What is finiteness in Dagaare?

2.1 The notion of finiteness across languages

According to Nikolaeva (2007: 1), the term “finite” originates from the Latin word finitus, the perfect of the verb finio ‘finish, set bounds to, circumscribe, bring to a close’. This etymology suggests that Classical grammarians used the term “finite” to indicate a property of clauses that characterised them as complete in terms of making a determined proposition or proposal. So far as Latin is concerned, the Classical grammarians were arguably right to identify inflectional categories of the verb such as tense, mood, and agreement features as markers of finiteness in the verbal domain since, in Latin, these grammatical categories sufficiently circumscribed clauses as full propositions.

The absence of these lexicogrammatical categories in languages other than the Standard Average European (SAE) languages has led some linguists to question whether finiteness should be considered a distinct grammatical system across languages (Bisang 2007). It has been argued that finiteness is an epiphenomenon of verb inflection or other grammatical categories rather than a distinct system in its own right (Cristofaro 2007). Others have contended that finiteness has an important role in the syntactic and semantic structure of clauses and should be considered a distinct grammatical system (Klein 2006: 2).

Several proposals have been offered to explain finiteness as a grammatical system. It has been considered as “a functional tendency defined by a cluster of correlating parameters” (Nikolaeva 2010: 1176, see also Cristofaro 2007). Across languages, parameters that have been identified include tense, aspect, modality, verbal modes, mood, person, agreement, politeness/honorification, and switch reference (see e.g. Estrada-Fernandez 2016; Nikolaeva 2010). Other parameters are assertion and information structure (Kalinina and Sumbatova 2007; Klein 2006). Important to the present study, Zhang (2019) identifies clause final aspect particles (CFAPs) as markers of finiteness in Mandarin Chinese. He provides robust evidence for the distinction between finite and non-finite clauses in Chinese: clauses that allow CFAPs behave like finite clauses in other languages and clauses that disallow CFAPs behave like non-finite clauses in other languages. Chinese finite clauses (i.e. clauses with CFAPs) are typically independent, allow pro-drop, allow speaker-oriented adverbs and epistemic modals, and their clause boundaries may be visible for binding. On the other hand, clauses that disallow CFAPs are intrinsically embedded, ban pro-drop, disallow speaker-oriented adverbs and epistemic modals, and their clause boundaries may be invisible for binding.

If one considers the grammatical resources associated with finite categories across languages, one will realise that they are resources that characterise the clause as an interactive unit, a move in exchange. They encode interpersonal meaning and ground the clause as something that can be argued about or negotiated (Halliday and Matthiessen 2014: 144). In other words, finiteness is often associated with obligatory categories that “locate the state of affairs being described with respect to the speech act situation” (Cristofaro 2007: 92). Sybesma (2017: 233) similarly defines a finite clause as a clause “that enables one to make a connection with the non-linguistic outside world”.

Unsurprisingly, the free declarative clause has been identified across languages as the prototype of finite clauses (Anderson 2011: 234–356; Nikolaeva 2010) since it is free clauses that have a full interactive potential in exchange. Thus, many studies on finiteness show that finiteness relates to the semantics of freedom in the clause and that reduction in finiteness is a signal of dependence on the textual context (Givón 1990: 853, see also Nikolaeva 2010). It has however been argued that in some languages (e.g. English and Tundra Nenets), forms associated with finiteness such as tense and agreement markers occur in both free and bound clauses (Anderson 2011: 234–356; Nikolaeva 2010). Indeed, across languages, there is the tendency that while non-finite forms are associated with bound clauses, finite forms may be found in both free and bound clauses. A counterargument is that finite bound clauses often require an additional marking through subordinators or some other binding items (Anderson 2011: 290–356; Matthiessen and Thompson 1988). In English, for instance, hypotactic finite clauses are always marked by a subordinator, while non-finite ones such as infinitival clauses normally do not require subordinators (Halliday 1985: 217–218; Matthiessen and Thompson 1988).

While some scholars (e.g. Bisang 2007) consider finiteness to be a discrete binary system of finite/non-finite, many studies maintain that finiteness is a cline, a scalar property characterised by a correlation of parameters (Anderson 2011: 290–356; Chamoreau 2016; Comrie 2016; Givón 1990). The cline or scalar view is that “finiteness and non-finiteness are scalar categories defined in terms of a variety of properties that may combine in different ways from one construction to another” (Cristofaro 2007: 92). This view is exemplified by Givón (1990: 852–860), who considers finiteness as a gradual category, namely a construction is “more or less” finite based on the number of properties it possesses that deviate from the pattern of prototypical free clauses.

Another issue has been the appropriate unit of analysis for finiteness. While some studies have considered it as a property of verbs (Anderson 2011: 357–395), others have analysed finiteness as a clause rank system (Anderson 2011: 234–356, see Nikolaeva [2010] for an overview). Sells (2007) makes a crucial distinction between finiteness as a property of the clause and its overt morphological form. He distinguishes among: (i) finite as a value of a form feature that verbs and possibly a few other lexical items have; (ii) finite as a formal grammatical property of clauses normally encoded by a Finite form or by other means); (iii) finite as a formal property that certain categories may be sensitive to, notably agreement, complementizers, and negation; and (iv) finite “as a property of clauses used to make an assertion” (Sells 2007: 86). Maas (2004) similarly makes a distinction between semantics finiteness and morphological finiteness using data from across languages. He characterises semantic finiteness as “the grounding of the sentence” (Maas 2004: 361). He explains further that semantic finiteness is the condition for an independent interpretation of a sentence (or clause) and concerns all questions of reference or deictic binding. Morphological finiteness, on the other hand, is marking of the verb, “the analysis of grammaticization in morphology” (Maas 2004: 381).

The present study is guided by insights from the crosslinguistic research on finiteness. Notably, it contributes to the dynamic or multifaceted perspectives on finiteness (see e.g. Maas 2004; Sells 2007). It however goes beyond morphological marking to provide a broader definition of grammatical finiteness and outline more explicit criteria in distinguishing between semantic finiteness and grammatical finiteness by further using insights from SFL language descriptions (see Section 3). I relate these criteria to the controversial issue of whether finiteness is binary or a cline, focusing on Dagaare.

2.2 The notion of finiteness in systemic functional language description

In SFL, finiteness is a descriptive concept rather than a theoretical concept. It is a concept that emerges from the empirical description of particular languages and from crosslinguistic comparison and not an a priori category given by the theory. Indeed, although finiteness has featured prominently in SFL language description, there is little dedicated SFL research on finiteness (see e.g. Yang 2004, 2022). The discussion of finiteness in SFL has rather largely been related to the description of mood in Indo-European languages. A possible explanation for this typological concentration is that most of SFL descriptions have focused on clause systems and finiteness tends to play a more crucial role in the mood systems of the Indo-European languages described (e.g. Bartlett 2021; Caffarel 2006; Figueredo 2021) than in the mood systems of other languages described in systemic functional terms (e.g. Akerejola 2005; Kim et al. 2023; Rose 2021; Teruya 2006; Wang 2021; Zhang 2020a, 2020b: 157–238).

In M. A. K. Halliday’s description of the English mood system (e.g. Halliday and Matthiessen 2014, see also Matthiessen 1995), he identifies a Finite operator that combines with the Subject of the clause to constitute the Mood (see e.g. Examples 1–3).^[2] Mood (with initial caps) is defined as the arguable or interactive component of the clause. Also, in a Mood Tag, it is the Finite and the Subject that are selected to appendage the clause as an interpersonal finale (see e.g. Examples 1–3). The indicative mood type in English is identified by the presence of Subject + Finite (e.g. She is coming here) and the imperative is identified by the absence of these two elements (e.g. Come here). Within the English indicative, the declarative is identified by Subject ^ Finite order (as in: She is coming here) while the interrogative is identified by Finite ^ Subject order (e.g. Is she coming here?).

The English Finite has been defined as the element that circumscribes the proposition and brings it down to earth as something that can be argued about. In other words, it relates the proposition to its context in the speech event by combining polarity with the specification of either temporal or modal reference (Halliday and Matthiessen 2014: 145). As Example (1) shows, in English, the Finite can be conflated with the Predicator when tense is realised by verbal inflection. Halliday and Matthiessen (2014: 396–398, 396) also posit Finite as an element of the English verbal group and identify a finite verb at word rank. Thus, in English grammar, the system of finiteness involves “double agnation” (Christian M. I. M. Matthiessen personal communication), defined in the context of the present discussion as a situation where a system of a higher rank such as the clause can also be posited as a system of a lower rank such as the group or word.^[3] At clause rank, Halliday and Matthiessen (2014: 162) posit finiteness as a delicate system of bound clauses since it is bound clauses in English that truly show a dichotomy between finite (e.g. What she should say) versus non-finite (e.g. What to say). Finite bound clauses are identified by the presence of Finite and Subject. All free clauses (e.g. She said it right) are naturally finite. The more fundamental point, however, is that, in English, a binary system of finiteness can be identified for the clause, verbal group, and the verb such that we can distinguish between finite clauses and non-finite clauses, finite verbal groups and non-finite verbal groups, and finite verbs and non-finite verbs. Martin et al. (2021) use the specific terms “Tense” and “Modal” as functional labels in place of Halliday and Matthiessen’s (2014) verbal group Finite. We can consider the term Finite in this context to be a higher level of abstraction combining “Tense” and “Modal”.

Fawcett (2000a, 2000b) challenges the representation of the interpersonal structure of the English clause by M. A. K. Halliday. He replaces the term Finite with Operator, which he defines as the element that expresses “the MOOD meaning of ‘polarity seeker’” in the clause (Fawcett 2000a: 172). Finiteness, according to Fawcett (2000a) is “time reference position” and is either encoded in the Operator (e.g. is in It is working well) or in the Main Verb/Predicator (e.g. It works well). One essential difference between Fawcett’s (2000a, 2000b) and Halliday and Matthiessen’s (2014) accounts is that Fawcett neither recognises the verbal group nor the concept of rank scale. Thus, the issue of finiteness serving as agnate systems across ranks does not arise in Fawcett’s version of the interpersonal structure of the English clause.

In Romance languages, typological differences have been identified in the encoding of finiteness. In French, Caffarel (2006: Ch. 3) identifies a Finite element as a constituent of the interpersonal structure of the clause. Like English, the order of the Finite in relation to the Subject shows a distinction between declarative (Subject ^ Finite) and polar interrogative (Finite ^ Subject) clauses although mood distinction in French can alternatively be indicated by intonation. The French Finite can also be conflated with the Predicator through verb inflection. In Portuguese, as Figueredo (2021: 209) shows, “the presence or absence of the Finite is the most important distinction between an indicative and an imperative clause”. Thus, in both French and Portuguese, the Finite is an integral component of the negotiability of the clause and is essential in making mood distinctions at clause rank. Like English, both languages also allow Mood tagging involving Finite ^ Subject at the end of a declarative clause.

Spanish is typologically different from French and Portuguese. Quiroz (2008, 2021) provides an extensive discussion of finiteness in Spanish. According to her, the Finite is not a discrete element of the clause in Spanish (see also Lavid et al. 2010). For instance, the language does not allow Mood tagging involving a Finite element. Rather, “the Predicator is the fundamental function grounding the Spanish clause to the speech event” (Quiroz 2021: 56). Spanish has rich verbal morphology and the verb realising the Predicator is inflected for person, mood, and tense. Quiroz (2021) thus identifies the Finite as an element of the verbal group, where it is always conflated with the Event, the verbal group function of the lexical verb. Bartlett (2021) shows that, in Gaelic (Celtic), a Finite can combine with any attendant Mood clitic to constitute the negotiation component of the clause, though, like English and French, the Gaelic Finite can be conflated with the Predicator.

It can be concluded that in most of the Indo-European languages described in SFL terms, the Finite is a crucial element in the modal structure of the clause and is essential in determining the mood type of a clause. The present study will show that Dagaare is typologically different from the Indo-European languages in the encoding of finiteness. Like Spanish, there is no justifiable reason for identifying a Finite element in the international structure of the Dagaare clause. Unlike Spanish, however, in the Dagaare verbal group, a Finite element is realised by a particle and Finite does not conflate with Event. The details will be discussed in Section 4.

4.1 The modal structure of the Dagaare clause

One notable strategy for determining the negotiatory elements of the clause in SFL language description is to identify those elements that are “replayed and adjusted in responding moves” in dialogues (Quiroz 2021: 48). A second strategy is to identify elements that serve as Mood Tag since tagging involves intersubjective elements at risk in negotiating meaning. Dagaare does not allow Mood tagging per se, although a similar function is performed by clause final modal particles. When we observe exchanges in Dagaare discourse, the crucial negotiatory elements we identify are Subject, Predicator, Mood Marker, and Negotiator. Let’s observe the dialogues in Examples (4) and (5) taken from a transcript of an unscripted play performed by school children:

In the responding moves in these dialogues, certain elements are replayed and adjusted through choices in person, polarity, and mood. In the responding moves in both Examples (4) and (5), the nominal groups realising the Subject are adjusted to show a shift in the modally responsible person in the clause, i.e. a shift from addressee (Fʋ̀, ‘you’) to speaker (Ɩ̃̀, ‘I’). In Example (4), Speaker B repeats the verbal group (wõ̀, ‘hear’) realising Predicator in Speaker A’s utterance. The clause final modal particle na serving as Negotiator asserts the proposition; it is an interpersonal finale. The Negotiator is obligatory in the clause structure as has been indicated by the asterisk against the brackets. It should be noted that there is a correspondence between the focus particle nι (clitic forms = n, = ι) in Speaker A’s question and the affirmative particle na in Speaker B’s response. In Dagaare, the focus marker, in addition to signalling newsworthiness, also contributes to mood distinction. That is, the focus marker is restricted to “indicative: affirmative” clauses and neither occurs in “indicative: non-affirmative” clauses nor in “imperative” clauses. Thus, in the interpersonal structure of the clause, its function is Mood Marker. In the absence of explicit focus marking in the responding clause by Speaker B (the clause is all-new focus), the clause final particle na, i.e. the Negotiator, is used to assert or affirm the proposition.

In Example (5), Speaker D’s responding clause shifts polarity, and negotiatory elements in the clause are appropriately adjusted. First, the verbal group serving as Predicator incorporates a negative particle bɛ́, which also marks mood (indicative) and tense (non-future). Crucially, the lexical verb bàw (‘know’) is maintained in the verbal group and the Complement is ellipted. Also, the clause ends with the non-affirmative Negotiator = ɛ́. It should be noted that the tense-mood-polarity (TMP) bearing particle bɛ́ is an integral component of the Predicator rather than a separate constituent of the clause. Notably, it cannot combine with the Subject alone to indicate negotiation (*Ɩ̃̀ bɛ́) as it would be the case with a Finite in English (No, I don’t). As the dialogues also show, in Dagaare, a Finite does not inverse with the Subject to distinguish between declarative and interrogative clauses. Thus, unlike most of the Indo-European languages (viz. English, French, Portuguese, and Gaelic) discussed in Section 2.2, there is no justifiable reason for positing a Finite element as a constituent of the Dagaare clause. Neither is the Predicator alone sufficient to enact negotiability in the clause as it is the case in Spanish (Lavid et al. 2010; Quiroz 2008, 2021). The negotiatory load of the clause is carried by the Subject, Predicator, and Negotiator or Mood Marker, whichever is present. These elements have been collectively labelled as the Mood Base in the analysis in Examples (4) and (5).

4.2 Finiteness in the Dagaare verbal group

This section proceeds to examine finiteness in the Dagaare verbal group. The verbal group functions as the Predicator/Process in the Dagaare clause, as Example (6) illustrates.

The Dagaare verbal group has a nucleus element in its structure, namely the Event, realised by the lexical verb serving as Head of the group (e.g. dɩ̀, ‘eat’ in Example 6). This nucleus can be expanded by the addition of modifying particles, which I categorise into Finite elements (e.g. the TMP marker nà in Example (6) and Extensions (e.g. the directionality marker tɩ́ in 6).^[5] The full structure of the Dagaare verbal group is posited as follows: (Finite ^) (Extension ^) Event (^ Extension), where the caret indicates that the elements are ordered, and the brackets indicate optional elements. This structure can be expanded by additional Finite and Extension elements. Examples (7) to (9) provide illustrations – the first row after the morpheme-to-morpheme interlinear glossing presents functional labels (i.e. Finite, Extension, Event) while the second row identifies the word class (e.g. particle, verb) of the items realising the functions. The final row indicates word rank functions of items in the verbal group.

As illustrated in Examples (8) and (9), the Dagaare verbal group can take multiple Finite elements in its structure (up to three elements). When a Finite is present in the verbal group, it is placed at the initial position of the group and is realised by particles indicating polarity, mood, tense, and modality. In Example (7), the Finite is the particle na, which encodes future tense, positive polarity and indicative mood. It contrasts with bɛ in Example (8), which indicates non-future tense, negative polarity and indicative mood. Example (8) has three Finite elements, namely the TMP particle bɛ́ and the habitual (mɩ) and remote past (tɩ) tense markers. In Example (9), the Finites are the modality particle naa, which encodes desirability and positive polarity, and the habitual (mɩ) and remote tense (tɩ) markers. The Finite enacts the verbal group as an interactive bundle in the clause. It grounds the clause in the temporal space of the speech event and (inter)subjectively through choices in polarity, mood, tense, and modality.

The verb serving as Event is obligatorily marked for aspect, the grammaticalization of the internal temporal structure of events (Comrie 1976: 6, cf. Dahl and Velupillai 2013). Aspect in Dagaare is an opposition between perfective and imperfective and it is encoded by verbal inflection. While the perfective is realised covertly (Examples 10a, 11a, and 12a), the imperfective is overtly marked by the suffix -rɛ/re (Examples 10b, 11b, and 12b), which has eleven allomorphic variants conditioned by various vowel and consonant harmony systems (see Mwinlaaru 2017: 74–79).

Since aspect marking is obligatory for every verb serving as Event irrespective of whether the clause in which it occurs is free or bound, aspect does not differentiate between finiteness and non-finiteness in Dagaare. There is thus no basis for distinguishing between finite and non-finite verbs in Dagaare. In other words, finiteness is not a word rank system in Dagaare.

Extensions are realised by verbal particles that contribute to meaning in the verbal group in one of two ways: they either extend the meaning of (i) the Finite (if it is present) or (ii) the Event (see Table 1). Particles that contribute to the meaning of the Finite comprise directionality, eventuality, and conditionality particles. These items are semantically similar to the finite markers in the sense that they relate to deicticity (directionality) or modal assessment (conditionality). Although eventuality is a non-deictic category, it shares a similar semantic space with tense by indicating temporality.

Extensions that contribute to the meaning of the Event are realised by the polysemous particle nɩ, indicating causative, comitative, transitivising, and instrumentality meaning. Interestingly, Extensions that contribute to the meaning of the Finite precede the Event, where they occur together with the Finite (if present) and those that contribute to the meaning of the Event follow the Event, sometimes forming one phonological word with the lexical verb as clitics (Examples 13b and 14b).^[6] They contribute to the valency of the lexical verb. In Table 1, I characterise these valence particles as afinite forms since they are not relevant to the differentiation between a finite verbal group and a non-finite verbal group. The widespread morphological similarity between the particles in the Dagaare verbal group as set up in Table 1 suggest that they most possibly share common grammaticalisation sources (Heine et al. 1991; Kuteva et al. 2019, see also Mwinlaaru and Yap [2017] on grammaticalisation in Dagaare).

The placement of elements in the verbal group arguably follows the principle of iconicity (Haiman 1985). The Finite and other particles semantically related to it are thematically placed at the beginning of the verbal group to orient the interpretation of the verbal group. Elements that contribute to valency are placed after the Event where they are closer to the participants serving as complementation of the verb.

In summary, the verbal group in Dagaare can be said to be finite or non-finite depending on whether a Finite (i.e. a TMP, modality, habitual, or remote tense particle) is present in its structure. A binary system of finiteness can thus be posited for the verbal group as represented in Figure 1.

Figure 1:

Finiteness in the Dagaare verbal group represented as a system network.

The section will proceed to discuss the items realising the elements in the verbal group in more detail to throw further light on finiteness.

4.2.1 The obligatory finiteness markers

A minimal realization of a finite verbal group in Dagaare will be Finite ^ Event (Example 15). The non-finite verbal group has the Event as the only obligatory element (e.g. wa, ‘come’ in Example 16). As will be noted frequently in subsequent discussion, the Finite can be covert in the structure of the verbal group (e.g. as indicated in Example 16 by the null symbol).

(15)

Dɛ̀r	nà	kʋ́	nà.
name	fut.ind.pos	kill.pfv	affr
Subject	Predicator		Negotiator
nominal group	verbal group
Thing	Finite	Event
noun	particle	verb	particle
‘Der will kill (it).’

(16)

Ʋ̀	Ø	nyɛ̃̀	nà	[[fʋ́	wà ]].
3sg	nfut.ind.pos	see.pfv	affr	2sg	come.pfv
Subject	Predicator		Negotiator	Complement
nominal group	verbal group			nominal clause
Thing	(Finite)	Event		Thing	Event
pronoun		verb	particle	pronoun	verb
‘S/he saw (that) you have come.’

The obligatory finiteness markers in the Dagaare verbal group are the tense-mood-polarity (TMP) and modality particles as well as remote and habitual tense markers (see Table 1). Each of the TMP particles (e.g. na, kʋ̃, bɛ) encodes three grammatical systems simultaneously: futurity (future vs. non-future), mood (indicative vs. imperative), and polarity (positive vs. negative). futurity in Dagaare is a contrast between future (marked by na, ‘positive’ and kʋ̃, ‘negative’) and non-future (marked by bɛ in the negative and covert in the positive). mood distinction in the verbal group, on the other hand, is between indicative (marked by na, kʋ̃, or bɛ) and imperative (marked by ta or taa). polarity is a contrast between positive (na) and negative (marked by kʋ̃ or bɛ in the indicative and ta or taa in the imperative) (Tables 2 and 3).

Table 2:

Realisation of polarity in the indicative clause.

Polarity	Futurity
	Non-future				Future
Positive	Dɛ̀r	cén	nɩ̀	súkùul.	Dɛ̀r	nà	cen	nɩ	súkùul
	name	go.pfv	foc	school	name	fut.ind.pos	go.pfv	foc	school
	‘Der’s gone to school.’				‘Der will go to school.’
Negative	Dɛ̀r	bɛ́	cén	súkùul=é.	Dɛ̀r	kʋ̃́	cén	súkùul=é.
	name	nfut.ind.neg	go.pfv	school=naffr	name	fut.ind.neg	go.pfv	school=naffr
	‘Der didn’t go to school.’				‘Der won’t go to school.’

Table 3:

Realisation of polarity in the imperative clause.

Polarity	Immediacy
	Immediate		Non-immediate
Positive	Cén! go.pfv ‘Go!’		–
Negative	Tá	cén = é!	Táa	cérè = ɩ́!
	neg.imp.im	go.pfv=naffr	neg.imp.nim	go.ipfv=naffr
	‘Don’t go!’		‘Don’t go (when he asks you out)!’

The imperative-polarity markers ta and taa are finiteness markers that occur in the imperative clause. They correspond to the TMP particles na, kʋ̃, bɛ in indicative clauses. While ta carries the meaning of immediate prohibition (Examples 17 and 18), taa indicates non-immediate prohibition (Example 19). The former can combine with either the perfective (Example 17) or imperfective form of the verb in (Example 18). The latter, however, occurs with only the imperfective form (Example 19):

(17)

Tá	tɛ́r	=ɩ́	nɩ́r	sùur	=é!
neg.imp.im	possess.pfv	=com	person	anger	naffr
Predicator			Complement	Complement	Negotiator
verbal group			nominal group	nominal group
Finite	Event	Extension	Thing	Thing
particle	verb	particle	noun	noun	particle
‘Do don’t have grudges against anybody!’

(18)

Tá	yɛ́rɛ̀	nɩ́	mã̀á	=ɩ́!
neg.imp.im	speak.ipfv	=com	1sg.emp	=naffr
Predicator			Complement	Negotiator
verbal group			nominal group
Finite	Event	Extension	Thing
particle	verb	particle	pronoun	particle
‘Don’t talk to me!’

(19)

Táa	yá-rɛ̀	yél-fàa	sàn	=ɛ́!
neg.imp.nim	pay-ipfv	matter-evil	debt	=naffr
Predicator		Complement		Negotiator
verbal group		nominal group
Finite	Event	Classifier	Thing
particle	verb	noun	noun	particle
‘Don’t pay back evil! (= Never pay back evil!)’

The use of a modality particle (specifically, naa, kʋ̃ʋ) in realizing the Finite in the verbal group is normally mutually exclusive with the use of the TMP particles (na, kʋ̃, bɛ). The exception is the median modality particle taa, which can occur with the TMP particles na, kʋ̃, bɛ (see Example 25 where modal taa co-occurs with positive future marker na). modality in Dagaare is a system of possibility and desirability . Possibility is the modalisation of propositions, distinguishing between what is likely and what is unlikely. The clauses in Examples (20) to (22) indicate different polarity values of possibility.

(20)

Dɛ̀r	nàá	wá	kʋ́-rɛ̀	nɩ́	bʋ̀ɔ.
name	mod.pos	evt	kill-ipfv	foc	goat.
Subject	Predicator			Mood Marker	Complement
nominal group	verbal group				nominal group
Thing	Finite	Extension	Event		Thing
noun	particle	particle	verb		noun
‘Der may be killing (a) goat.’

(21)

Dɛ̀r	kʋ̃̀ʋ́	wá	kʋ́-rɛ̀	bʋ̀ɔ	=ɩ́.
name	mod.neg	evt	kill-ipfv	goat	=naffr.
Subject	Predicator			Complement	Negotiator
nominal group	verbal group			nominal group
Thing	Finite	Extension	Event	Thing
noun	particle	particle	verb	noun	particle
‘Der may not be killing (a) goat.’

(22)

Dɛ̀r	táa	wá	kʋ́-rɛ̀	bʋ̀ɔ.
name	mod	evt	kill-ipfv	goat.
Subject	Predicator			Complement
nominal group	verbal group			nominal group
Thing	Finite	Extension	Event	Thing
noun	particle	particle	verb	noun
‘Der might be killing (a) goat.’ (‘= in case Der is killing a goat’)

Naa in Example (20) encodes positive possibility – ‘it is possible that Der will be killing a goat’. The clause also has a focused element, indicated by the focus particle nɩ, highlighting what the listener should consider as newsworthy. Modality particle kʋ̃̀ʋ́ in Example (21) encodes negative possibility; it is the proposition that is negated: ‘it is possible that Der will not be killing a goat’. Example (22) indicates a median value of possibility, the speaker enacts an uncommitted middle ground, making no claim to the polarity value of the proposition. The clause has a reduced assertion since there is no focus marker nor a clause final Negotiator. The speaker suspends asserting the proposition, a strategy used to show weak possibility.

Desirability, on the other hand, is the modulation of proposals, characterising them as either desirable or undesirable. Examples (23) to (25) illustrate the encoding of desirability in Dagaare:

(23)

Ɩ̃̀	nàá	tɩ́	tɛ̀r	ʋ̀
1sg	mod.pos	pst.rem	possess.pfv	3sg
Subject	Predicator			Complement
nominal group	verbal group			nominal group
Thing	Finite	Finite	Event	Thing
pronoun	particle	particle	verb	pronoun

nɩ́	à	ɩ̃̀	zìe	kà.
foc	def	1sg	place	here
Mood Marker	Adjunct
	adverbial group
	Deictic	Location
particle	determiner	adverb complex
‘I should have liked to keep him here with me.’ (Bible.is, Fɩlɩmɔ 1: 13a)

(24)

Fʋ̀	kʋ̃̀ʋ́	cén	=é.
2sg	mod.neg	go.pfv	=naffr
Subject	Predicator		Negotiator
nominal group	verbal group
Thing	Finite	Event
pronoun	particle	verb	particle
‘I wish you don’t go.’/‘I wish you hadn’t gone’.

(25)

Nɩ̀	táa	nà	zɩ̀	ʋ́?
2pl	mod	fut.ind.pos	sit.pfv	int
Subject	Predicator			Negotiator
nominal group	verbal group
Thing	Finite	Finite	Event
pronoun	particle	particle	verb	particle
‘You may want to sit down?’/‘I don’t know if you want to sit down’

Negative desirability Example (23) is a negation of the proposal and not the modality – ‘I wish you don’t go’ – and the modal particle taa encode weak desirability (Example 24), the speaker is not committed to the desire for the actualisation of the proposal or otherwise. It is up to the addressee to decide (Example 25).

Tense markers other than those encoding futurity consist of the habitual (mɩ́) and remote past marker (tɩ) and I will call them secondary tense markers because they are placed after the future/non-future (encoded by na, kʋ̃, and bɛ). My use of the term secondary tense is thus different from its use for English (Halliday and Matthiessen 2014: 396–410) where it implies a hypotactic relationship between a primary tense marker and a secondary tense marker. Habitual and remote tense markers can co-occur with the primary tense or futurity markers (na, kʋ̃, and bɛ), as shown in Example (26). Also, both the habitual marker mɩ́ and remoteness marker tɩ can co-occur in the same verbal group (Example 26):

(26)

Ɩ̃̀	bɛ́	mɩ́	tɩ́	cén	=ɩ́
1sg	nfut.ind.neg	hab	pst.rm	go.pfv	=com
Subject	Predicator
nominal group	verbal group
Thing	Finite	Finite	Finite	Event	Extension
pronoun	particle	particle	particle	verb	particle

líbìr	=é.
money	=naffr
Complement	Negotiator
nominal group
Thing
noun	particle
‘I used not to go with money.’

The habitual marker mɩ́ can occur in imperative clauses as well (Example 27):

(27)

Mɩ́	bín	dàɩ!
hab	put.pfv	fermented local beer
Predicator		Complement
verbal group		nominal group
Finite	Event	Thing
particle	verb	noun
‘Always reserve fermented beer/pito (for us)!’

The fundamental point in the use of multiple Finite elements in the Dagaare verbal group structure is that Dagaare grammar has evolved disassociated systemic variables of tense, namely futurity, remoteness, and habituality. Since these temporal variables are disassociated, they must be realized by different grammatical items (see also Mwinlaaru [2021] on an analogue of multiple Deictic elements in the Dagaare nominal group). As Example (26) illustrates, it is essentially choices from the three tense systems, or their combination with modality, that results in multiple Finite elements in the verbal group. Choices from each of the three tense systems together ground the clause to the temporal context of the speech situation. I illustrate this point in the system network presented in Figure 2. The figure shows that while futurity and habituality are simultaneous systems, remoteness is a delicate system available after the choice of non-future.

Figure 2:

A system network of tense and its interaction with polarity in the Dagaare verbal group.

4.3 Finiteness in the semantics of the Dagaare clause

This sub-section proceeds to discuss finiteness at the clause rank in Dagaare. As already mentioned, finiteness is not a grammaticalized system of the Dagaare clause. Finiteness in the Dagaare clause is rather a fuzzy semantic category, which we defined as the aggregate of linguistic resources that are required to ground a proposition or a proposal to the speech event and/or its textual context as a move in exchange and a message (Section 3). Clausal finiteness in Dagaare therefore has to do with interpersonal and textual semantics and clauses form a cline from finite to non-finite. Since the free indicative clause has been identified crosslinguistically as the prototype finite clause (Anderson 2011: 234–356; Nikolaeva 2010), we will first identify the obligatory grounding devices of this prototype finite clause. In Dagaare mood system, the “indicative” is a supercategory consisting of declarative and interrogative clauses, i.e. “indicative”: “declarative”/“interrogative” (see Mwinlaaru 2017: 138–161; Mwinlaaru et al. 2018).^[7]

Three indicators of finiteness can be identified in the free indicative clause in Dagaare: (i) a finite verbal group, i.e. a verbal group with a Finite in its structure; (ii) clause final negotiation, and (iii) information focus or the presence of a Mood maker. These are characterised as indicators of finiteness in the clause because they are required in the clause to ground it as a free or an independent unit. As Examples (40) and (41) show, the clause final Negotiator and information focus marking (or the presence of Mood Marker) are mutually exclusive.^[8]

A more extended discourse example of focus and negotiation as markers of semantic finiteness is given in (42), taken from a film:

Line 001 in this dialogue is a biclausal construction. The initial clause is the main clause, and it is a finite indicative clause. Focus is placed on the wh- item (Bʋʋ) through a cleft-construction (see Mwinlaaru 2017: 234–237). The verbal group include a covert Finite, indicating non-future tense and positive polarity in the indicative mood. The complement clause, on the other hand, is non-finite, the verbal group is also non-finite, and neither is there independent focus marking or clause final negotiation. Also, the initial clause in the clause complex in line 002 is a projecting desiderative clause and takes the clause final Negotiator = a. The projected proposal is an imperative clause. The declarations made by the excited Batolemi who has been healed of his blindness in lines 003 and 004 all end with a clause final Negotiator. The verbal groups in these clauses all have a covert Finite, indicating non-future tense and positive polarity. The crucial role of the clause final modal particle in indicating clausal finiteness in Dagaare makes Dagaare typologically similar to Chinese where clause final aspect particles are noted to encode finiteness (Zhang 2019).

Now the question is how do other clause patterns fit into the characteristics of the free indicative clause in Dagaare? I outline the characteristics of various clause patterns in Dagaare below, showing the extent to which they are semantically finite.