1 Towards the Reconstruction of the Tense-Aspect-Mood

Towards the reconstruction of the tense-aspect-mood (TAM) system in early Bantoid, with particular attention to the category “tense” John R. Watters, SIL International September 2012 Proto-Niger-Congo Congress, Paris

1.0 Definition of the question The general purpose of this study is to contribute to our understanding of the tense-aspect- mood (TAM) categories and systems in proto or early Bantoid. Within that broad purpose, this study will give particular attention to the status of tense in Proto-Bantoid or early Bantoid. All Bantoid languages today systematically mark aspectual categories (e.g. imperfective) and modal categories (e.g. subjunctive). However, only some Bantoid languages mark tense as a grammatical category such as the Grassfields Bantu and Narrow Bantu languages. In other Bantoid languages, tense is absent such as in the Ekoid Bantu languages.

This variation in the distribution of tense raises the question whether proto or early Bantoid was a tense-aspect language in which tense was lost in some subgroups over time, or was an aspect-prominent language in which certain subgroups innovated tense. If the latter case holds, I would want to ask additional questions: Where and when did the marking of tense as a verbal category emerge? What would it mean for our understanding of the early years of Bantoid, its subgroupings, and its internal migrations? Did the development of tense originate within one sub-set of languages from which it spread to others, or did various subgroups innovate tense independently of the others?

I would also want to compare the distribution of tense in Bantoid languages with the results of lexicostatistical studies. These studies to date form the basis upon which researchers identify Bantoid subgroups today. So how does the distribution of verbal systems with tense and without tense correspond to the results of lexicostatistical studies? Do they corroborate their conclusions or only make the picture of early Bantoid more complex? I should note that in terms of historical time Proto-Bantoid dates back possibly five millennia from the present (see Blench (2006: 126-138), Nurse (2008), and Ehret (2016: 106-116).

When referring to tense as a marked, systematic, grammatical, verbal category, I expect that an affix, clitic or particle relative to the verb will morphologically mark it. Otherwise, a compound verbal construction will indicate the tense. The construction will consist of an auxiliary verb indicating the tense followed by the main semantic verb, either in finite or non-finite form.

Combinations of these realizations or grammatical strategies may mark tense in the same system. These strategies will produce an inflectional verbal category within a paradigm in which at least past, present and future are formally distinguished. In addition, the tense paradigm must minimally involve both perfective and imperfective aspects. The perfective aspect may carry a label like ‘factative’ or ‘completive’ and the imperfective a label like ‘incompletive’ or ‘continuous’. Therefore, if there is a past tense, there will be both a past perfective and a past imperfective.

So we return to the basic question of this study: Was Proto-Bantoid or early Bantoid an aspect- prominent or tense-aspect language. From the answer to this question, I want to consider the implications for Proto-Bantoid or early Bantoid and the Bantoid subgroups.

2.0 Possible results in answering the question The first step in this study will be to determine the distribution of aspect-prominent Bantoid languages and tense-aspect languages. The distribution and attribution of verbal systems will be presented in terms of Bantoid subgroups rather than individual languages, but the verbal system typical of a given subgroup will be represented by at least one language within the subgroup. Given certain assumptions to be noted in section 3.0, this distribution could help us to:

1. determine the likely TAM system of Proto-Bantoid relative to tense, whether aspect- prominent or tense-aspect in form, and 2. determine the likely geographical area for the genesis of tense in Bantoid languages, whether the area where Proto-Bantoid was spoken or a subarea within the Bantoid region.

The second step will compare the results of this distributional study of aspect-prominent systems and tense-aspect systems within Bantoid with the results of recent lexicostatistical studies. We will see that there is a lack of alignment between the results of the distribution of TAM systems with the lexicostatistical results at two points. This lack of alignment actually provides a richer tapestry from which we can speculate about the early location of Proto- Bantoid and migrations of the Bantoid subgroups. So we will seek to:

3. derive a plausible identification of the geographical area where Proto-Bantoid was spoken, and 4. derive a plausible scenario for the migration of the early Bantoid subgroups.

The third step will be to range more widely than Bantoid. It has been proposed that the Bantoid languages are paired at the next point higher on the eastern Benue-Congo tree with the Cross River languages (see (1) below). Again, the distribution of aspect-prominent and tense-aspect languages will be mapped for the Cross River subgroups, providing evidence that could allow us to:

5. determine the likely TAM system of Proto-Cross River relative to tense, whether aspect- prominent or tense-aspect in form, and 6. determine whether any language(s) of the Cross River group that are tense-aspect in form today shared in the possible innovation of tense in Bantoid as noted as a possibility by Nurse (2008).

3.0 Assumptions underlying this study The possible results listed in 2.0 depend on certain assumptions that operate in this study.

First, following Nurse and Philippson (2003) and Nurse (2008), I assume that Proto-Niger- Congo was an aspect-prominent language. Even today most languages in West Africa are aspect-prominent. Yet in the central and eastern regions of Niger-Congo is a large contingent of over 500 Bantu languages that are well known for their elaborate tense systems. Tense appears in a few other areas within Niger-Congo (Supyire, eastern Kru, Nupoid languages, maybe Igbo, Ijo, and Zande), but tense in these other areas seems to be a local development in each case, unrelated to Bantu and the other Bantoid languages with tense (Nurse 2008). In addition, they only account for a small number of the total number of Niger-Congo languages outside Bantu and Bantoid. Most Niger-Congo languages outside Bantu and Bantoid have not innovated tense. The hinge between the western and the central and eastern portions of Niger-Congo is the Bantoid set of languages found along the Nigeria-Cameroon border. We will see that they in fact define the major border between aspect-prominent languages and tense-aspect languages. And even though Narrow Bantu that spreads across central, eastern and southern Africa has over 500 of the nearly 700 Bantoid languages within its subgroup, the number of subgroups parallel with Narrow Bantu within Bantoid is greater in the Nigeria-Cameroon borderland than elsewhere. In other words, we will see that most of the Bantoid groups stayed close to the Bantoid homeland while Narrow Bantu spread across Africa to the east and south of the Bantoid homeland in relative recent times, perhaps over the past two millennia.

Second, I assume that the expansion of Niger-Congo (excluding discussions about Kordofanian) was from west to east. Closer to the Bantoid region this means that the development of Benue- Congo subgroups was due to a migratory process that was generally west to east across Nigeria into Cameroon. The Bantoid languages would end up being the most easterly of these Benue- Congo subgroups, found somewhere along the Nigeria-Cameroon borderland. Its current 691 languages would mean that Bantoid accounts for 45% of the 1,532 Niger-Congo languages (Lewis 2009). This easterly migration pattern is critical to determining the most plausible location for Proto-Bantoid and a plausible location for the innovation of tense relative to Bantoid.

Third, I assume that Bantoid is a coherent grouping of languages, sharing in a common history. The important remaining question concerning the coherency of Bantoid involve one subgroup, namely, the place of the Dakoid within Bantoid. The coherence of Bantoid has been reinforced through various lexicostatistical studies over the past forty years that have included the Bantoid languages within their scope (Henrici 1973; Heine 1973; Coupez, Evrard and Vansina 1975; Bennett and Sterk 1977; Piron 1995 and 1997; and Grollemund 2012). Bantoid as such includes not only a collection of various groupings of languages in Cameroon and Nigeria, but also all of the Bantu languages spread across central, eastern and southern Africa. Based on this coherence, we understand that all of the various subgroups of Bantoid, including Bantu, derive from the variety of speech forms of an earlier Bantoid period perhaps five millennia ago in the Cameroon-Nigeria borderland (Nurse 2008:283 ??).

Fourth, I assume at least operationally that Proto-Bantoid Cross is a unit higher on Benue- Congo tree, above Bantoid and Cross River. Williamson and Blench (2000) proposed that the Bantoid languages share this wider common history with the Cross River languages. This is a proposal that has a certain plausibility but currently has less evidence supporting it than that supporting the unitary nature of Bantoid. However, if these two Benue-Congo groups do form a unit, it then widens the question of the distribution of tense by bringing the Cross River languages into the discussion moves the question further back in time to Proto-Bantoid Cross.

Based on these assumptions, I will take the steps mentioned in 2.0 by using the following process. First, I will consider in section 4.0 the best current hypothesis as to the origin of tense in Bantu and Bantoid. Second, in section 5.0 I will briefly take into account the status of tense and aspect in the various subgroupings of Bantoid. This brief tour through the Bantoid subgroups will lead us to a divided Bantoid. Third, in section 6.0, I will consider what the Cross River languages tell us about the place of tense in their history and their close contact with

4 some Bantoid languages. Fourth, in section 7.0, I will compare these results with those of the two most recent lexicostatistical studies, one by Piron (1999) and the other more recent one by Grollemund (2012). The lack of alignment between the results of those studies and the results of this study on tense will lead us to some hypotheses about the possible origin of tense in Bantoid, the possible location of the homeland of the Bantoid, and the possible migrations of Bantoid sub-groups in the its early millennia. We will finish with a summary of conclusions in 8.0.

4.0 Best current hypothesis The best current hypothesis regarding the development of tense in Bantu and Bantoid is found in Nurse (2008). The short form of this hypothesis is the following.

“…it would seem most likely in the present state of knowledge that tense was innovated within the community ancestral to today’s Bantu languages (2.10.2(iv, vii)) (Nurse 2008:282-283).”

In this study I hope to clarify what and where “the community ancestral to today’s Bantu languages” is. Nurse continues:

“Although most Bantu languages have multiple pasts and futures today (3.4, and Table 3.3 [and p. 279, 6.4]), careful examination of what can be reconstructed for Proto- or early Bantu suggests it most likely that a simple tense system developed first and became progressively and differentially more elaborate (Nurse 2008:285).”

Determining the plausibility of this statement will require a different study involving the other Bantoid languages with tense-aspect systems. So this study attempts to build a framework within which such a study could be pursued.

Now to provide more specifics: Nurse notes that three groups of languages along the southern Nigeria-Cameroon borderland are recognized as having integrated tense as a grammatical category into their verbal systems. These are Narrow Bantu (Zone A), Grassfields Bantu, and Lower Cross River. Assuming the ancestors of these three groups were in the same area in the early Bantoid era, Nurse says it is likely that it was in this region that tense was integrated into the verbal systems of at least some Bantoid languages, including Narrow Bantu and Grassfields Bantu. Nurse goes on to write that it is best to assume this scenario, given the current state of our knowledge, that it was within the Bantu ancestral community and its neighboring

5 community of Lower Cross River that tense was innovated within the verbal systems for the first time within Bantoid (and presumably Lower Cross River):

“Whether it [the innovation] started in one of them [Narrow Bantu, Grassfields Bantu or Lower Cross River] and spread to the others later, or whether it was an innovation shared at some level of the Bantoid-Cross River tree cannot be determined until we know more of the distribution and nature of tense in Grassfields and southeastern Nigeria. It is unlikely that different subsets of Bantu innovated the general phenomenon of tense reference at different times and places in eastern and central Africa. Since we can be fairly sure of that, but not sure when tense was innovated in Grassfields, Bantoid, and Cross River, it would seem most likely in the present state of knowledge that tense was innovated within the community ancestral to today’s Bantu languages (2.10.2(iv, vii)). The innovation of multiple tense distinctions characterizes all Bantu languages, but not in the same way. This suggests that once tense distinctions had developed while early Bantu communities and their languages were still in western Cameroon and eastern Nigeria, some five millennia ago, they were then carried and elaborated as communities moved east and south. The table above [cf. p. 279, 6.4] is based on ‘comparative Bantu internal evidence and suggests that a single simple pattern developed initially and ramified subsequently and kaleidoscopically (Nurse 2008:282-283).” (Italics mine)

This study seeks respond to and follow up on Nurse’s challenge that “until we know more of the distribution and nature of tense in Grassfields and southeastern Nigeria” we will not know how and where tense was innovated in the ancestral community of Bantu. Such a “ancestral community” could be found within the Bantoid grouping or it could be found in the larger Bantoid-Cross grouping.

One of the major complicating factors in the discussion about verbal systems in Bantoid is Zone A of Bantu found mostly in Cameroon, and more broadly, the northwest languages of Bantu (Zones A, B, C and beyond). The “northwest Bantu languages do not always behave as the others (Nurse 2008:124).” They do not have the same morphemes to mark tense and they are analytic in structure rather than synthetic and agglutinating like the rest of Bantu.

“In sum, examination of past and future reference in the northwest leads to the conclusion that, while the number of tense contrasts and their time reference in Zones A, B, C, D10, much of D20, D30, and H10 are similar to what occurs elsewhere in

Bantu, their morphological exponence is different. Phonological loss has removed or reduced traditional markers so that languages of the northwest have compensated by innovating new markers for old categories (Nurse 2008:126).”

Not surprisingly this region of greatest variety within Bantu, namely its northwest region, borders on the most diverse region of the wider Bantoid group. One can imagine the Bantoid groups in Nigeria and Cameroon and the Bantu languages of Zones A, B and C as possibly languages with older histories than those of central, eastern and southern Bantu that have resulted from more recent Bantu migrations perhaps over the past two millennia. In fact, Nurse notes:

“If we ignored the conventional classification (Narrow versus Grassfields), and if we ignored possible (?) diagnostic lexical and phonological features, and concentrated only on these verbal distinctions, it would be hard to uphold the unity of Zone A and the distinction between Zone A and Grassfields. These criss-crossing patterns must result partly from shared inheritance and partly from contact, and make it hard to be sure whether tense distinctions originated in Narrow Bantu or in southwestern Cameroon in general (Nurse 2008:126-127).”

It is to this question that we now turn. What is the evidence from the various subgroups of Bantoid languages, and what does that evidence tell us?

5.0 Evidence from Bantoid sub-groups In this section I provide brief presentations of the verbal systems of the various sub-groups within Bantoid, focusing on whether they are aspect-prominent or tense-aspect systems. In most cases the given Bantoid subgroup is exemplified by one language because of the lack of (in depth) studies on the languages of the region. At the same time, I recognize the inherent risks in limiting it in many cases to one language.

We will use Piron’s (1997:625) Bantoid classification to identify those subgroups. The conclusions of her classification in (1) are fairly conservative in that Narrow Bantu, including Zone A, remains an integral subgroup as opposed to the rest of Bantoid. Bantu is the first break from South Bantoid, though Jarawan Bantu shares in this first division from the rest of South Bantoid. However, one significant change for our understanding of Bantoid is that her study indicated there really is no sharp distinction between and South and North Bantoid, but that they the languages along this boundary form a continuum of related languages.

For purposes of presentation, however, we will start with what have been referred to as ‘South Bantoid’ (Watters and Leroy 1989, Watters 1989). In this case South Bantu included Jarawan, Narrow Bantu, Tivoid, Beboid, Wide Grassfields, Ekoid-Mbe, and Nyang. We will then turn to what have been referred to as ‘North Bantoid’, particularly Tikar and Mambiloid which are clearly part of Bantoid. From there we will consider the Cross River languages to see how they might inform the Bantoid results and how Cross River might be informed by Bantoid.

(1) Piron’s Bantoid classification (Piron 1997:625)

Bantoid Cross River

South Tikar Mambiloid (Narrow) Bantu

Dakoid Jarawan

Northwest

Tivoid Beboid Wide Grassfields Mbam B10-30

Narrow Grassfields Ekoid-Mbe Nyang Other Bantu

Ring Eastern Ndemli Mundani Menchum Momo

Before launching into the details of each subgroup, it should be pointed out in broad terms how the decision was made to consider a given subgroup as being aspect prominent and another as being a tense-aspect system. Aspect and tense are considered to be categories that participate in paradigmatic systems and not an ad hoc collection of categories. Because of the tendency in the Anglo-Saxon world to refer to almost anything inflectional having to do with verbs as ‘tense’, the various sources were considered carefully. If the author used terms such as ‘past tense’ and

‘present tense’ but the forms did not really fit into an integrated system of tense and aspect system, the categories were re-stated in aspectual terms. This should become clear in the presentation below and in the appendices. Before we go there, however, it would be helpful to provide an example of a system of categories that I would say demonstrates an aspect prominent system and one the demonstrates a tense-aspect prominent system.

An aspect-prominent system is demonstrated in (2a). A Bantoid example is given in (2b), using the Tiv language to demonstrate such a system. In the my analysis of these systems as in (2b) and in the appendices, what I considered to be the aspectual category was given in the abbreviated form such as PFV for ‘perfective’ or IMPV for ‘imperfective’. At the same time I noted the original author’s term by using double quotes as in “Past” tense (see 2a and 2b). This is to say it is not uncommon for some to analyze a perfective verb form as the “Past” tense even when tense is not systematically integrated into the verb system as a marked category. It is usually true that perfective (and ‘factative’) verb forms in their most neutral context have a past tense meaning since most complete events or situations have taken place in the past. The same goes with IMPFV ‘imperfective’ and its common identification with “Present” tense, ANT ‘anterior’ (also ‘perfect’) with “Recent past” and POT ‘potential’ with “Future”.

Looking at the Tiv verbal system from the perspective of a tense-aspect system we can note that there is only one form of the so-called “Past,” “Present,” and “Future” for example. In a tense=aspect system I would expect to find something on the order of (3a) and (3b). For example, I would expect minimally to find a “Past Perfective” and a “Part Imperfective.”

(2) Example of an aspect-prominent system – positive, indicative categories

a. A generalized form of an aspect-prominent language

ASPECT PFV (“Past”) IMPFV (“Present”) ANT (“Recent Past”) POT (“Future”)

b. An overview of the Tiv aspect-prominent system (Abraham 1940)

ASPECT

FACTATIVE (“Past”) ḿ-yèm ‘I went’ í-vɔ̂r ‘he was tired’ OR ‘he is tired’ CONTINUOUS (“Present”) ḿ ŋ̀gù érè-n ‘I am doing’ (lit: ‘he is in the act of doing’) I Cl1.is do-NOM.sfx ANTERIOR (“Recent Past”) ḿ kìmbí ú ‘I have paid you’ I pay.ANT you POTENTIAL (“Future”) m-á bê tóm ‘Tomorrow I shall finish the work’ I-shall finish work NOTE: m-á = [mə́] HABITUAL (“Present Habitual”) ḿ kìmbî hǎm mà ìyaŋge ‘I pay every day’ I pay.HAB every day

The aspect prominent system contrasts significantly with a tense-aspect system exemplified in (3).

(3) Example of a tense-aspect prominent system – positive, indicative categories a. A generalized form of an aspect-prominent language

ASPECT

T Past 2 (P2): Perfective (PFV) Past 2 (P2): Imperfective (IPFV)

E Past 1 (P1): Perfective (PFV) Past 1 (P1): Imperfective (IPFV) N Present (P0): Anterior (ANT) Present (P0): Imperfective (IPFV) S Future 1 (F1): Perfective (PFV) Future 1 (F1): Imperfective (IPFV) E Future 2 (F2): Perfective (PFV) Future 2 (F2): Imperfective (IPFV)

In the tense-aspect system in (3a), note that at least the major aspectual categories, namely perfective/factitive and imperfective, participate with tense in a systematically integrated system. The system in (3a) is only an example of one possibility. Consider the Tikar example in Appendix H as a Bantoid example. There are many different ways in which the actual tense- aspect system may be structured. For example, the aspectual distinctions may only be found in the past tense, while in the present and future tenses the difference between perfective and imperfective aspects is reduced to just imperfective aspect. The point here is that tense must be

10 demonstrated to play an integral, systematic role in the verbal system at some point that cuts across aspectual categories and perhaps even modal categories such as subjunctive for it to be considered a tense-aspect system. Without that integration, systematization the given language is actually working off tense meanings that are derived from the given aspects e.g. you can often infer past time from a perfective aspect. In an aspect prominent system tense is derived semantically rather than grammatically from actual inflected forms of the verb or particular auxiliaries sued in verbal compound constructions. The so-called “Past” in some analyses only indicates a complete situation, whether perfective or factitive, and the so-called “Present” only indicates a situation that is in progress or continuing, whether it is the imperfective or a related category such as the progressive or continuative.

So what do we find the in the Bantoid subgroups? The data for some, such as Grassfields Bantu and Narrow Bantu is extensive, while for others it is extremely limited such as for Jarawan, Tivoid and Mambiloid.

5.1 Jarawan There are no grammars or articles that present the tense-aspect-mood (TAM) system of a Jarawan language. What is mostly available are some wordlists and information on noun classes. However, the Jaar and Kulung communities are developing “Reading and Writing” booklets for their own communities. The information on tense and aspect in these booklets shed some light on the situation.

Jarawan languages appear to be aspect prominent languages. They do not mark tense systematically through inflection or constructions with auxiliaries. Instead, references to “tense” have less to do with grammatical systems than to the most neutral reading of the given aspect. So the following categories are identifiable:

Aspect category: May be identified as: Perfective/factitive aspect “Simple past tense” Imperfective aspect “Present continuous” Anterior “Perfect tense” Potential “Future”

The verbal morphemes for are primarily suffixes except for the imperfective which use a pre- verbal particle in an analytical structure. See Appendix A for more details.

5.2 Tivoid The language representing Tivoid is Tiv. The source is Abraham’s (1940) grammar of Tiv.

Tiv is also an aspect prominent language. The following categories are prominent: factitive, continuous, anterior, potential, and habitual. The factitive is synthetic in form with a subject prefix and verb root. However, the other forms are given in a more analytic form. See Appendix B for more details.

5.3 Beboid The language representing Beboid is Noni. The source is Hyman (1981).

Noni has a complex verbal system involving aspect, tense, mood, polarity as well relative and consecutive clause forms. It is definitely a tense-aspect language. Noni distinguishes eight time distinctions, four past (P0, P1, P2, P3) and four future (F0, F1, F2, F3). So Noni is like its Grassfields Bantu neighbors, being rich in tenses that cut across aspectual categories.

At the same time, Noni, a Beboid language, is said to relate closely to Tivoid based on lexicostatistical studies (Piron 1997, Grollemund 2012). Yet, it contrasts with Tiv or Tivoid in terms of its tense-aspect system. The best explanation at this time is that Beboid, i.e. Noni, gained its rich tense system from contact with the Grassfields Bantu languages. The Tivoid languages are further to the west and largely out of contact with the Grassfields languages because of distance and rugged terrain (i.e. mountains), and so never participated in the tense innovation that took place in the Cameroonian highlands and neighboring areas.

5.4 Ekoid Ejagham represents Ekoid Bantu. The sources are Watters 1981, 2010 and 2012a.

Ejagham is an aspect prominent language. Using terms from Nurse 2008, Ejagham marks factitive, imperfective, progressive/continuous, habitual, anterior, conditional and subjunctive. More details can be seen in Appendix D.

5.5 Nyang or Mamfe Bantu Denya and Kenyang represent this small group of three languages. Sources include Abangma 1987 (Denya), Mbuagbaw 1998 (Kenyang), and Watters (2012b).

The Nyang or Mamfe Bantu languages are aspect prominent even though Kenyang (Mbuaghaw 1998) has possibly developed some potential future auxiliaries that for now can be treated as marking potentials. The question about these future auxiliaries is whether they are really marking tense or simply adverbs occurring between the subject marker and the verb root. Also, there is no evidence that “past tense” is marked. These forms appear as perfectives or factatives. Abangma (1987) treats the forms of Denya as aspect. Neither language has a complex system involving tense with the major aspectual categories.

5.6 Grassfields Bantu Grassfields Bantu includes a large number of languages. Wide Grassfields numbers 60 plus languages. The sources would include Anderson 1979, 1983, Hyman 1979, 1980, Parker 1985, 1991a, 1991b, and Watters 1979, 1980, 2003.

There is no doubt that the Grassfields Bantu languages are tense-aspect languages, just like their neighbors, the Bantu languages of Zone A, to their southeast.

5.7 Narrow Bantu The 500 plus Bantu languages are well documented in Nurse’s 2008 extraordinary contribution on tense and aspect in Bantu. Nearly all Bantu languages are tense-aspect languages. Nurse (2008: 279ff) suggests that Proto-Bantu was a tense-aspect language with one past, one future, and a present tense.

5.8 First interlude: South Bantoid and distribution of verbal systems (4) South Bantoid

(Narrow) Bantu Jarawan tense aspect

Tivoid Beboid Wide Grassfields aspect tense tense Ekoid-Mbe Nyang aspect aspect

In (4) we now gain a sense of the distribution of aspect-prominent and tense-aspect languages in South Bantoid and the division that exists. We see that the following are aspect prominent: Jarawan, Tivoid, Ekoid, and Nyang/Mamfe. On the other hand, the Beboid, Grassfields, and Narrow Bantu are all tense-aspect languages. If we look for a common denominator among these languages we see that the aspect-prominent languages are primarily located in Nigeria, and the tense-aspect languages in Cameroon. We will come back to this topic below.

In light of this distribution it might be helpful to consider the so-called North Bantoid languages to see where they fit within this division. We will look at Tikar and the Mambiloid languages Vute and Mambila. No data is available on Dakoid, and that begs the question as to whether Dakoid is a Bantoid subgroup.

5.9 Tikar Tikar is a single language given its own status on the Bantoid tree. The source of the data is Stanley 1991.

Tikar is a tense-aspect language. Its word order in the perfective aspect is S V O and in the imperfective is S AUX O V. Tense is integrated into the aspect system of perfective and imperfective. There are four pasts including P0 and two futures. There is also a P0 narrative form. All imperfective forms take a suffix of a possible underlying shape –CV but with realizations as -ˊ , -li, -a, -ni, -e-, -bi, and -mi. The P1 and P2 each use a suffix that has similar surface variations: P1 uses a suffix -í but with variations -V, -lV, -bV, -mV, -nV, and P2 uses a suffix -é with variations -e, -le, -be, -me, and –ne. The phonological variation of these suffixes suggest the possibility that they represent older morphology than that used for P3 and F1 and F2. See Appendix H for more details.

5.10 Mambiloid - Vute Vute is included within the Mambiloid subgroup. The source of data is Thwing and Watters (1987).

The perfective aspect is marked by a high tone / ˊ / pre-verbal morpheme and a t ́ post-verbal particle. The imperfective is marked by a pre-verbal particle á and a n ́ post-verbal particle. The only exception is in F2 in which no post-verbal is present. Tense is marked pre-verbally, with yi for P2, tɨ for P1, and null ø for P0. In the futures, the distinction between perfective and imperfective is lost. The F1 looks like the imperfective, using the n ́ post-verbal particle. The F2

14 looks more like the perfective with the high tone pre-verbal morpheme. It also takes kwâ as a pre-verbal particle. See Appendix I for more details.

5.11 Mambiloid - Mambila It is difficult to know exactly what the TAM system is in Mambila since the data available (Perrin 1972) was presented around the topics of transitivity, grammatical relations, and ergativity. These categories do not require a systematic analysis of the TAM system in the language. However, from the data presented it is possible that Mambila has three degrees of past (P1, P2, and P3), a present (P0), and two degrees of future (F1 and F2) that have been integrated into an aspect system of perfective and imperfective, though on closer analysis it may prove otherwise. These parameters are not so different from those for Vute, the other Mambiloid language in this study. See Appendix J for more details.

5.12 Second Interlude: Piron (1997) in her study notes that the boundary between North Bantoid and South Bantoid is not remarkable, contrary to earlier presentations (Watters 1989, Watters and Leroy 1989). The three languages just reviewed, Tikar, Vute and Mambila, have been considered North Bantoid. What we see in terms of tense-aspect in each case appears to be a robust tense-aspect system.

Map 1: Bantu, Bantoid, Cross River & Jukunoid in the Cameroon-Nigeria borderland (J.R. Watters)

Tikar, Vute and Mambila represent speech forms (or ‘lects’) that broke off from the early Bantoid language possibly before any of the other Bantoid subgroups formed. However, they group formally and geographically with subgroups that innovated tense. First, they have tense- aspect verbal systems like Beboid, Grassfields and Narrow Bantu. Second, none of these languages are present on the plains of the Nigeria, along the eastern perimeter of the Benue Trough. Instead they are found on the eastern side of the Cameroon Volcanic Line inside Cameroon. Thus, they are also geographically and topographically closer to Beboid, Grassfields, and Narrow Bantu than they are to the South Bantoid languages with aspect-prominent systems.

The Cameroon Volcanic Line (from now on the “Cameroon Line” – see Map 1) runs from the Gulf of Guinea in the Atlantic to the Adamawa and Ngaoundere Plateaus in central Cameroon. In southern Cameroon and Nigeria they serve as a natural boundary between the eastern plains of Nigeria, and the highland regions in Cameroon just to the east of the border. In Cameroon this volcanic line includes Mount Cameroon, the highest mountain in West Africa, as well as the Manengouba, Bamboutos and Oku mountains. It also includes the South Cameroon Plateau, the Western High Plateau and the Adamawa and Ngaoundere Plateaus. These mountains and high plateaus present a very different terrain than the plains that stretch across most of West Africa and Nigeria.

To the east of these plateaus is the Sanaga River Basin. The Sanaga River separates the Cross- Sanaga-Bioko coastal forests to the west and north of the Sanaga from the Atlantic Equatorial coastal forests to the south and east (http://en.wikipedia.org/wiki/Sanaga_River). The Ejagham (Ekoid Bantu), Mamfe Bantu as well as the Cross River languages are found in the Cross- Sanaga-Bioko coastal forests. Most of these languages are in inland areas of the forest. The Mbam languages are found near the confluence of the Mbam and Sanaga Rivers, and the Narrow Bantu are present to the south and east of the Sanaga.

6.0 Cross River languages and Bantoid We now turn to the Cross River languages with which Bantoid is supposed to share a common history with East Benue-Congo.

First, recall that Nurse referred to three groups of languages along the Nigeria-Cameroon border area that had developed grammatical tense: Zone A of Narrow Bantu, Grassfields Bantu, and Lower Cross River. The possibility noted was that the ancestors of these three groups may

16 have been in the same area in the early Bantoid era when tense was innovated. In other words, there would have been a single genesis of tense among these three.

Second, Williamson and Blench (2000:31) joined Bantoid with the Cross River languages as a single subgroup of East Benue-Congo, one of three that make up East Benue-Congo. This was the first time the two groups were considered to be a unit. Crabb (1967) proposed that the Bendi languages near Ogoja and Ikom in Nigeria, as a subset of Cross River, was actually more closely related to Ekoid and therefore Bantoid than it was to other Cross River languges. His proposal has not been generally adopted. Instead it has been proposed that all of Cross River languages (including the Bendi languages) form a common unit with the Bantoid languages at a higher node in the East Benue-Congo tree. Given this proposal as well as the proximity of the Cross River languages to Ekoid, Tivoid, and a couple Bantu Zone A languages, and given Nurse’s reference to Lower Cross, it would be worth reviewing what we know of the tense- aspect in the Cross River languages.

Watters (2012b) has noted that not only are Ekoid and Mamfe Bantu aspect prominent, but all languages in the upper regions of the Cross River Basin (Manyu River in Cameroon) are aspect- prominent. Thus, for the Bendi languages that border on both Ekoid and Tivoid languages, Stanford (1967) in his grammar of Bekwarra only mentions aspectual and modal notions regarding the verb. Tense does not enter into his presentation. In addition, the largest Bendi language, Bokyi, is also apparently aspect prominent (Kierien Ekpang, p.c.) with no systematic integration of tense.

Turning to the Upper Cross River languages, just to the south of Bendi, and bordering Ekoid on the east and north, and the Igboid languages on the west, they also appear to be aspect prominent. Two examples of this would be Barnwell (1969) who like Stanford in Bekwarra does not mention tense, only aspectual and modal notions, and she has since confirmed the aspect prominence of the language (p.c.). The second example is Hyman et al (2002) in which Leggbo is demonstrated to be aspect prominent. So I conclude at this point that both Bendi and Upper Cross are aspect-prominent, forming an areal feature that is true for Tivoid, Ekoid, Bendi, and Upper Cross, all neighboring areas in Nigeria along the Cameroon border.

The case of the Lower Cross languages is different. Both Efik (Welmers 1973:408 ) and Ibibio (Essien 1983, 1990) are reported as tense-aspect systems. Efik has past, present and future tenses. Ibibio distinguishes P2, P1, Present, F1 and F2. Tense is marked by prefixes, and aspect uses prefixes and reduplication. The aspects reported are inceptive, habitual, and completive.

Meanwhile, another example is Obolo. Aaron (1999) claims Obolo makes a Future/Non-Future distinction, and therefore has tense. In terms of the analysis I am pursuing, I would still say that Obolo is an aspect-prominent language, with some incipient formation of tense, marking future. The future forms might also function as Potential aspect within the larger aspectual system. The place of tense is still not a thoroughgoing integration across past, present and future. So the presence of tense in the Efik-Ibibio (Central Lower Cross) shows at least a partial development of tense-aspect systems in this region, but it does not seem to have spread into Obolo, for example1.

In conclusion it would appear that Proto-Cross River was an aspect-prominent language like most languages to its west and likely its ancestral language, Proto-Benue-Congo, and that the development of tense in Central Lower Cross is a local development, not related to tense in Bantoid. What is the evidence? First, given the geographic distance between Lower Cross and the Grassfields and Narrow Bantu languages as well as the significant topographical distance with forest and a massive escarpment between, it is unlikely that Lower Cross participated with those languages on the eastern side of the Cameroon Volcanic Line. Second, all other Cross River languages appear to be aspect-prominent, and the Ekoid and Mamfe Bantu languages that separate Lower Cross from the lower reaches of Grassfields are also aspect-prominent. Third, the eastern spread of some Lower Cross languages along the coast into the region of tense rich Narrow Bantu Zone A languages in Cameroon is probably a relatively recent development.

At the same time, as a side note regarding Obolo and other languages said to be future/non- future in tense marking, it makes sense cognitively that an aspect prominent language might be able to begin developing tense through two routes. One route is to build on the perfective /factitive as the past tense and the imperfective as the present/future tense. The other route is to build on the fact that the perfective, though not a tense, provides for past tense situations, and the imperfective, though not a tense, provides for present tense situations. Cognitively this could leave a gap when it comes to the future. This future gap could then be filled with special constructions not found with the perfective or imperfective, thus explaining why some African languages have been analyzed as future/non-future languages.

7.0 Current lexicostatistical results and tense-aspect distribution The lexicostatistical study of Piron (1997) and the very recent study of Grollemund (2012) have provided us with important insights into the relationship between the non-Bantu Bantoid

1 Bruce Connell states that he does not see any other clear references to tense as a grammatical category in other Lower Cross River languages that have been recently studied (p.c.).

18 languages and Bantu, as well as the relationship of the various Bantoid languages to one another. One important insight is the likelihood that there is not a clear distinction between North Bantoid and South Bantoid (Piron 1997) contrary to earlier presentations (Watters 1989, Watters and Leroy 1989). Instead, there are only graded differences from north to south. The second important insight (Grollemund 2012) follows in a similar vein but relates Bantu Zone A to the rest of Bantoid. Here too there is not a clear distinction between non-Bantu Bantoid and the Bantu of Zone A but instead a graded, complex relationship between Zone A and the other Bantoid languages.

(5) South Bantoid (Grollemund 2012) – compare Piron’s classification in (1)

Nyang aspect Ekoid-Mbe aspect Tivoid Beboid aspect tense Grassfields tense A40-60-Jarawan+A31 tense- aspect + tense Bantu tense

What is striking about the analysis by Grollemund for this study is the placement of the Jarawan Bantu languages within a group of Bantu Zone A languages. A very different break up of Bantu relative to other Bantoid languages is found in Bennett and Sterk (1977). Grollemund goes beyond Piron (1997), Williamson and Blench (2000), and Blench (2006) who had placed Jarawan ever closer to Bantu within the larger South Bantoid group, but always distinct from Bantu. Grollemund incorporates Jarawan alongside a set of Bantu Zone A languages, namely A40-60 (Mbam) and A31 (Bubi on the island of Malabo).

Looking more closely now at the distribution of aspect and tense-aspect languages and their alignment or non-alignment with the results of lexicostatistical studies by Piron and Grollemund, at least two questions are raised for this study. These relate to the distribution of tense in (5).

1. Why does Jarawan relate so closely to Northwest and Central Bantu when it does not share any of Bantu’s innovation of tense? 2. Why does Tivoid and Beboid share such a close relationship and yet have very different verbal systems relative to tense?

Both these questions relate to the distribution fact noted earlier, that the aspect-prominent languages are largely Nigerian-based, and the tense-aspect ones are Cameroonian-based. We know that the distinction between Nigeria and Cameroon is irrelevant when we consider language groups going back four or five millennia.

So what is the distinguishing factor? I would suggest that it is the Cameroon Volcanic Line running from Mt. Cameroon all the way to the Tibetsi mountains of northern Chad. The relevant portion are the high mountains running from Mt. Cameroon through the Manengouba, Bamboutos and Oku mountains. In contrast to the plains on the eastern side of the Benue Trough, the mountains of the Cameroon Volcanic Line present a significant barrier to migration and to contact between communities on both sides of the ridges. So how does this relate to the Jarawan-Bantu Zone A relationship and the Tivoid-Beboid relationship?

In the case of Tivoid and Beboid, the explanation seems fairly straightforward. The Beboid are the eastern section of this unit and the ones that did migrate across the ridges and valleys into the highlands on the east side of the Cameroon Volcanic Line. They either arrived in time to participate in the innovation of tense with the Grassfields Bantu, Narrow Bantu, and possibly others, or they arrived later and adopted the tense-aspect systems they found among their Grassfields Bantu neighbors. Meanwhile, the Tivoid languages remained in the plains on the eastern side of the Benue Trough, although some migrated a little distance eastward into the mountains around Akwaya, but not far enough to have regular contact with those developing tense.

The Jarawan-Bantu Zone A relationship is the more problematic one. Two possibilities come to mind. The first is that the ancestors of the Jarawan were part of the migration across the Cameroon Volcanic Line and were part of the group that included the ancestors of the Bantu Zone a40-60 languages as well as A31 (See Grollemund 2012). They may have migrated as far as the Bamileke highlands or beyond to the region around the Mbam and Nkam rivers. However, before experimentation in tense began the ancestors of what became the Jarawan reversed their eastward migration and returned via a more northerly route into northern

Nigeria. Grollemund (2012:348) takes this perspective to explain the seeming close relationship between Jarawan with Zone A40-A60 and A31 languages. The one question mark is posed by the lack of any uniquely shared cognate or innovation between this group of languages. On the other hand, the distribution of the Jarawan languages suggests the possibility of a long migration northward in Cameroon, with possibly at least three language locating in eastern Cameroon and the rest continuing to the northeastern reaches of the Benue River and Benue Trough where they distributed themselves along the Benue River Valley, particularly into what is now Nigeria.

The second possibility is that the Jarawan ancestors were more closely related to those that migrated across the Cameroon Volcanic Line than the other Bantoid groups, but did not migrate eastward. Instead, they moved northward on the western side of the Cameroon Volcanic Line and became conservative in their lexical innovation and borrowing, displaying what we might call the “Icelandic effect”. They lost contact with all other Bantoid languages and were eventually separated from one another due to migrations of other language families such as the Chadic and Adamawa-Ubangian languages.

8.0 Implications: Proto verb systems, location of tense innovation, migrations and the evolution of Bantoid In section 2.0 six possible results were listed for this study. In this section these and other results are presented as plausible implications from what was presented in sections 4.0 through 7.0, if we take the assumptions listed in 3.0 as valid. They are plausible implications in that additional data could modify them. They are presented under three themes: Proto verb systems relative to tense and aspect, the location of tense innovation in Bantoid, and the geographic locations and migrations of Proto languages and subgroups.

1. Proto verb systems relative to tense and aspect: a. Proto-Bantoid was aspect-prominent. b. Proto-Bantoid Cross, if this is a coherent subgroup at the next higher level, was also aspect-prominent. c. Proto-Cross River was also aspect-prominent. d. Implications a.-c. are consistent with what would have been their likely inheritance from Proto-Niger Congo via Proto-Benue Congo, both assumed to have been aspect-prominent. e. Within Bantoid the following Proto languages were likely aspect-prominent: Proto-Nyang (Mamfe) Bantu, Proto-Tivoid, Proto-Ekoid-Mbe, and Proto-Jarawan.

f. Within Bantoid the following Proto languages likely had tense-aspect systems: Proto-Grassfields, Proto-Beboid, Proto-Mbam, Proto-Bantu. g. The development of tense in Lower Cross was separate from that in Bantoid and was a more local innovation even within Lower Cross. Thus, Proto-Lower Cross was likely aspect-prominent.

2. The location of the innovation of tense in Bantoid: a. Tense as a systematic grammatical category in Bantoid languages emerged east of the Cameroon Volcanic Line, in the high plateaus and slopes and the lower riverine regions around the Mbam River in the Sanaga River Basin. b. The Cameroon Volcanic Line serves as the boundary between aspect-prominent and tense-aspect languages, with the distribution of aspect-prominent Bantoid languages being spoken primarily in Nigeria today, west of the Cameroon Volcanic Line, and the tense-aspect systems in Cameroon to the east of the Line. c. It is impossible to know at this time the time depth for the innovation of tense, exactly what motivated its innovation, and exactly which group’s ancestors were involved and which groups might have borrowed and adopted tense from their neighbors east of the Cameroon Volcanic Line. Most likely ancestors of Grassfields, Mbam, and Narrow Bantu were involved, and possibly those of Beboid, Mambiloid and Tikar, or these may have borrowed and adopted what originated among the ancestors of Grassfields, Mbam and Narrow Bantu. d. Through the Bantu migrations, the originating center of which was the Mbam region (Grollemund 2012), tense-aspect systems spread throughout central, eastern and southern Africa.

3. Geographic locations and migrations of Proto languages and subgroups: a. The Proto-Bantoid Cross, assuming this is a valid unit, likely broke off from the Proto-Benue Congo group of languages somewhere in the Benue Trough and took up residence on the plains at the eastern edge of the Benue Trough near what is today the Cameroon-Nigeria border, perhaps between Ikom and Gboko. b. The ancestors of the Cross River languages were those who eventually migrated south from this Proto-Bantoid Cross homeland into the Cross River Basin in Nigeria in the southeastern region of the Benue Trough but the western side of the Cross River. This migration resulted in the differentiation of Cross River languages from Bantoid languages.

22 c. The Cross River languages moved south along the Cross River and then spread along the Atlantic coast first west and then east. The ancestors of the Bendi languages might have been more closely related to the Ekoid than the Cross River and the Bendi languages may be Bantoid instead of Cross River ones. At least the Bendi eventually became neighbors to the Ekoid and Tivoid languages. d. The ancestors of some of the Bantoid subgroups migrated east across the mountains and ridges of the Cameroon Volcanic Line uplift. Some remained in the mountains but others migrated to the eastern side of the Cameroon Volcanic Line. e. Bantoid migrations over the Cameroon Volcanic Line may have been multiple and periodic over the millennia, more than once moving from the plains of eastern Nigeria to the mountains and further onto the high plateaus on the eastern side of the Cameroon Volcanic Line. f. Some of the first migrations into the mountains and the high plateaus may have involved the ancestors of the Mambiloid and Tikar since they apparently are more distantly related to the other Bantoid subgroups. Reverse migrations may have taken place as well – not everything would necessarily be east and south. This would be a possible explanation for the apparent Jarawan connection with Bantu Zone A40-A60 languages. They migrated north either from the Mbam region in Cameroon or from the Proto- Bantoid region in Nigeria. The Cameroon origin is the more plausible. g. Grollemund (2012:348) argues that the fact that Jarawan does not share any unique lexical innovations with the languages of Bantu Zone A40-A60 while it does share an important number of lexical items suggests the Jarawan did not stay long in the Mbam region. We could add that they definitely left before the innovation of tense took place in the Grassfields (Western Plateau) and Mbam region. h. Some of the ancestors of what became Tivoid and Ekoid stayed close to the Proto-Bantoid homeland. The Ekoid might have moved south but north of the Cross River sometime after the ancestors of the Cross River languages migrated south. The Proto-Ekoid likely settled near what is today Ikom. From there, during the past millennia, the Ndoe cluster stayed close to home, the Mbe language and the Bakor cluster spread north from the home region and the Ejagham cluster spread east and south into the rainforest, some reaching the Calabar region before 1600 (Watters 1978, 1981). The Tivoid spread out east, west and north from the homeland over more recent millennia.

i. According to Grollemund (2012:403), the Proto-Bantu location was likely the Mbam region in Cameroon. This lies just below the highlands in western Cameroon which is on the east side off Cameroon Volcanic Line. Grollemund claims that the Zone A40-A60 languages form a clear transition zone from the Bantoid subgroups to the west and the Narrow Bantu languages to the east and south of the Mbam region. j. Grollemund (2012:339) concludes that Beboid, Tivoid, Ekoid and Grassfields are definitely part of the southern Bantoid languages. Kenyang, a Mamfe/Nyang Bantu language, presents a much weaker lexical relationship with these southern Bantoid languages even though it is neighbor to the Ejagham (Ekoid) to the west and the Momo languages (Grassfields) to the east. k. The Kenyang may be an early migration up the Cross/Manyu River to the base of the escarpment on the west side of the Bambouto Massif (or Bamboutos Mountains). The migration of the Eastern Ejagham up the same river took place later, sometime in the past millennia, probably at least more than 500 years ago.

Appendix A: Jarawan Bantu Jaar Jaar is a SVO language. It uses two verbal suffixes and a pre-verb form as the basic aspectual markers. The Jaar booklet (Usman and Hamidu 2012) recognizes these as aspects and does not refer to them as “tense”. The two suffixes are, taken verbatim from the booklet:

-am indicates “completed aspect, marking that the action has finished. This usually has a past tense meaning”, and -an indicates “potential aspect, marking that the action is potential, it may happen. This usually has a future tense meaning.”

In addition there appears to be what we might call within the terminology being used here an incompletive aspect that uses the morpheme ban with the meaning “is currently X-ing”. The language seems to be analytical rather than synthetic, except it does use the two suffixes noted above.

The vowel of the completed suffix –am varies in relation to the root vowel. The subject pronoun yi ‘3s’ is used in each case. Consider:

yi ɗaɓ-am ‘he took’

yi sən-em ‘he saw’

yi ka-ma ‘he went’

yi li-im ‘he ate’

yi li-im gərliim ‘he ate food’

The vowel of potential suffix –an also varies in relation to the root vowel.

yi ɗaɓ-an ‘he will take’

3s take-POT (=potential)

yi sən-en ‘he will see’

yi ka-na ‘he will go’

yi li-in ‘he will eat’

yi li-in gərliin ‘he will eat food’

The metathesis of the suffixes in relation to the verb ka ‘go’ is unexplained. Note that the object noun gərli- ‘food’ is also suffixed with the completed and potential aspect suffixes.

The morpheme ban ‘is-currently X-ing’ occurs between the subject pronoun. Whether there is any internal constituency to the morpheme ban or what its origin might be is not addressed.

yi ban ɗab ‘he is currently taking’

yi ban sən ‘he is currently seeing’

yi ban ka ‘he is currently going’

yi ban li ‘he is currently eating’

yi ban wal-a ‘he is beating me’

3s ICOM beat-3s

yi ban wal-ya ‘he is currently beating them’ 3s ICOM beat-3p

Kulung From the source available (Kulung Language Committee 2012), the category ‘tense’ is used extensively. That may be due to English being used in formal education in Taraba State in Nigeria where Kulung is spoken. It is not unusual for speakers of aspect-dominant languages who are educated in English to attach labels such as ‘tense’ to the aspect-marked forms of their language.

The Kulung source presents what is likely a perfective/factitive – imperfective distinction, as well as an anterior form and potential form. i ɓug-i ‘he beat him’ perfective/factitive (“Simple Past Tense”) 3s beat-3s

i ya ɓug-i ‘he is beating him’ imperfective (“Present Continuous Tense”) 3s be beat-3s

i ɓug-e yi ‘he has beaten him’ anterior (“Perfect Tense”) 3s beat-ANT 3s

i ɓug-na yi ‘he will beat him’ potential (“Future Tense”) 3s beat-POT 3s

yir-i ɓug-na yi ‘he will beat him’ potential (“Future Tense” with yiri) AUX-3s beat-POT 3s

The perfective/factitive aspect (“simple past tense”) is the least marked form of the verb. Since tone is not marked in the source, the best that can be said is that the simple verb root is used. The examples above combine the verb ɓug ‘to beat’ with the short (suffixed) form of the direct object pronouns.

The imperfective (“present continuous tense”) resembles the perfective/factitive. The only difference is the AUX ya ‘to be’ which occurs between the subject pronoun and the verb stem.

The anterior aspect (“Perfect Tense”) is marked by a suffix that resembles the short form of the possessive pronoun. The suffix agrees with the subject in person and number. Object pronouns take the long form of the object pronoun, at least in part since the ANT suffix has taken the suffix position of the verb stem.

The potential aspect (“Future Tense”) takes the suffix –na. Optionally, the auxiliary verb yiri may occur before the subject pronoun. If there is a semantic or pragmatic difference in the use and non-use of yiri it is not clear.

Appendix B: - Tivoid

ASPECT

Notes: Continuous (“The Present Tense” “better styled the ‘continuous’”) Copula followed by the “Continuous Verbal Noun” = simple verbal stem suffixed with –n ŋ̀gù érè-n ‘he is doing’ (lit: ‘he is in the act of doing’ is do-NOM.sfx The copula with the a ‘with’ following the copula has the sense of ‘to possess’

NOTE: Abraham writes: “we get the sense of the imperfect” when the verb lù is used instead of the copula. The verb lù ‘was’ (cf. 50(b)) has also the sense of ‘had’ ḿ lù érèn tóm ‘I was working’

I was do-NOM.sfx work

Potential/Conditional (“Definite Future” and “Indefinite Future”) Only distinguished in classes 1 and 2 (persons), otherwise all other classes use one form (the indefinite form) for both definite and indefinite uses a –á suffix on the verbal subject prefix, m-á [mé] ‘I shall …’ “Definite future” kpér m-é bê tóm ‘Tomorrow I shall finish the work’ tomorrow I-shall finish work the indefinite adds another –ꞌá má-ꞌá [méꞌá] ‘I will …’ “Indefinite future”

Appendix C: Beboid

See Hyman 1981 for details.

Appendix D: Ekoid Bantu

See Watters 1981, 2010 and 2012a.

(1) Examples of basic Ejagham aspectual categories in the Indicative, Positive and Negative

Affirmative Negative Anterior (‘Perfect’) á-rə̀bhè á-ká-rə̀bhé ‘they have opened’ ‘they have not opened’

Perfective Op Focus á-rə́bhè á-ká-rə̀bhé ‘they opened’ ‘they did not open’ Perfective Const Focus á-rə́bh-ˈé á-ká-rə̀bhé ‘they opened X’ ‘they did not open X’ Progressive á-kí-rə̀bhé á-bhó-rə̀bhé ‘they are opening’ ‘they are not opening’ Habitual/Concomitant á-rə̀bh-á á-bhó-rə̀bhé ‘they open’ ‘they do not open’ Imperfective Const Focus á-rə́bh-ˈá á-bhó-rə̀bhé ‘they are opening X’ ‘they are not opening X’

Appendix E: Nyang or Mamfe Bantu

See Abangma 1987 (Denya), Mbuagbaw 1998 (Kenyang), and Watters 2012b.

Appendix F: Grassfields Bantu

See Anderson 1979, 1983, Hyman 1979, 1980, Parker 1985, 1991a, 1991b, and Watters 1979, 1980, 2003 – particularly Watters 2003 for a summary.

Appendix G: Narrow Bantu

See Nurse 2008 – the definitive work.

Appendix H: Tikar

ASPECT

Perfective Imperfective S V O / S AUX O V / suffixes vary by degree of past new AUX + suffixes: (pp. 90-102) -ˊ , -li, -a, -ni, -e-, -bi, -mi used by every degree of past (pp. 102-128)

P3 -kàꞌ bɛ́kàꞌ … VERB-sfx T -kɛ̀n-kàꞌ ‘to leave’ à bɛ́kàꞌ mɛ̀-nye ɓwɛ́m-mí -wòe-kàꞌ ‘to open’ 3s IPFV.P3 houses build E ‘3s was building houses’ N P2 -é (=e, le, be, me, ne, e) ɓé … VERB-sfx S today -yɛn-ne ‘to see’ à ɓé gwè fyàeb-bi E -kɔ-e ‘to write’ 3s IPFV.P2 maize harvest ‘3s was harvesting maize’ (ɓé = copula PFV.P2) P1 -í (=V, lV, bV, mV, nV) ɓí … VERB-sfx -ɓì-i ‘plant’ à ɓí plè tae-li -yì-li ‘sweep’ 3s IPF.P1 dress sew-sfx ‘3s was sewing a dress’ (ɓí = copula PFV.P1)

-mɛ́ tǎ … VERB-sfx P0 à ɓɛ̀n-mɛ ‘he arrived’ à tǎ nyé wǒ il arriver-P0 3s IPFV.P0 house open ‘3s is opening the house’ P0 swum ndɔꞌ gwan-ø Nar- thing certain go.bad rative ‘something went wrong (in the vehicle)’ F1 kà … VERB-kà yɛ̀ … VERB-sfx à kà yɛn-kà ɓyin à yɛ̌ cìꞌ tɛ̀-lí 3s PfvF see-PfvF you 3s IPFV.F1 canoe steer-sfx ‘he will see you’ ‘3s is going to steer the canoe’

F2 ywìmɛ́ … VERB-sfx à ywìmɛ́ nún yɛ́n-ní 3s IPFV.F2 3s see-sfx ‘he/she will be seeing him’ A Brief of the Tikar Tense-Aspect System (Stanley 1991)

Imperfective P0 can express either a progressive/continuous or habitual sense. Semi-transitive verbs have a semantic shift depending on the presence or absence of the object. If the object is present then in P0 it has the sense of progressive/continuous such as ‘she is sewing a dress’. If

30 the object is not present, it gains a more general sense of habitual action, such as ‘she sews dresses (=she is a tailor/seamstress)’.

Appendix I: Vute (Mambiloid) ASPECT Perfective Imperfective T P2 yi ˊ VERB t ́ yi á VERB n ́ E P1 tɨ ˊ VERB t ́ tɨ á VERB n ́ N P0 ˊ VERB t ́ á VERB n ́ S F1 ɓa VERB n ́ E F2 kwâ ˊ VERB

An overview of the Vute tense-aspect system (Thwing and Watters 1987)

Appendix J: Mambila – Atta dialect (Mambiloid)

It is difficult to know exactly what the TAM system is in Mambila since the data available (Perrin 1972) was presented around the topics of transitivity, grammatical relations, and ergativity. These categories do not require a systematic analysis of the TAM system in the language. However, from the data presented it is possible that Mambila has three degrees of past (P1, P2, and P3) and two degrees of future (F1 and F2) that have been integrated into an aspect system of perfective and imperfective, though on closer analysis it may prove otherwise. There is a null ø form that may be the P0. (The four tone levels are indicated by numbers with

1 being the highest and 4 the lowest.)

Word order appears to be S V O or S AUX V O. a. be1 kee3 ‘they look’ they look b. be1 ngene2 yar23 ‘they see the buffalo’ they see buffalo

The following comment is provided in presenting the verb with various tense (and probably aspectual) morphemes but few examples are provided to demonstrate something of the overall system and how these morphemes actually co-occur. I quote (square brackets are my personal comments or queries]:

“Following the auxiliary, or if there is none, the verb, is an optional tense marker. It is convenient to think of these in two groups, as certain co-occurrences are permitted:

1. nde21 – future [F] 2. a21 – immediate (past and future) [P0?] [IMM]

na21 – past [P] ne1 – present [IMPFV? = Tikar n ́ ?]

ba21 – continuous [IMPFV: CONT]

[ø] – neutral? PFV? e.g. be1 kee3 ø ‘they look’ Co-occurrences: (note the changes in tone) nde2 a21 - near future [F1]

nde2 ne1 - future (unspecified) [F2]

na2 ba21 - past [CONT PAST??]

a4 nde2 cu2 a21 ‘he returned’ he go again tense [IMM?]

“d) Particle specifying time in the past, may precede the verb, in which case a [post-verbal?] tense [aspect?] marker, either a21 , ba21 or na21, is obligatory.

la3 – recent past (i.e. ‘this morning’) [P1] le1 – less recent past [P2] lo2 – distant past (i.e. the times of the ancestors) [P3] e.g. a4 le1 nde2 cu2 a21 ‘he returned’ 3s P2 go again tense [IMM?]

“e) As final item in the verb phrase, the following adverbial particles may occur: ken4 - completely

ka2 - still

be32 / yee32 - habitual

le2 / re2 - then

“There are certain restrictions on their occurrence: ken4 can only occur when tense [aspect?] markers a21, ba21, or na21 are present; ka2 may be followed by either be32, yee32, or le2, but no other combinations of adverbial particles within one VP are found e.g. a4 le1 nde2 cu2 a21 ka2 ‘he still returned’ 3s part. [P2?] go again tense [IMM?] still a4 le1 kuch21 gi2 a21 ken32 ‘he died’ 3s part. [P2?] die finish tense [IMM?] completely

References Abangma, S. N. 1987. Modes in Dényá Discourse. Dallas: SIL International. Abraham, R. C. 1940. The Principles of Tiv. London: Crown Agent for the Colonies. Republished 1968. Farnborough, Hants, England: Gregg International Publishers Limited. Aaron, Uche E. 1999. Tense and Aspect in Obolo: Grammar and Discourse. Dallas: SIL International. Anderson, S. C. 1979. Verb Structure in Aghem. In Hyman ed., Aghem Grammatical Structure. Los Angeles: University of Southern California. SCOPIL 7:73-136. Anderson, S. C. 1983. Tone and morpheme rules in Bamileke-Ngyemboon. Doctoral dissertation, University of Southern California. Barnwell, K. 1969. A grammatical description in Mbembe (Adun dialect), a Cross River language. Doctoral dissertation. SOAS. Bennett, Patrick and Jan Sterk. 1977. South Central Niger-Congo: a reclassification. Studies in African Linguistics, Volume 8:240-274. Blench, Roger. 2006. Archaeology, Language, and the African Past. Lanham, Maryland: Alta Mira Press. Coupez, A., E. Evrard and J. Vansina. 1975. Classification d’un échantillon de langues bantoues d’après la lexicostatistique. Africana linguistica 6:133-58. Tervuren: Musée Royal de l’Afrique Centrale. Crabb, David W. 1967. Noun classes of Cross River I (bè-ndì) languages. Mimeo. Ehret, Christopher. 2016. The civilizations of Africa: A history to 1800. 2nd edn. Charlottesville: University of Virginia Press. Essien, Okon E. 1983. The tense system in Ibibio. In J. Kaye, H. Koopman, D. Sportiche, and A. Dugas (eds), Current Approaches to African Linguistics (vol. 2), 329-344. Essien, Okon E. 1990. A Grammar of the Ibibio Language. Ibadan: University Press Limited. Grollemund, Rebecca. 2012. Nouvelles approaches en classification: application aux langues bantu du nord-ouest. Thèse pour Docteur en Sciences du Langage, Université Lumière Lyon 2. Henrici, Alick. 1973 Numerical classification of Bantu-languages. African Language Studies 14:82-104. Heine, Bernd. 1972/73. Zur genetischen Gliederung der Bantu-Sprachen. Afrika und Übersee 56:164-85. Hyman, L. M., ed. 1979. Aghem Grammatical Structure. Los Angeles: University of Southern California. SCOPIL 7. Hyman, Larry M. 1980. Relative time reference in the Bamileke tense system. SAL 11.2: 227- 237.

Hyman, Larry M. 1981. Noni Grammatical Structure. Southern California Occasional Papers in Linguistics 9. Los Angeles: University of Southern California. Hyman, L. M., H. Narrog, M. Paster, and I. Udoh. 2002. Leggbo Verb Inflection: A Semantic and Phonological Particle Analysis. In J. Larson and M. Paster (eds.), Proceedings of the Twenty-Eighth Annual Meeting of the Berkeley Linguistics Society. 28:399-410. Lewis, Paul. 2009?? Ethnologue. Dallas: SIL International. Mbuagbaw, Tanyi Eyong. 1998. Kenyang Lexicon. Youndé: Society for Kenyang Literature and CABTAL. Nurse, D. 2003. Aspect and Tense in Bantu Languages. In D. Nurse and G. Philippson, eds. The Bantu Languages, 90-102. London: Routledge. Nurse, D. 2008. Tense and Aspect in Bantu. Oxford: Oxford University Press. Nurse, D., S. Rose, and J. Hewson. 2010. Verbal categories in Niger-Congo languages. Kulung Language Committee. 2012. Reading and Writing Kulung. A proposal for writing the Kulung language. Draft edition. Taraba State, Nigeria: Kulung Language Committee. Parker, Elizabeth. 1985. Mood, tense and aspect in Mundani. Manuscript. Parker, Elizabeth. 1991a. Conditionals in Mundani. In Stephen C. Anderson and Bernard Comrie, eds. Tense and Aspect in Eight Languages of Cameroon, 189-210. Dallas: SIL. Parker, Elizabeth. 1991b. Complex sentences and subordination in Mundani. In Stephen C. Anderson and Bernard Comrie, eds. Tense and Aspect in Eight Languages of Cameroon, 189-210. Dallas: SIL. Perrin, Mona. 1972. Transitivity in Mambila. M.A. dissertation, SOAS(?). [Atta dialect, Cameroon]. Piron, Pascale. 1995. Identification lexicostatistique des groupes bantoïdes stables. Journal of West African Languages 25.2:3-39. Piron, Pascale. 1997. Classification interne du groupe Bantoid. 3 vols. Munich, Newcastle: LINCOM Europa. Stanford, Ronald. 1967. The Bekwarra language of Nigeria: a grammatical description. Doctoral dissertation. SOAS. Stanley, Carol. 1991. Description morpho-syntaxique de la langue tikar (parlée au Cameroun). Yaounde: SIL. (Thèse 1986). Thwing, Rhonda and John R. Watters. 1987. Focus in Vute. Journal of African Languages and Linguistics 9: 95-121. Usman, Ahmed Madiya and Musa Hamidu (Representing the Jaar Language Committee). 2012. Reading and Writing Jaar. A proposal for writing the Jaar language. (First Edition). Bauchi State, Nigeria: Jaar Language Project. Watters, John R. 1978. A reconstruction of Proto Ekoid phonology. Manuscript.

Watters, John R. (1979) Focus in Aghem: a study of its formal correlates and typology. In Larry M. Hyman, ed. Aghem Grammatical Structure. Southern California Occasional Papers in Linguistics (SCOPIL), Volume 7:137-198. Watters, John R. (1980) Notes on Ngie. Ms. UCLA and SIL. Watters, J. R. 1981. A Phonology and Morphology of Ejagham. Doctoral dissertation, UCLA. Watters, J. R. 1989. Bantoid Overview. In J. Bendor-Samuel (ed.), The Niger-Congo Languages. Lanham: University Press of America. 401-20. Watters, J. R. 2003. Grassfields Bantu. In D. Nurse and G. Philippson (eds.), The Bantu Languages. London: Routledge. 225-56. Watters, J. R. 2010. Focus and the Ejagham verb system. In I. Fiedler and A. Schwarz eds., The Expression of Information Structure: A documentation of its diversity across Africa. Amsterdam: John Benjamins Publishing Company. 349-376. Watters, J. R. 2012a. Watters, John R. Tone in Western Ejagham (Etung): Lexical tone on the “minimal" verb forms. In M. Brenzinger and A.-M. Fehn, eds. Proceedings, 6th World Congress of African Linguistics. Cologne 2009. pp. 317-327. Watters, J. R. 2012b (manuscript to be published) Ejagham without tense: historical implications for Proto-Bantoid. Watters, J.R. and J. Leroy. 1989. Southern Bantoid. In J. Bendor-Samuel (ed.), The Niger-Congo Languages. Lanham: University Press of America. 430-449. Welmers, Wm. E. 1973. African Language Strictires. Berkeley: University California Press. Williamson, K. and R. Blench. 2000. Niger-Congo. In B. Heine and D. Nurse eds., African Languages. Cambridge: Cambridge University Press. 1-41