
In the first instance, the habitual marker -a assimilates to -e before the pronoun “e” in a habitual construction. In the second instance, the -a assimilates to -e and then finally coalesced with the final sound “a” of the verb “kpa” to -ε before the pronominal object “e”. Thus, “Kofi kpɔa Ama” and “Kofi kpaa de” become “Kofi kpɔee” and “Kofi kpεe” respectively.
This is summarised in Table 1 below.
![]()
Table 1. Lexical verbs and habitual forms in Tɔŋu.
From the discussion so far, we can say that marking habitual in Tɔŋu differs from the Standard Eυe. The study has shown that the full form of -na does not exist in the Tɔŋu dialect. -na is always truncated to -a in Tɔŋu. Nonetheless, the truncated form has different pronunciations as follows:
For instance, we noticed that -na is always truncated to -a in Tɔŋu whether in a transitive construction or in an intransitive construction. Interestingly, -a is realised as -a only when the final sound of the verb that -a is affixed to is “a” or when the next word after the verb begins with “a”. Example 8 occurs as a result of the verb final “a” of the verb. This indicates that in Tɔŋu the final “a” of a verb is always lengthened to determine a customary action. Phenomenally, the final “o” of verbs sometimes raises to /u/, and then -a changes to low back vowel /ɔ/ with the insertion of /w/ between the “u/o” and “-a” in habitual marking as demonstrated in example 7. Example 7 is thus an instance of -a changing to “ɔ” to collocate with the rounding of /w/ and “u” to mark a customary action. There is also a total change of “-a” to “-e” when the final vowel of a verb is “e”. The final vowel “e” of the verb is normally first replaced with “i” before taking the habitual marker -e. However, we noticed that in this instance, there is always an insertion of /j/ whose height is similar to that of /i/. Similarly, i-stem verbs collocate with -e as their habitual markers. This is shown in example 8. It has been noticed that the variants of the habitual marker -a are not triggered by the consonants of the verbs but their final vowels. What accounted for the replacement and positional change in the verbs ending in -e and -i are the insertion of /j/.
6. Conclusion
The paper examines how the habitual marker -na is realised in Tɔŋu dialect. The study draws a conclusion that the full form of the habitual marker -na does not exist in Tɔŋu dialect, except its truncated form. But the truncated form “-a” exhibits itself in various forms depending on the verb-stem-final sound, a submission contrary to the assertion that marking of customary actions in Tɔŋu is merely a prolongation of the last vowel of a lexical verb. These findings, are nevertheless, at variant with the recognised realisations in the SE, which are fundamentally morphological (Ameka, 2008; Aziaku, 2012). Indeed, this is in confirmation of the observation made by Westermann (1907) as far back as 1907 that the habitual marker varies greatly in the Eυe dialects.
In Standard Eυe, as explained earlier, the full form -na occurs if there is no object in the construction, but not in Tɔŋu. The presence of an object, however, necessitates the use of the non-full form -a in the SE as well as in Tɔŋu. We also noted earlier that a third category of -na exists in the form of -nε in the SE, which is normally triggered by the use of a pronominal object in a habitual construction. This study, however, shows that the -nε undergoes different processes in Tɔŋu. In one instance, the habitual marker -a assimilates to -e before the pronoun “e”. In another instance, “a” assimilates to -e and then finally coalesces with the final sound of the verb to -ε before the pronominal object “e”.
The realisation of -na in Tɔŋu as found in this study therefore has the forms: -a, -e, -ε and -ɔ, depending on the kind of sound that ends the lexical verb. The study observes four phenomena that result in the variation in the use of -na as vowel lengthening, labialisation, palatalization and coalition. We noticed that vowel lengthening is one of the ways Tɔŋu marks habitual. For example, the final sound of the lexical verb is lengthened when it ends with -a or -ɔ in both transitive and intransitive constructions, that is vowel lengthening, a phenomenon misconstrued to be the situation of the Tɔŋu dialect in all cases. Another way in which -na is realised in Tɔŋu is by labialisation. We observed that when a verb ends with “u” or “o”, a /w/ is always inserted between the last sound of the verb and the habitual marker -a. The -a subsequently changes to “ɔ”. Under this phenomenon, “-a” which is unrounded changes to rounded “ɔ” after “w” to collocate with the rounding feature of “w”. The -a is however maintained if the initial sound of the following noun is “a”. Thus, we have a deletion of “n” and subsequent insertion of /w/, which results in labialisation, raising and rounding of -a. In the case of palatalization, the truncated habitual marker -a changes to “-e”. But before this, a palatal /j/ is introduced between the last sound of the verb, mostly if the last sound of the verb is “e”, thus -a changes to collocate with /j/. The introduction of /j/ causes both the habitual marker and the last sound of the verb to raise towards /j/. The realisation of -na as -ε, as we observed, happens due to a process known as coalition. The coalition of -a to -ε among Tɔŋu as noticed in this study can be said to be due to the presence of a pronominal object “e”. We observed that in Tɔŋu dialect, -a assimilates to -e before the pronoun “e” in a habitual construction. Both the -e and “a” of the verb coalesce finally into -ε.
We can therefore say that Tɔŋu dialect does not permit a cluster of vowels. Thus, if two vowels of different qualities follow each other, they form a hiatus. The hiatus is broken either by lengthening the final sound of the first word or the hiatus is broken by the insertion of /w/ or /j/ between the vowels. Consequently, either the first vowel or both vowels change their quality to look similar to the inserted consonant. Thus, the impression that -na is always realised by lengthening the final vowels of a verb is found not to be the case, therefore the assertion that vowel harmony does not exist in Eυe must be looked at again.