FOCUS: Nowhere

Chapter 10

FOCUS: Nowhere

1. Introduction In addition to the opposition between paradigmatic and syntagmatic FOCUS (cf. Chapter 9), there is another opposition, one that depends upon, and is limited to occurrence within, syntagmatic FOCUS. In this chapter, we turn to the expression of FOCUS that has been labelled “in situ”. To assert the existence of FOCUS in situ languages presupposes that we can determine what their neutral, basic or unmarked word order is. Unless we can do that, there is no situs to speak of.1 We will quickly discover that the “FOCUS in situ” label is too broad. Languages that have been given this appellation differ significantly among themselves, and some are in fact not FOCUS in situ at all. Languages are usually identified as FOCUS in situ by reasoning through the following four steps: (i) There is an overt expression of FOCUS; (ii) There is an identifiable range of usages for that expression, e.g. answers to wh-questions; (iii) The range of usages of (ii) are additionally associated with an in situ expression; (iv) The conclusion is that the in situ expression must share a sense of FOCUS with the overt expression of FOCUS, which in (i) was associated with the usages. It is the fact that in situ FOCUS references order that limits it to syntagmatic FOCUS. The contrasting non-in situ FOCUS of (i) is usually called ex situ, even though order may not be the actual morphosyntactic mark of its presence.2 The existence of FOCUS in situ languages forces us to modify our original claim that the semantics of language can only be expressed by three morphosyntactic means: syntagmatic contrast in order, paradigmatic contrast in alternate forms, and contrasting suprasegmentals. In “true” FOCUS in situ languages, FOCUS is nowhere to be found in the morphosyntax, and such languages depend upon contrasting contexts to distinguish different semantics.

1 Within the frame of the opposition between paradigmatic and syntagmatic FOCUS (cf. Chapter 10 and below), in situ FOCUS is confined to the latter. This seems self-evident since the use (or non-use) of order that is the sine qua non of in situ FOCUS presupposes a syntagmatic choice, which is absent from paradigmatic FOCUS.

2 Cf., for example, Northern Sotho in section 3.1 below. 2 SYNTAX & SEMANTICS

Examination of FOCUS in those languages which appear to have FOCUS in situ has the beneficial effect of placing in sharper relief the role which the semantics of ASSERTION plays in shaping the semantics of FOCUS in general. ASSERTION is at the foundation of paradigmatic FOCUS (cf. Chapter 9), but it may also appear in syntagmatic FOCUS. The exploitation of ASSERTION for the purpose of expressing syntagmatic FOCUS is a graded phenomenon. While, at one extreme, there are languages whose syntagmatic FOCUS relies exclusively on ASSERTION, at the other extreme, there are languages which configure syntagmatic FOCUS while completely avoiding any appeal to ASSERTION. Still others will restrict ASSERTION to specific places in their systems of syntagmatic FOCUS. Recognizing this, we may distinguish languages along this dimension, i.e., the degree to which the semantics of ASSERTION is present in the organization of syntagmatic FOCUS. The first type will rely entirely on ASSERTION. Bella Coola, Wolof, Eastern Armenian, and Mupun (section 3.3 below) are of this type. Reliance on ASSERTION does not imply that its morphosyntactic expression will be the same in each language. Bella Coola, Wolof, Somali (Chapter 9, section 3.2.1), and Mupun are strikingly (and interestingly) different in how they have gone about formally integrating ASSERTION into FOCUS. The second type is split. These languages will employ ASSERTION (in paradigmatic FOCUS) as the sole mark of focused Agents, while the other functions will have another FOCUS that is ASSERTION free.3 The third type of language will give expression to ASSERTION in a way tangential to FOCUS. ASSERTION is present and intersects orthogonally with it, but ASSERTION does not itself give expression to FOCUS. Hausa (section 3.2 below) is a language of this type. While not relying on ASSERTION to compose FOCUS, Hausa permits the expression of ASSERTION as an independent modulation of each of the three degrees of FOCUS which the language does express. The fourth type is like Hausa, but the tangential use of ASSERTION is restricted and does not occur with all degrees of FOCUS. Modern Standard Arabic (section 2.1) illustrates this type. The fifth type separates the semantics of ASSERTION from the semantics of syntagmatic FOCUS entirely. Except for coexisting in the same utterance, they do not interact at all. Miya, Modern Greek and (probably) Haida are of this type.4 Rather than grouping all languages which fail to use order to signal FOCUS together, with no further distinction, and instead of using only the formal

3 We will return to try to explain why this should be so in the following chapter as part of the explanation of why FOCUS has an affinity for EVENTS and certain positions in word order.

4 I am not aware of any reason to connect the morphology of -¢uu with the semantics of ASSERTION. FOCUS: Nowhere 3 criterion that FOCUS is expressed in the same word order position which the focused content would otherwise occupy, were it not focused, I now propose a functional semantic behavior of in situ FOCUS. The outcome will be that some, but not all of the FOCUS in situ appearing languages are truly FOCUS in situ. The proposal uses the typology of the preceding pragraph. This is the FOCUS in situ Conjecture:5 (i) A language has more than one degree of FOCUS in terms of a scale of LAX versus INTENSE FOCUS; (ii) The semantic composition of the force of FOCUS is not structured in terms of ASSERTION (e.g., as it is in Wolof); (iii) When this occurs, the weaker degree of FOCUS will be expressed by a “weaker” morphosyntax, i.e., in situ, with or without the assistance of suprasegmentals. The FOCUS in situ Conjecture makes predictions that are easily falsifiable.6 It also raises further questions. If a language satisfies conditions (i) and (ii), is it necessarily the case that it will be FOCUS in situ? And unless we allow some nonarbitrary connection between meaning and expression, why should the weaker FOCUS be expressed by the formal means that is FOCUS in situ? We will begin to address these questions in section 3.3 in a discussion of Mupun. The involvement of ASSERTION in FOCUS will be fairly easily to detect, either by the specific morphosyntax that is used,7 or by the semantics associated with the morphosyntax,8 or both.9 Conversely, the absence of ASSERTION from the constitution of FOCUS will also be easily identified.10 The languages we discuss in this chapter will be presented according to the neutral word order attributed to them in the literature: VSO, SVO, and SOV.

2. In Situ Languages that are VSO

2.1 Modern Standard Arabic11

5 Recall that since in situ FOCUS is dependent upon syntagmatic FOCUS, this conjecture applies only to that kind of FOCUS.

6 So it may have a very short life span.

7 Bella Coola.

8 Hausa (below).

9 Mupun (below).

10 A useful heurisic in recognizing the presence of ASSERTION is that an utterance that can be questioned yes/no is one that is asserted. But ASSERTION, like FOCUS, comes in degrees. Cf. Davis Ms.c. There are many other treatments of ASSERTION in the literature.

11 The following sources provide the material for this section: Bahloul 1993, Bakir 1980, 4 SYNTAX & SEMANTICS

We shall first examine Modern Standard Arabic (MSA). Our reason for beginning with MSA is that it appears in some ways to challenge our previous conceptualization of FOCUS. Moutaouakil (1989.1) describes MSA thusly:

Modern Standard Arabic, essentially a written language, is the vehicle of culture in the Arabic world; it is the language of contemporary Arabic literature, of the press and of education. Although Modern Standard Arabic is markedly different from ‘Classical’ Arabic, it continues to display the same fundamental structure, for largely cultural reasons.

The language is also known as Literary Arabic, about which Bakir (1980.1 & 3) writes:

... no Arab acquires this form of the language at home and therefore there is no NATIVE speaker of it ... By Literary Arabic I mean the variety of Arabic used by educated Arabs for a variety of linguistic functions. It is used exclusively in books, newspapers and writing media; radio, television and even in personal letters. Literary Arabic represents a continuation of Classical Arabic . Diferrent terms have been used to name this form of Arabic: Standard Arabic’ Modern Standard Arabic; Modern Written Arabic ....

There is unanimity in attributing a neutral word order of VSO to the language:

... although Arabic allows a number of variations in the word order of the sentence, the VSO order should be taken as ... the basic order .... (Bakir 1980.3)

Il existe en Arabe un ordre non marqué (ou neutre) selon lequel les constituants sujet et objet occupent, dans la phrase verbale (don le prédicat est un verbe), les positions VSO .... (Moutaouakil 1984.144-145)

I assume ... that SA [Standard Aabic] is underlyingly an SVO language and VSO word order is achieved through verb movement to INFL. Such a movement is rendered obligatory in SA so verbs get tense features ....” (Bahloul 1993.214)12

I accept the fact that VSO is indeed the discourse neutral word order (Mohammad

Mohammad 2000, Moutaouakil 1984 & 1989, Ouhalla 1993, 1994a, 1994b, 1997a, 1997b & 1999, Plunkett 1993, and Shlonsky 2000. Mohammad 2000 manages to discuss variation in word order in MSA and Palestinian Arabic for 180 pages without once mentioning stress, FOCUS, or even [+f].

12 Which is an oblique way of saying the language is VSO. FOCUS: Nowhere 5

2000.1)13

It is well-known that in Standard Arabic (SA), the verb occurs clause-initially and the unmarked order of consitutents is VSO. (Shlonsky 2000.325)

Some examples are the following:

(1) (Ouhalla [see-MASC:3SG the-boys-NOM Zayd-ACC 1997a.201) ‘The boys have seen Zayd’

(2) (a) (Plunkett 1993.242) [study-3MS-IMP the-students-NOM] ‘The students are studying’

(b) Qadima al-walad-u (Bahloul 1993.209) [came-3MS the-boy-NOM] ‘The boy came’

(3) (a) Al-walad-u fii al-bayt-i (Bahloul 1993.211) [the-boy-NOM in the-house-GEN] ‘The boy (is) at home’

(b) Kaana al-walad-u fii al-bayt-i14 (Bahloul [was the-boy-NOM in the-house-GEN] 1993.212) ‘The boy was in the house’

(4) Zayd-un mudiir-un (Bahloul 1993.211) [Zayd-NOM director-NOM] ‘Zayd is the director’

Wh-questions in MSA are formed by placing a question word sentence-

13 “However, I see no a priori reason for assuming that the discourse neutral surface word order should be the underlying word order” (Mohammad 2000.1).

14 “... the presence of the copula in a present/timeless context is ... obligatory ... (i) [may be the-earth-NOM rounded-ACC] ‘The earth may be rounded’ The obligatory presence of the copula is also observed in nonpresent time contexts, such as those referring to remote or forthcoming events ...” (Bahloul 1993.210). 6 SYNTAX & SEMANTICS initially, and only initially (Ouhalla 1997b.24):

... preposing of a wh-phrase (in simple wh-questions) is necessary for a (information-seeking) wh-question reading ... SA is not a wh-in situ language. In SA, a sentence with a single wh-phrase in situ cannot have a wh-question reading.15

These are from Moutaouakil (1984.124 & 1989.22-23):16

15 What this means is that there are no true questions on the model of (i), which is an echo question (Ouhalla 1993.285): (i) Qaabal-a Zaydun man! [meet-3M Zayd-NOM who] ‘Zayd met who!’ There are in situ questions if there is another wh-word sentence initially (Ouhalla 1993.285 & Moutaouakil 1984.139): (ii) Man qaabal-a man? [who met-3M who] ‘Who met whom?’ (iii) [qui a-visité qui] ‘Who visited whom? Cf. also Ouhalla 1994a.69. Moutaouakil (1989.47) adds, “the number of Focus terms (=the number of interrogative pronouns) may never exceed three”: (iv) [who informed who with-what] ‘Who informed whom of what?’ (v) * [who informed who with-what where] ‘Who informed whom of what where?’ But ‘When?’, ‘How?’, ‘Where?’, and ‘How?’ are not possible in situ questions anyway, so the number three probably has nothing to do with (v) (Moutaouakil 1989.48): (vi) * [when came when] ‘When did who come?’

16 I am not certain how to match the French grammatical glosses with English, so I have left them in French. I have, however, converted the free glosses from French to English. Neither Moutaouakil 1984 nor Moutaouakil 1989 segments Arabic noun stems from their inflection. I have taken the liberty of doing so. FOCUS: Nowhere 7

(5) (a) [qu’est-ce-que as-mangé-tu] ‘What did you eat?’

(b) [ai-mangé-je tharid-acc] ‘I ate tharid’17

(d) % [lequel ai-mangé-je tharid-nom] ‘What I ate was tharid’

The same pairings of answers to questions are not provided for Agent questions, but Moutaouakil 1989.40 does have this question and answer:

(6) (a) [who wrote book-ACC] ‘Who wrote a book?’

(b) [wrote Zayd-NOM book-ACC] ‘Zayd wrote a book’

(c)

17 “Tharid ... is [a] traditional Arabic dish made of pieces of bread in vegetable or meat broth. The dish is notable in that it has been mentioned in a number [of] hadith attributed to the Prophet Mohammed. According to one tradition, he liked the dish so much that he compared it to his wife Aisha, saying: ‘The superiority of 'Aisha over other women is like the superiority of Tharid to other meals.’” (http://en.wikipedia.org/wiki/Tharid)

18 Sentence (6c) is not provided in Moutaouakil 1989, and I have constructed in on the model of others. But its status is uncertain. Moutaouakil (1989.62) writes: PØ-placement [= “Top/Foc Position” or preverbal position] of a Focus is also sensitive to the syntactic function assigned to the term in question. Thus, if an 8 SYNTAX & SEMANTICS

[Zayd-NOM wrote book-ACC] ‘Zayd wrote a book’

And (Ouhalla 1994a.66-67):19

(7) (a) [when return Zayd-NOM] ‘When did Zayd return?’

(b) ‘Aad-a Zayd-un LAYL-AN [return-3MSG Zayd-NOM night-ACC] ‘Zayd returned during THE NIGHT’

In sentences on the model of (5b) and (6b) “the Focus constituent is indicated solely by means of the tonic accent.”20 The notation implies a contrast

Object is assigned Contrastive Focus, it can occupy PØ quite unproblematically; a Subject, however, cannot occupy this position. He follows with this example: (i) [Zayd-NOM came not ‘Amr-NOM] ‘It was Zayd that came, not ‘Amr’ We return to the difference below.

19 Ouhalla does not provide the MSA question ‘When did Zayd return?’, which is taken from Moutaouakil 1984.123, hence the difference in transcriptions.

20 In Mouataouakil 1984, this is “le constituent focalisé est indiqué, uniquement, par l’accent tonique” (123). Ouhalla (1997b.11) says “... f-phrases are written in capital letters, following the tradition. In pronunciation, f-phrases are marked by a special pitch accent (also called focal stress), known in the Arabic tradition as l-nabr.” Sometimes, confusingly, Ouhalla (1997b) will use uppercase in the English gloss, but there will be nothing in the MSA example (27): (i) [Q Zayd-nom at-you or Laylaa] ‘Is it ZAYD who is at your place or LAYLAA?’

Moutaouakil (1989.21) uses italics to indicate “tonic accent.” However, in Moutaouakil FOCUS: Nowhere 9 between sentences (5b) & (6b) and (8) (Moutaouakil 1989.6):

(8) [met ‘Amr-NOM -ACC] ‘‘Amr me in which the italics are absent, and supposedly the “tonic accent” as well.21 Sentences without the tonic accent are responses to questions such as ‘What happened?’, while those with tonic accent are not (Moutaouakil 1984.128):

‘What happened?’

(b) Raja‘a22 Zayd-un mina s-safari [est-revenu Zayd-NOM de voyage] ‘Zayd returned from his trip’

(d) % [a-voyagé Zayd-NOM hier] ‘Zayd left yesterday’

Sentence (9b), but not (9c) nor (9d), is a response to (9a). Sentences (9c) and (9d) are acceptable MSA utterances, but as responses to wh-questions, ‘When did Zayd leave?’ and ‘Who left yesterday?’

1989, Chapter 2 on Focus, “the italicized constituents ... receive the pragmatic function Topic” (69). And sometimes the presence of FOCUS is not noted at all (80), e.g. (ii) [has-seen ‘Amr-NOM Zayd-ACC] ‘‘Amr saw Zayd’ in response to ‘Who has seen Zayd?’

21 I have not been able to find in the literature a statement about any default placement of accent in sentences like (8).

22 The example from Moutaouakil 1984.128 has this as rajaca. Moutaouakil 1989.25 has the same example with raja‘a. I have chosen the later transcription. 10 SYNTAX & SEMANTICS

In response to wh-questions in (5a), (6a), and (7a), the answering content must be located in the position in which it would appear in the neutral expression, with the exception that it now carries focal stress. Thus, although (5c) and (6c) and (7c) have focal stress, they are not acceptable answers to their respective questions.24 Moutaouakil (1984, 1989) describes them as

23 This is a bit of an extreme conclusion and not one justified by evidence in the literature. As in Modern Greek above, and elsewhere, these are probably implementations of informational FOCUS, in which the entire utterance shares equally in the content of FOCUS.

24 There is a problem at this point. Moutaouakil (1989.22) is very explicit that the sentence- initial FOCUS is not a possible answer to a wh-question: “It is not possible for such questions to be answered by ... sentences in which the Focused constituent has been placed in preverbal position.” Some of the motivation for this conclusion is the belief that sentence-initial FOCUS is “contrastive” and, hence, inappropriate to respond to wh-questions. Ouhalla (1999.340), however, points out that “The term ‘contrastive focus’ is itself confusing and in some instances perhaps outright inaccurate. Not all f-phrases [FOCUSES, PWD] necessarily have a ‘contrastive reading’ ... [A] ‘contrastive reading’ is actually a function of the negative continuation rather than of the f-phrase ....” Although Ouhalla is, at this point discussing an example from Moroccan Arabic, it seems clear that he is exending the judgment to Modern Standard Arabic. However, in five papers on FOCUS in MSA (Ouhalla 1993, 1994a. 1994b, 1997b, 1999), Ouhalla himself fails to explain how to answer a wh-question! The problem is that Bakir (1980.21-22) provides a wh-question (i)

[book-ACC bought Muhammed-NOM] ‘Muhammed bought a book’ using preverbal sentence-initial FOCUS, precisely the form Moutaouakil has proscribed from this context. The so-called “contrastive responses” (Bakir 1980.23) are found when the (non- S) constituent is postverbal. Sentence (iii) contradicts and corrects sentence (iv): (iii) [bought book-ACC muhammed-NOM] ‘Mohammed bought a book’ (iv) [bought Mohammed-NOM watch-ACC] ‘Muhammed bought a watch’

Nevertheless, Bakir (9180.52) later adds that “we may find a sentence like ...[(ii)] with the initial constituent receiving contrastive focus rather than providing new information, depending on the context in which it is uttered....” This is basically what Ouhalla has said. Yet Bakir contends that a sentence-initial content answers a wh-question, while Moutaouakil counters that a sentence-initial FOCUS will not do. Only an in situ constituent FOCUS: Nowhere 11 expressions of “focus de contraste” (Moutaouakil 1984.120 et passim) or “Contrastive Focus”: “... the constituent with the pragmatic function of ‘ ’ (equivalent to ‘Contrastive Focus’) was seen generally to occupy initial position” (Moutaouakil 1989.1). The supposition of ‘contrast’ is based on utterances such as these (Ouhalla 1997b.18, 20):25

can answer a wh-question. Obviously, Moutaouakil and Bakir cannot both be correct. Moutaouakil (1984, 1989) does not cite Bakir’s (1980) work. Focal stress plays a large role in others’ (not Bakir’s) description of MSA FOCUS. Bakir mentions stress only three times in 185 pages. In a footnote (Bakir 1980.52), Bakir implicates stress without further comment: A sentence like (i) will show one constituent under contrastive focus: (i)

[Mohammed-NOM left to DEF-Iraq] ‘Mohammed left to Iraq’ Bakir (1980.130) notes, ... the NP receives a primary stress ... In contrast to this, a sentence where the NP to the left is a ‘topic’ denoting something about which something is said ... will have the same surface structure, but will have no primary stress assigned to the ‘topic’ NP, and may have a pause separating the ‘topic’ from the ‘comment’. Certainly arguments for distinctive features based on phonology carry little weight in the discussion of the form of Arabic we are concerned with here, since this form of Arabic is mainly used in writing. It is clear that the biases underlying Bakir’s work make it of little relevance to a discussion of MSA FOCUS. I will assume that the other literature is more accurate, and I will follow it.

25 Recall Moutaouakil’s comments (in the footnote to [6c] above). Sentence (10) should fail in MSA, but Ouhalla (1997b.28) asserts, “virtually any [What can’t?, PWD] category can be (focused) and focused moved [i.e., to Moutaouakil’s “PØ position”].” Something is not right. If sentences like (6c) and (10) are not possible, it would be such a large hole in the morphosyntax, someone would certainly have commented on it at length. 12 SYNTAX & SEMANTICS

(10) [Zaynab-NOM wrote3FS the-novel-ACC not Laylaa] ‘It was ZAYNAB who wrote the poem [sic] not LAYLAA’

(11) LAYL-an wasalat Zaynab-u laa NAHAAR-an [night-ACC arrive3FS Zaynab-NOM not day-ACC] ‘Zaynab arrived during the NIGHT not during the DAY’

(12) [read3FS Zaynab-NOM the-poem-ACC not wrote3FS-it] ‘Zaynab READ the poem. She did not WRITE it’

(13) ZAYNAB-u fii l-bayt-i laa LAYLAA [Zaynab-NOM in the-house-GEN not Laylaa] ‘It is ZAYNAB who is in the house not LAYLAA’

and (Moutaouakil 1989.23):

(14) (a) * [met Zayd-NOM Halid-ACC not ‘Amr-NOM] ‘It was Zayd who met Halid, not ‘Amr’

(b) [ate-1S bread-ACC not meat-ACC] ‘I ate bread not meat’

It is the fact that one can have an ‘X not Y’ with the sentence-initial FOCUS in (10) - (13) but not in the in situ expression in (14) that suggests that the former is ‘contrastive’.26 The conclusion that sentence-initial FOCUS is ‘contrastive’ and the in situ FOCUS is not must come with at least two caveats. First, there are these comments from Ouhalla (1997b.21-22):

... it is conceivable that the contrastive reading arises only in the presence of a negative/adversative continuation included in the senence (and introduced with the adversative/negative particle laa) or understood in the context. In the absence of the continuation, preposed f-phrases [i.e. a constituent bearing focal stress, PWD] have precisely the reading conveyed by the word namely

26 The in situ expression of FOCUS is commonly called “le Focus de nouveau” (Moutaouakil 1984.121 et passim), “New Focus” (Moutaouakil 1989.17 et passim), or “new information focus” (Ouhalla 1997b.11). FOCUS: Nowhere 13

specification (focus).27

Ouhalla (1997b.22) goes on to say “Assuming this conclusion to be on the right track ...,” but in the remaining 22 pages, there is no discussion of examples that might support the assumption. Nevertheless .... Second, the sense of ‘contrastive’ can be evoked — in the proper context — even for in situ FOCUS. Consider these two utterances (Ouhalla 1993.287):28

(15) (a) [NEG write-3F Zaynab-NOM novel-ACC] ‘Zaynab has not written a novel’

(b) [NEG novel-ACC write-3F Zaynab-NOM but

poem-ACC] ‘It was not a novel that Zaynab has written but a poem’

Sentence (15b) would seem to exemplify ‘contrastive’ FOCUS in the manner of (10) - (13), the difference being that in (15b) the context is negative, not positive. Now let us add (15c) (Ouhalla 1993.287):

(15) (c) [NEG write-3F Zaynab-NOM novel-ACC but qasiidat-an poem-ACC] ‘It is not a novel that Zaynab has written but a poem’

Ouhalla (1993.277) writes “As shown by the English translation and the

27 Moutaouakil 1984.119 has the term and in Moutaouakil 1989.18, it is

28 Given the gloss ‘It was not a novel ...’ in (15b) and given the preverbal position of riwaat- an, it is almost certain that riwaat-an is bearing focal stress in (15b), although Ouhalla does not note it here with uppercase as he does in (15c). In Ouhalla 1997b.39, there is a more consisent use of uppercase: (i) [not write-3FS Zaynab-NOM the-poem-ACC but read-3FS-it] ‘Zaynab did not WRITE the poem. She READ it’

The EVENT is focused in (i) and not the Patient, but otherwise (i) is parallel to (15c). Notice also that is uppercase in (i), but qasiidat in (15c) is not. 14 SYNTAX & SEMANTICS adversative continuation ... maa ... negates and contrastively focuses ... the object ....” I have purposely deleted some of the quotation. Ouhalla is actually writing about (15b). The portions deleted from the quotation are simply pieces of information that identify (15b) by its form. What remains of Ouhalla’s description is, of course, as applicable to (15c) as it is to (15b). Both are ‘contrastive’, and ‘contrastive’ cannot then be what distinguishes sentence- initial ex situ FOCUS from in situ FOCUS.29 But why would in situ FOCUS encourage contrastive senses in the negative, and not in the positive? The apparent answer lies in the relative contextless nature of VSO utterances minus focal stress that we witnessed in (9). In the positive, that contextless nature of VSO utterances minus focal stress persists, but when negative, the fact of negation imputes a context. Why say — out of the blue — that Zaynab did not write a novel unless something else happened? If Zaynab did not write a NOVEL, she may have written something else, and hence (15a) allows the FOCUS of (15c), which has the potential of being contrastive. The FOCUS of VSO without focal stress — outside of some determining context such as a negative maa — is then not Focus de nouveau, but a FOCUS without a limiting context, without a constraining locus.30 The content is presented as such, on its own, and without

29 The morphosyntax of word order, other than in situ, used to express FOCUS has recently come to be called ex situ.For example Aboh, Hartmann & Zimmermann 2007.5 et passim.

30 Note further in regard to Focus de nouveau that a definite Patient is possible using VSO morphosyntax (Mohammad 2000.9), e.g., l-walad-a ‘the boy’: (i) FOCUS: Nowhere 15 relation to anything else. FOCUS is distributed across the proposition, and that is what makes (14) fail. The contrasts implied by ‘Amrun and are not supported by the preceding assertion and its context. The environments of ‘Amrun and in (14) are not rich enough to make ‘contrast’ plausible.31 But on the other hand, why should sentence-initial position not be the

31 Another pattern that supports the basically contextless nature of VSO utterances without focal stress is the following. Sentence (i) is appropriately answered by (ii) and not (iii) (Moutaouakil 1984.143): (i) 16 SYNTAX & SEMANTICS morphosyntax used in answering wh-questions?32 The explanation has to lie somehow in the nature of the semantic contrast between ex situ FOCUS and in situ FOCUS. In situ FOCUS seems to operate on a constituent which stands alone against an empty background. There are no others with which might compete for the orientation that is FOCUS. In contrast, the constituent which carries ex situ FOCUS exists in a contextual matrix that is missing from in situ FOCUS. There are others about. To make the term of the ex situ FOCUS notable against its background, a stronger touch is required. In situ FOCUS needs a mere gesture, while the ex situ FOCUS demands a stronger touch. Because the in situ FOCUS exists in isolation, it is the appropriate FOCUS to respond to wh- questions. No candidate answers exist, and the contextless character of in situ FOCUS matches this context. Ex situ FOCUS is now inappropriate to answer wh-questions precisely because its does designate one among others ... which is not appropriate to wh-questions, which use the VSO syntax and bear no focal stress. Ex situ FOCUS is then appropriate to uses where a contrast is to be expressed (as we have seen), while in situ FOCUS requires the facilitation of negation to suggest even the possibility of contrast. The ex situ FOCUS does not, however, “mean contrastive FOCUS”. It may be used in such contexts, but it is also used in places which are not contrastive, but as long as they still reference a context of alternatives. As an example, MSA has two ways of composing yes-no questions. One queries the validity of the proposition as a whole, and another that queries the proposition with respect to one of its components (Ouhalla1997b.26):

(16) (a) [Q wrote-3FS Zaynab-NOM the-poem-ACC] ‘Did Zaynab write the poem?’

(b) [Q wrote-3FS Zaynab-NOM the-poem-ACC] ‘Did Zaynab WRITE the poem?’

The contrast in glosses indicates the difference between the two questions. It is the latter that selects one component as the point of interest — the pivot — in the query: ‘Is it the poem ... or something else?’ Hence the continuations of

32 This is what I had in mind in 4.1.1 in saying that MSA challenges our conceptualization of FOCUS. FOCUS: Nowhere 17

(17) (Ouhalla1997b.26):33

(17) (a) [Q wrote-3FS Zaynab-NOM the-poem-ACC or ? read-3FS-it] ‘Did Zaynab WRITE the poem or READ it?’

(b) [Q Zaynab-NOM wrote-3FS the-poem-ACC or ? Laylaa] ‘Did ZAYNAB write the poem or LAYLAA?’

Two observations.34 First, the term selected for FOCUS is the ex situ sort, and second, the context which is a part of ex situ FOCUS is explicit in the

33 Again Ouhalla indicates the presence of focal stress only in the uppercase of the English gloss and not in the transcription of the Arabic. Moutaouakil (1989.25) confirms the presence of focal stress with his use of italics: (i)

34 Actually three. The yes-no questions with hal, as one might expect, do not permit this morphosyntax (Moutaouakil 1989.37): (i) [Q Zayd-ACC met-2s] ‘Was it Zayd you met?’ and they admit no instance of focal stress; “ ... hal cannot appear in a sentence one of whose constituents has been assigned focus” (Moutaouakil 1989.27):

(ii) * [Q will.travel-2S tomorrow] ‘Are you setting out tomorrow?’ Unfortunately, neither Ouhalla nor Moutaouakil tells us how to answer these quesions. 18 SYNTAX & SEMANTICS continuation ‘or ...’ Although Moutaouakil (1989.67) contends “It is a Contrastive Focus since the speaker already posesses information which he wishes to have confirmed or refuted,” it is, as he actually suggests, an affirmation or denial. Not a contrast. If ex situ FOCUS is not suited as a response to a wh-question because of its greater integration with a context, it may nevertheless answer a question if the question itself is based on a context, and is not out of the blue. Consider a ‘Which ...?’ question. Here the question is not without context. There is a sense that the candidates are known, and what is wanted is a specification of the choice. It is not absolute ignorance that is assuaged. It is a contextualized ignorance. In this, a sentence-initial FOCUS might serve. Because in situ FOCUS is more lax, less intense than ex situ FOCUS, the former will permit more than one occurrence of it in a single proposition, i.e., more than one focal stress so that more than one constituent in the utterance may have focal stress as in (18a) and (18b), or be entirely focused as in (18c) (Moutaouakil 1989.47 & 1984.128):35

(18) (a) [met Zayd-NOM Halid-ACC ‘Zayd met Halid’

(b) [informed Zayd-NOM Halid-ACC with-success-3SG] ‘Zayd informed Halid of his success’

35 We know that there are two focal stresses in (18a) and three in (18b) because they are the respective answers to (i) and (ii): (i) Man man? [who met whom] ‘Who met whom?’ (ii) Man man FOCUS: Nowhere 19

And because ex situ FOCUS is more concentrated and pointed, it will permit but a single expression of FOCUS (Moutaouakil 1984.154):

(19) [Zayd-acc le-livre-acc ai-donné-je] ‘It’s Zayd that I gave the book’

FOCUS in MSA appears to be structured semantically along two dimensions. The first is that of GLOBAL (informational) and PARTICULAR (identificational). The grammatical means of the opposition is absence of focal stress for GLOBAL FOCUS and the use of focal stress for a PARTICULAR FOCUS.36 The second dimension is that of INTENSE versus LAX. When the FOCUS is PARTICULAR, INTENSE FOCUS is expressed by sentence-initial position. When the PARTICULAR FOCUS is LAX, the neutral VSO grammar is used. The same opposition between INTENSE versus LAX can be implemented when the FOCUS is GLOBAL (Moutaouakil 1989.29)

(20) (a) [wrote halid-NOM book-ACC] ‘Halid wrote a book’

(b) [ wrote halid-NOM book-ACC] ‘I assure you that Halid wrote a book’

Neither has focal stress. The first will answer ‘What happened?’ The second will not. This four-way organization of MSA FOCUS might be summarized as in Figure 1. This figure looks forward to further discussion below of the semantic typology of FOCUS. The implementation of INTENSE and LAX FOCUS in the company of GLOBAL FOCUS will reveal itself as variation in the force of ASSERTION. The two dimensions, which appear orthogonal in Figure 1, may align themselves congruently so that GLOBAL FOCUS coincides with LAX FOCUS while PARTICULAR FOCUS aligns with INTENSE FOCUS, and the whole FOCUS semantics plays out as variation in ASSERTION.

36 I assume that (18c) is not global FOCUS, but what happens when the proposition is exhaustively populated by particular FOCUS. 20 SYNTAX & SEMANTICS

INTENSE LAX FOCUS FOCUS

GLOBAL Morphology with VSO without focal stress VSO without focal stress FOCUS

PARTICULAR Ex situ — with focal stress In situ — VSO with focal stress FOCUS

Figure 1: An Organization of FOCUS in Modern Standard Arabic.

If the ex situ position of FOCUS is correctly characterized by its being embedded in a context, it is interesting to find that the ex situ position has another use without focal stress (Ouhalla 1997b.12):

(21) (a) [the-novel-NOM wrote3FS-it Zaynab-NOM] ‘As for the novel, Zaynab wrote it’

(b) [Laylaa loved-3MS-her Qays-NOM] ‘As for Laylaa, Qays loved her’

Ouhalla (1997b.12) describes the morphosyntax: 37

First, LD-phrases [which express content that is TOPIC, PWD] express old information, that is, they refer to individuals which are familiar to the conversants and who may already be the topic of discussion. This is not the case with f-phrases [i.e., FOCUS expressed preverbally, PWD] ... Second, LD-phrases do not bear l-nabr,

37 Moutaouakil (1989. 69ff) depends on an a priori definition of TOPIC as “the entity ‘about’ which the predication predicates something in the given setting.” Without establishing how TOPIC is manifest grammatically, TOPIC is potentially anywhere in a verb-initial proposition not populated by FOCUS or the EVENT. For example, any term following an initial verb, if it does not bear focal stress, may by this calculation be TOPIC. Which, if any, are TOPIC goes without morphosyntactic marking. And if VSO utterances are correctly taken to be contextually neutral — as they have been alleged to be above — and to be the only answer to ‘What happened?’, then any attribution of the semantics of TOPIC to them seems truly arbitrary. Moutaouakil (1989.102) defines a “theme [which] specifies the universe of discourse with respect to which the subsequent predication is presented as relevant.” It is his “theme” that matches the morphosyntax of others’ TOPIC in MSA. FOCUS: Nowhere 21

contrary to f-phrases which invariably do ... Third, LD-phrases are separated from the rest of the sentence by an intonational break (or a pause), represented ... with the comma. Preposed f-phrases, in contrast, are not separated from the rest of the sentence by a similar intonational break. Fourth, LD-phrases are related to a pronoun located inside the sentence ... whereas the preposed f-phrases are not ....

The similarities with Modern Greek (Chapter 9) are obvious. Both use a sentence-initial position before the verb in combination with a focal stress to implement INTENSE FOCUS. Both use that same position without the focal stress for TOPIC. Both will separate TOPIC from what follows with a pause (although Modern Greek has some variation on this). Both represent the TOPIC — if it is an actant in the proposition — with a pronoun (although Modern Greek has some variation on this). And in the presence of a preverbal INTENSE, PARTICULAR FOCUS, it is the TOPIC that precedes in MSA, as in Modern Greek (Ouhalla 1994a.71):

(22) (a) [Zayd-NOM novel-ACC give-1SG-him] ‘Zayd, it was A NOVEL I gave him’

(a) [novel-ACC Zayd-NOM give-1SG-him] ‘Zayd, it was A NOVEL I gave him’

3.2 Modern Greek In addition to being VSO languages, Modern Standard Arabic and Modern Greek seem to share one additional property. They both have an in situ expression of FOCUS. Tsimpli (1995.188) assigns Modern Greek to the group of languages “where the focus phrase is [can be, PWD] in situ.” It is clear that what appears to be an in situ focused constituent can respond to a question, in two ways (Philippaki-Warburton 1985.122-123):39

38 Sentence (22b) would be incorrect even if it were *

39 As is frequent in the literature, there is a disagreement about what can be said. Although Philippaki-Warburton affirms (23b) and (23c) as acceptable answers to the question in (23a), Keller & Alexopoulou (2001.308) report this failed exchange:

(i) (a) Pjon apelis-e i Maria? [who fired-3s the-nom Mary] ‘Whom did Mary fire?’

(b) ??Apelis-e i Maria TO JANI 22 SYNTAX & SEMANTICS

(23) (a) Pjon filis-e o Janis? [whom kiss-3s the-nom John] ‘Whom did John kiss?’

(b) O Janis filis-e / TI MARIA [the-nom John kiss-3s the-acc Mary] ‘John kissed Mary’

Philippaki-Warburton (1985) does not, that I can find, explain wthat the solidus is used for in the answers. It is potentially important to know because there exists this question and answer pair (121):

(24) (a) Ti ekan-e o Janis [what did-he the-nom John] ‘What did John do?’

(b) O Janis / filise TI MARIA [the-nom kiss-he the-acc Mary] ‘John kissed Mary’

The formal difference between (24b) and (23b) lies solely in the position of the solidus. Semantically, the difference between them is what is considered to be TOPIC. In (24b) (122):

The subject NP ... is both given and theme [i.e., topic, pwd], and the VP is rheme with focus TI MARIA ... The common features of the answers ... [(24b) and (23b)] are that the subject NP is the theme [i.e., TOPIC, PWD] and TI MARIA is the focus, while the verb is within the new (Rheme) part of ... [(24b)] but within the given in [(23b)] ....

The description here derives from two mistaken presumptions about morphosyntax. The first is that morphosyntax is precise and the boundaries are neat. This belief is one of those that has prompted the postulation of

[fired-3s the-nom Mary the-acc John] ‘Mary fired John’ FOCUS: Nowhere 23 hierarchical structure. The evidence here from Modern Greek is that that is not so. Morphosyntax can be vague. There is no a priori requirement that we know from the expression where TOPIC and FOCUS begin and end. To be effective, grammar must alert us to the presence of a meaning. We must know which meanings are present, but not necessarily where the exact beginnings and ends are. If we know (or can determine) which meanings are, that understanding may come in part from the context. The sentence (23b)/(24b) resolves itself into either (23b) or (24b) depending on having heard the preceding question. The second mistaken presumption is this. When one speaks of Topic + Comment, Theme + Rheme, or the equivalent, the two must exhaust the utterance with no remainder.40 Historically, this has been a common belief in linguistics. But again, there is no a priori reason that TOPIC and FOCUS cannot both be present along with content that expresses neither. Modern Greek appears not to distinguish formally between “new” and “given”. That contrast is probably not there, and Modern Greek has a more lax sense of what counts as a continuation. Ti Maria satisfies the wh-request in both (23b) and in (24b), while O Janis filise in (23b) and O Janis in (24b) connect the answers to the context. While both (23b), (23c), and (24b) appear to be in situ expressions (i.e., the O is final), notice that (23b) and (24b) seem to be SVO. It may be that these three are analogous to (15c), repeated along with its question as (25):

(25) (a) Pjos efage tin turta? [who ate-3s the-acc cake] ‘Who ate the cake?’

(b) Tin efage tin turta, o Janis41

40 Holton, Mackridge & Philippaki-Warburton (1997.430) Use “topic” and “comment” when describing the exhaustive partition of the morphosyntax, and “topic” and “focus” when there is some remainder: FOCUS “carries more prominent stress and is part [Emphasis mine, PWD] of the comment ....”

41 In their discussion of VOS grammar, Georgiafentis & Sfakianaki (2004.951) make these remarks about the presence and absence of clitics like tin in (24b): In the VOS order the clause final subject scores 18%, whereas in the clVOS order it scores 14%. A One-Way Analysis of Variance shows that this difference is not statistically significant (P = 0.651). Thus, the DP-subject [i.e., the “S”, PWD] receives focus with almost the same frequency in both cases (i.e. with or without clitic). This suggests that we can probably treat the two as the same here. Whatever their differences, they are both sufficiently TOPIC for the use they fulfill here. 24 SYNTAX & SEMANTICS

[cl-acc ate-3s the-acc cake the-nom John] ‘John ate the cake’

The sameness across them is that the initial portions of (23b), (23c), (24b), and (25b) occur sentence-initially, without strong stress, and act as TOPIC continuing from their respective questions.42 Keller & Alexopoulou (2001.306) add this example:

(26) (a) Pjos apelis-e ti Maria? [who fired-3s the-acc Mary] ‘Who fired Mary?’

(b) Tin apelis-e O JANIS ti Maria [her fired-3s the-nom John the-acc Mary] ‘John fired Mary’

Here, in (26b) Tin apelise, the sentence-initial pronoun Patient and the EVENT act as TOPIC to the following FOCUS, o Janis. The material completing the utterance, ti Maria, is probably comparable to o Janis in (27) and (28) (Georgiafentis 2001.138-139):

(27) Efage tin turta(,) o Janis [ate-3s the-acc cake the-nom John] ‘John ate the cake’

In sentence (27), o Janis “is unstressed, constitutes background information and is right dislocated,” i.e., is preceded by a pause (Georgiafentis

42 Philippaki-Warburton (1985.123) reports that O Janis filise is “much more natural, in fact, to answer” (23a), and Filise o Janis is “not the best answer.” This makes sense since the former has O Janis in the identified sentence-initial TOPIC position, and the latter uses a verb initial “neutral” order. Philippaki-Wurburton does not report any oddity in the strategy of responding to a wh- questioning of the O; (23b) is “natural.” However, when the S is wh-questioned, Georgiafentis & Sfakianaki (2004.952 et passim) “the VOS order is dispreferred for subject focussing.” Their subjects “opted for the OclVS or the SVO order for subject focussing.” This is the same as Philippaki-Warburton’s conclusion about O’s. In both cases, place the focused term last with focal stress. Precede it with a TOPIC and identify the content as TOPIC by placing unstressed material (the O or the S, depending on what is being questioned) sentence-initially before the EVENT/Verb. The alternative of placing the material after the EVENT/Verb invokes the “neutral” grammar and does not convey the meaning of TOPIC. Hence, it does not work (very well). FOCUS: Nowhere 25

2001.138).43 Some “constituent other than ... [o Janis] receives the main accent” in (27) (Georgiafentis 2001.144). Georgiafentis (2005.165) has this similarly formed utterance:

(28) O Janis tin efaje tin turta [the-nom John it ate-3s the-acc cake] ‘What John did was eat the cake’

O Janis appears sentence-initially without focal stress functioning as TOPIC. The preverbal pronoun identifies ‘the cake’, thus making it TOPIC as well. Tin turta appearing sentence-finally cannot then be FOCUS and carry focal stress. This leaves only efaje to function as FOCUS and it “will receive the main accent.” The English equivalent is a syntactic pseudocleft. As in (25b), (26b), and (27) the final term in (28) simply reprises content and falls into a Dead Zone (is right dislocated, is a Tail, etc.). The content of (28) is like (23b), (23c), (24b), (25b), (26b) in being organized as TOPIC + FOCUS. Haidou (2000.167) has these:

(29) Efage ti supa, o Janis [ate-3s the-acc soup the-nom John] ‘John ate the soup’

(30) Efage ti supa o Janis [ate-3s the-acc soup the-nom John] ‘John ate the soup’

“with two possible intonations. Either the subject is preceded by a pause or not.” Sentence (29) has “comma intonation,” and (30) has “flat intonation.” On the contrast between (29) and (30), Joseph & Philippaki-Warburton (1987.97) write, “Noncontradictory sentence emphasis [stress, PWD] can only be expressed by placing emphatic stress on the item which, under normal circumstances carries the center of the intonation. This is usually the last constituent of the sentence. However, there are situations where the final position is occupied by a right-dislocated topic [sic], which cannot carry emphasis. Therefore, in this case, emphasis is carried by the constituent

43 The comma is in parentheses suggesting that the pause may be present or not, as a free variation. In (25b), the pause is more assuredly present, and the subject is “preceded by comma intonation” (Haidu 2000.180). The material following the pause (comma), or following an in situ FOCUS, as in (26b), i.e., ti Maria, are likely candidates for the Dead Zone described for Warao, section 2.8 in Chapter 8. 26 SYNTAX & SEMANTICS before the final topic [sic].” “The last constituent of the sentence” describes Georgiafentis’ “default sentence-final stress.” Joseph & Philippaki- Warburton then discuss this example:

(31) Tha páo na dhó to Jáni ávrio [fut go-2s prt see-1s the-acc John tomorrow] ‘I will go see John tomorrow’

They comment, “In ... [(31)] the last word, the adverb ávrio, carries emphatic stress. In a normal, nonemphatic rendering ávrio would also carry the main stress of the sentence. Example ... [(31)] may be interpreted either as giving emphasis to the whole sentence or to the item ávrio alone. In the former case, however, the whole sentence is uttered on a slightly higher stress/pitch level and the tempo is faster and steadier in reaching the final peak. On the other hand, if ávrio is intended only as the only emphatic element, the first part of the sentence is delivered on a lower stress/pitch level and there is more of an abrupt rise in stress on ávrio.” The “lower stress/pitch level” is what combines Tha páo na dhó to Jáni together and marks it as TOPIC to the FOCUS ávrio. This appears to describe the phonological contrast between (29) and (30) above. If we collect all the purported examples of FOCUS in situ and arrange

TOPIC Vague FOCUS DEAD ZONE

O Janis filis-e TI MARIA Filis-e o Janis TI MARIA O Janis filise TI MARIA Tin efage tin turta o Janis Tin apelis-e O JANIS ti Maria O Janis tin efaje tin turta Tha páo na dhó to Jáni ávrio

Figure 2: TOPIC & FOCUS expressions. them as in Figure 2, we discover that they can be reasonably interpreted as variations of a morphosyntax that orders TOPIC first and FOCUS second. Copular sentences reduce the TOPIC + FOCUS syntax to a bipartite one (Joseph FOCUS: Nowhere 27

& PhilippakiWarburton 1987.34):44

(32) O Rigan próedhros [the-nom Reagan president] ‘Reagan is president’

“The usual order in such sentences has the subject nominal occurring before the predicate nominal.” The picture of Modern Greek in situ FOCUS that is emerging is that the “neutral” VSO order fails to provide adequate responses to wh-questions. If the sentence-initial FOCUS is not used, then one of the TOPIC + FOCUS expressions in Figure 2 must be used. In response to (33a), the VSO syntax of (33b) fails, as does (34b) fail to answer (34a):45

(33) (a) Pjos apelis-e ti Maria? [who fired-3s the-acc Mary] ‘Who fired Mary?’

(b) ??Apelis-e O JANIS ti Maria [fired-3s the-nom John the-acc Mary] ‘John fired mary’

(34) (a) Pjon apelis-e i Maria? [whom fired-3s the-nom Mary] ‘Who did Mary fire?’

(b) ??Apelis-e i Maria TO JANI [fired-3s the-nom Mary the-acc John] ‘John fired mary’

44 “In main clauses, under certain conditions, specifically when the copula is understood as present tense and third person, it can be omitted” (Joseph & Philippaki-Warburton 1987.34).

45 I did not find the information to solve this puzzle in the published literature. The speakers of Modern Greek whom I consulted commented on (33b) as follows, “I wouldn’t use it, but it makes sense.” Another ranks the response as “So and so,” and the third speaker thinks it is “quite strange.” The response in (34b) is “Strange,” and another commented “No. It is correct grammatically (and even semantically and pragmatically) but why start with the verb, which was just uttered in the question? ....” Speakling of an answer analogous to (34b), Philippaki- Warburton (1985.123) says that “... although ... [it] is a possible answer to ... [the question], it is not the best answer.” I am grateful to the three native speakers of Modern Greek who responded to an inquiry posted on the Funknet Listserve. 28 SYNTAX & SEMANTICS

However, those VSO sequences with their varying placements of focal stress, which failed in (33) and (34), do appear to be satisfactory answers to ‘What happened?’:46

(35) (a) Ti ejine? [what happened-3s] ‘What happened?’

(b) Apelise o Janis TI MARIA [fired-3s the-nom John the-acc Mary] ‘John fired Mary’

(d) APELIS-E o Janis ti Maria [fired-3s the-nom John the-acc Mary] ‘John fired Mary’

From this inquiry, it seems clear that VSO orders do not respond to wh- questions. Focal stress appears to function as it did above in (19a). It identifies the point of interest, either the entire utterance or some component within it. This leaves the TOPIC + FOCUS and the sentence-initial FOCUS as the devices to respond to wh-questions. Given the TOPIC + FOCUS sentence in (36a) and the sentence-initial FOCUS of (36b),

(36) (a) Aghoras-e to aftokinito O JANIS [bought-3s the-acc car the-nom John] ‘John bought the car’

(b) O JANIS aghoras-e to aftokinito [the-acc John bought-3s the-acc car] ‘John bought the car’

46 One speaker says that all “are acceptable answers to the question.” The other two speakers concur with the reservation that (35c) seems “strange” or “rather not” an answer. It is interesting that one speaker ranks (35b) as “fine,” (35c) as “so and so,” and (35d) as “strange,” as focal stress moves further from the default final position. FOCUS: Nowhere 29

Alexiadou & Anagnostopolou (2000.176) remark:47

... the preverbal foci are not equivalent to postverbal ones: while in ... [(36b)] the subject receives a contrastive of narrow-focus reading, ‘it was John and not Paul’, in ... [(36a)] the subject simply receives a ‘new’information focus.

If all this is accurate, we end with a construction of FOCUS in Modern Greek as depicted schematically in Figure 3. There is a system of four seman-

INTENSE LAX FOCUS FOCUS

GLOBAL VSO with default stress I FOCUS VSO with focal stress II TOPIC plus FOCUS III PARTICULAR Sentence-initial FOCUS IV FOCUS

Figure 3: Grades of FOCUS in Modern Greek.

tic grades of FOCUS.48 Beginning at the upper right in the Figure 3, FOCUS becomes increasingly INTENSE as it simultaneously narrows its compass to a single term. As FOCUS becomes increasingly INTENSE, it becomes increasingly PARTICULAR. While Modern Greek and Modern Standard Arabic seem, on the surface, to be quite similar in the configuration of FOCUS, they end by being significantly different. Compare Figures 1 and 3. In situ FOCUS has a much

47 Citing Alexiadou 1999a.

48 There continues to exist some lack of clarity in the description of Modern Greek. Tsimpli (1998.203) remarks of an apparent VSO utterance:

(i) Sinandis-e TI MARIA (oxi tin Eleni) [met-3s the-acc Mary not the-acc Helen] ‘He met Mary, not Helen’ that “For many Greek speakers ... a contrastive reading can also be construed for in-situ focus phrases ... The implication is that the in-situ focus is ambiguous between a presentational and a contrastive reading whereas a preposed focus phrase cannot be construed as presentational.” What is not clear is the number of utterances lurking behing the transcription in (i). Recall from above that the default position for sentence stress is sentence-final, and it is apparently possible to invoke a phonological contrast between two degrees of stress in that position. 30 SYNTAX & SEMANTICS more minor presence in Modern Greek than in Modern Standard Arabic. Finally, Modern Greek adds support to our initial FOCUS in situ Conjecture that languages which have an in situ FOCUS will exhibit at least two degrees of focal INTENSITY, that the in situ FOCUS will be the weaker, and that ASSERTION will be absent from the constitution of FOCUS. The morphosyntax of Modern Greek betrays no involvement of ASSERTION. There is only (i) the neutral VSO word order, (ii) the same plus focal stress, (iii) TOPIC, and finally (iv) ex situ FOCUS itself.

3 In Situ Languages that are SVO Like the VSO in situ languages Modern Standard Arabic and Modern Greek, in situ languages that are SVO will also have an alternative morphosyntax of FOCUS which is more INTENSE than the in situ expression.

3.1 Northern Sotho49

Northern Sotho (Sesotho sa Leboa, also known as Sepedi after its standardized dialect) is one of the eleven official languages of the Republic of Africa.50 It is spoken in the northern provinces of South Africa by approximately 4,208,980 speakers .... (Zerbian 2006b.385)

Northern Sotho “displays [an] SVO basic word order” (Zerbian 2006b.235), and “The basic or dominant word-order in sentences with a neutral word order pattern is SVO” (Louwrens, Kosch & Kotzé 1995.39). These are typical Northern Sotho transitive utterances (Zerbian 2006b.388 & 2007b.325):51

(1) [CL1-old.man SC1 greet CL9.doctor] ‘The old man is greeting the doctor’

49 The bibliography that is the basis of this section is Louwrens 1987 & 1991, Louwrens, Kosch & Kotzé 1995, Lombard 1985, Steyn 1992, Steyn and Prinsloo 1995, van Wyk 1967, and Zerbian 2006a, 2006b, 2007a & 2007b.

50 The other ten are Afrikaans, English, Ndebele, Pedi, Swazi, Tsonga, Tswana, Venda, Xhosa, and Zulu.

51 “CL” stands for ‘Noun Class.” Northern Sotho has 18 noun classes of grammatical gender (Louwrens, Kosch & Kotzé 1995.18-20). “SC” stands for ‘Subject Concord.” “Essentially, the verb consists of a SAG [subject agreement] and a verb stem” (Louwrens, Kosch & Kotzé 1995.30). Zerbian writes the agreement as a separate element, not as an affix. Her transcription also uses underlining to note “underlying high tones” (Zerbian 2006b.388). I have omitted that. FOCUS: Nowhere 31

(2) [CL1-man SC1 greet-APPL CL2-woman today] ‘The man writes to (the) women today’

The “basic word order in double object constructions ... [has the] beneficiary preced[ing] ... the patient (or the animate object preced[ing] ... the inanimate object)” (Zerbian 2006b.395):

(3) O fa mo-kgalabje se-hla:re [SC1 give CL1-old.man CL7-medicine] ‘He gives the old man medicine’

“Objects and adverbials are questioned in situ, i.e. the position of the interrogative word in a question corresponds to its syntactic position in basic word order in a declarative sentence” (Zerbian 2006b.394, 396):

(4) O bona mang ka mehla? [SC2.SG see who always] ‘Who do you always see?’

(b) *O bona ka mehla mang? [SC2.SG see always who] ‘Who do you always see?’

(5) (a) O fa mang sehla:re? [SC1 give who CL7-medicine] ‘Who does he bring medicine to?’

(b) *O fa sehla:re mang? [SC1 give CL7-medicine who] ‘Who does he bring medicine to?’

(6) (a) O fa mo-kgalabje eng? [SC1 give CL1-old.man what] ‘What does he give the old man?’

(b) *O fa eng mo-kgalabje ? [SC1 give what CL1-old.man] ‘What does he give the old man?’ 32 SYNTAX & SEMANTICS

(7) (a) O bona ngaka ne:ng? [SC2.SG see CL9.doctor when] ‘When do you see the doctor?’

(b) Neng o bona ngaka? [when SC2.SG see CL9.doctor] ‘When do you see the doctor?’

Sentence (4b) fails because the wh-word questioning the O is not in the immediate postverbal position of the neutral order. Sentences (5b) and (6b) fail because the wh-words are not in the neutral positions illustrated in (3). Sentence (7b) fails because the wh-word questioning the adverb ‘when’ is not in its sentence final neutral position. Although we are not told in Zerbian 2006b how to answer questions (4) - (7), Zerbian 2007a and Zerbian 2007b illustrate some possible answers (Zerbian 2007a.61, 71):

(8) (a) O mémá má:ng? [SC2.SG invite who] ‘Who do you invite?’

(b) Ke mémé mohú:mi [SC1.SG invite CL1-rich] ‘I invite the rich man’

Sentence (8b) is also an acceptable response to ‘What do you do?’ and ‘What are you doing with the rich man?’ Sentence (8b) is, moreover, phonologically identical to the same utterance when it is not a response to any of these questions: “... in syntactically identical sentences there is no suprasegmental marking of focus in Northern Sotho” (Zerbian 2007a.73). A possible reaction to this discovery is to conclude that Northern Sotho does not signal FOCUS (and probably, hence, that it lacks FOCUS), but a more reasonable conclusion is that everything from the verb to the end of the utterance is equally FOCUS and that the specific term(s) that FOCUS augments — if any — is determined solely from the context.52

52 Zerbian (2007a.74-75) does not conclude that FOCUS is missing in these utterances. Instead, “From the study of African languages such as ... Northern Sotho (Bantu), evidence against the obligatoriness of marking focus in grammar emerges” (75). It is not really that FOCUS is not obligatorily marked, it is just that the morphophosyntax of in situ FOCUS is vague in that it is not explicit in selecting just one term for FOCUS. You have to know what you are talking about. FOCUS: Nowhere 33

The reason for asserting that FOCUS is present morphosyntactically exactly from the verb to the end of the utterance is that S’s do not (for the most part) participate in the in situ grammar of FOCUS, while everything following a sentence-initial S does. Agents differ from the other functions in having a focal expression that resorts to a distinct morphosyntax. Northern Sotho has a system of copulative expressions (Lowrens 1991.63-90) that distinguishes three kinds of ASSERTION: the identifying, the descriptive, and the associative.53 In the first, “the subject and the complement are identified with one another ... as being semantically equal” (Lowrens 1991.69). In the descriptive, “the complement describes the subject in terms of a particular feature” (Lowrens 1991.69), and in the associative, “the subject and the complement are associated with one another” (Lowrens 1991.70). The identifying copulative in (9) asserts an identity that is noun-like; the associate asserts in (10) an adjective-like property; and the associative in (11) effects possession (Lowrens 1991.64, 65, 70, 79):

(9) (a) Yó a hlôkôfét ego ke mmakgolo [ CL1 passed.away COP grandmother’ ‘The one who passed away is grandmother’

(b) Motho yô ke lehôdu [person this COP thief] ‘This person is a thief’

(10) (a) Moruti o bogale [teacher COP bad.tempered] ‘The teacher is bad tempered’

(b) Tau e bogale [lion COP vicious] ‘The lion is vicious’

(11) (a) Kgômo e na le namane [cow COP calf] ‘The cow has a calf’

(b) Moruti o na le mmôtôrô wô mohubêdu [teacher COP car red]

53 Lombard (1985.193-197) describes the same three-way distinction. 34 SYNTAX & SEMANTICS

‘The teacher has a red car’

The identifying copula is said to be “invariable” (Lowrens 1991.73) meaning that it does not reflect the varying noun classes to which the subject belongs. The other two copulas are “variable” (Lowrens 1991.74-75). Without providing a complete paradigm, Lombard (1985.193) records variation in the identifying copula by person and number:

(12) (a) Kè moruti [COP preacher] ‘I am a preacher’

(b) O moruti [COP teacher] ‘You are a teacher’

(d) Re [COP workers] ‘We are workers’

(e) Le baithuti [COP students] ‘You are students’

In replying to a wh-question, the identifying copula can suffice to provide an acceptable response, and the indentifying copula must be used when the Agent is questioned and when an Agent is the answer. Consider (Zerbian 2006b.397):

(13) (a) Ké mang (yo) a nyaka-ng nga:ka?54 [COP who RPRN.CL1 CL1 look.for-REL CL9.doctor] ‘Who is looking for a doctor?’

54 Zerbian 2006b.397 has nga:ka in this example, while the same question in Zerbian 2007b.326 has ngaka. FOCUS: Nowhere 35

(b) %Mang o nyaka nga:ka?55 [who SC1 look.for CL9.doctor] ‘Who is looking for a doctor?’

Sentence (13b) suggests that true in situ questions of Agents are not possible in Northern Sotho and that (14)

(14) o nyaka ngaka [SC1-old.man SC1 look.for CL9.doctor] ‘The old man is looking for a doctor’ is not an answer to (13a) (Zerbian 2007b.326), although it does answer ‘Who is the old man looking for?’ The appropriate answer to (13a) is (Zerbian 2007b.326):

(13) (c) Ké a nyaka-ng ngaka [COP SC1-old.man SC1 look.for-REL CL9.doctor] ‘The old man is looking for a doctor’

Morphosyntactic in situ questioning of the Agent exists, but they are not actually wh-questions (Zerbian 2006b.399):

(15) (a) o hlokomela ngaka [SC1-old.man SC1 look.after CL9.doctor] ‘The old man is looking after the doctor’

(b) Mang o hlokomela ngaka? [who SC1 look.for CL9.doctor] ‘Who is looking after the doctor?’

Sentence (15b) is an echo question, and with that meaning — a lesser question and a lesser FOCUS — it is possible to question the Agent without the morphology of ké ... (yo).56 The sense of ‘echo’ suggests a more LAX FOCUS for the in situ expression, with a more INTENSE FOCUS for the grammar of ké ... (yo).

55 The reason for placing a “%” before (13b), suggesting an inappropriate usage in place of out and out grammatical malformation, is explained in (15) below.

56 We are not told whether Mokgalabje o hlokomela má:ng? is also an echo question meaning ‘The old man is looking after who!?’ as well as the expected ‘Who is the old man looking after?’ And if not, whether there is such an echo question. 36 SYNTAX & SEMANTICS

The contrast between a LAX and an INTENSE FOCUS is confirmed by the occurrence of non-Agents initial with ké ... (yo) (Zerbian 2006b.398 & 2006a.372):57

(16) Ké mang o mo-kgalabje a mo nyaka-ng? [COP who RPRN.CL1 CL1-old.man SC1 OC1 look.for-REL] ‘It is who that the old man is looking for?’

(17) Ke kae (mo) go dula-ng di-nonyana? [COP where RPRN SM17 live-REL CL10-bird] ‘Where are there singing birds?’

(18) Ke neng mo go fihla-ng ba-na? [COP when RPRN SM17 arrive-REL CL2-child] ‘When will the child arrive?’

Actants and adverbs may be used in this way, but other than being told that the morphosyntax “is more marked pragmatically and hence more rarely used in Northern Sotho” and that the expressions “cannot be used interchangeably [with in situ] for questioning non-subjects” (Zerbian 2006b.399), there is no precise characterization of the contrast.58 We may expect that the answers to (16) - (18) would use the same grammar and parallel (13c), and not the in situ expressions, but we are not told that either. The copulative expressions ASSERT the term following the copula as the one that exemplifies the term that precedes as the (as yet) unidentified term. In (9a), Yóa hlôkôfét ego is ‘the one who passed away’ and ke mmakgolo names the one. When used to ask or to respond to questions, the phrase Yó a hlôkôfét ego would follow the ASSERTION, as in (13c) and (16) - (18). the FOCUS expressed is paradigmatic in that it is an exclusive, either-or choice. This relation contrasts with the syntagmatic FOCUS ngaka ‘a doctor’ of (14), as an answer to ‘Who is the old man looking for?’, in that ngaka ‘a doctor’ is the FOCUS without excluding expression of the non-FOCUS content, o nyaka ‘The old man is looking for ...’ There are notable parallels between the syntax and semantics of Modern

57 Zerbian 2006a does not write the high tone on ké. The mo that is optional in (17) is obligatory in (18).

58 “For the semantics of object clefts,” we are referred to Zerbian’s 2006, unpublished and inaccessible, Ph.D. dissertation, Expression of Information Structure in the Bantu Language Northern Sotho, at Humboldt-University, Berlin. FOCUS: Nowhere 37

Greek, Modern Standard Arabic, and Northern Sotho.59 The existence of two modes of FOCUS in Modern Greek and Modern Standard Arabic is repeated in Northern Sotho. But Northern Sotho differs from Modern Greek and Modern Standard Arabic in using a form of paradigmatic FOCUS (the grammar of the identifying copulative) to effect what Modern Greek and Modern Standard Arabic achieve by an ex situ FOCUS. Secondly, in Modern Greek and Modern Standard Arabic, in the absence of overt marks of FOCUS, preverbal, sentence-initial content is taken as more TOPIC-like, and the result is a morphosyntax that expresses TOPIC + FOCUS.60 The following Northern Sotho utterance illustrates an analogous usage (Zerbian 2006a.374):

59 This, despite the assignment of these languages to two distinct word order types: MG & MSA are VSO and Northern Sotho is SVO. Taking the typological measure of these languages using the order of V, S, and O will obscure the similarities they plainly exhibit.

60 Another indication that utterances are more or less TOPIC preverbal and then FOCUS, comes from the observation of verbal agreement with Patients. Patients, like Agents may appear preverbally, but when they do, there must be a verbal affix marking the gender class of the Patient (Louwrens, Kosch & Kotzé 1995.41-42): (i) [child dog SAG-OAG-bit] ‘As for the child, the dog bit him/her’ (ii) [dog child SAG-OAG-bit] ‘As for the dog, it bit the child’ It is not clear what the contrast between (i) and (ii) is (One one page, [i] has a comma after ngwana, and on the next, the same example omits the comma.). It is certain that “object nouns which are moved from the postverbal to the preverbal position also cannot be subjected to questioning.” In postverbal position the OAG affix is also possible, and in that use, the Patient also cannot be questioned (Louwrens, Kosch & Kotzé 1995.40): (iii) [dog SAG-bit child which] ‘Which child did the dog bite?’ (iv) [dog SAG-OAG-bit child which] ‘Which child did the dog bite?’ This suggests that the agreement affixes ground their referents in the context, recording the fact that they are established in the discourse, hence unquestionable in (iv) and TOPIC-like if preverbal. This further suggests that since Agents always are recorded by agreement, that they are always grounded or established in the discourse. The semantics of grounded/ established vs. ungrounded/unestablished is the basis for the binary appearance of Northern Sotho sentences. This effect of verbal affixes is not uncommon. Cf. the Rwandan examples (27) & (28) in Chapter 3. 38 SYNTAX & SEMANTICS

(19) Ba-sádi bá a bá bé:tha [CL2-woman SM2-A-OM2 beat] ‘The womeni, theyi are beating themj’61 ‘The womeni, theyj are beating themi’

Sentence (19) is ambiguous because both the Agent and the Patient belong to the same gender class and take the same agreement markers on the verb and because there is no use of suprasegmentals, as there can be in Modern Greek and Modern Standard Arabic, to assist in locating a sentence-initial TOPIC.62 Although the two glosses differ in what is the Agent and what, the Patient, they are alike in taking basádi ‘the women’ as TOPIC in both glosses, hence the identity in syntax. Northern Sotho FOCUS may be compared to Modern Greek as in Figure 4.The three degrees of FOCUS identified in Modern Greek as I-II-III in Figure 3 are, in Northern Sotho, expressed by a single morphosyntax. Lacking the

INTENSE LAX FOCUS FOCUS

GLOBAL TOPIC plus FOCUS I FOCUS or SVO II III PARTICULAR Paradigmatic FOCUS IV FOCUS

Figure 4: FOCUS in Northern Sotho. suprasegmental distinctions available in Modern Greek, Northern Sotho exploits a single grammar — preverbal, sentence-initial position and a remainder, the verb and what follows — to give form to a TOPIC FOCUS

61 I have altered Zerbian’s first gloss, which is ‘The women are beating them’, in order to underscore the TOPIC sense of this grammar.

62 On the matter of agreement, Louwrens, Kosch & Kotzé (1995.41) conclude: “The reason why object agreement becomes compulsory in sentences such as ... [(19)], is because the argument roles of subject and object can become confused when NP’s are presented in non- basic positions ... Due to the role played by agreement in this regard, the syntactic position which NP’s occupy becomes of secondary importance [To what?, PWD] and the word order pattern followed in the sentence doesn’t really matter ....” Misinterpretations like this one are fairly frequent. FOCUS: Nowhere 39 semantics, a range of FOCUS that is distributed over three contrasting grammatical expressions in Modern Greek. The TOPIC FOCUS grammar and its semantics appear also in what Zerbian (2006a) calls “inversion” or “impersonal constructions”. Consider these (Zerbian 2006a.367):

(20) (a) Go direg-ile eng? [SM17 happen-PERF what] ‘What happened?’

(b) Go fihl-ile mang? [SM17 arrive-PERF who] ‘Who arrived?’

The utterance (20c) is an appropriate response to either (20a) or (20b). Since ‘the uncle’ is not yet grounded and established in the context of the conversation, placing malome in sentence-initial position would imply the wrong thing.63 The answer of (20c) resolves that by placing the Agent following the verb among the non-TOPIC, FOCUS content. As noted above, the vagueness with which Northern Sotho encodes FOCUS in the TOPIC FOCUS grammar permits (20c) to respond both to the more GLOBAL inquiry of (20a) and to the more PARTICULAR one of (20b), i.e., to be I in Figure 4 or to be II or III. VS utterances also appear outside the context of questions (Zerbian 2006a.362):

(21) Go-fula di-kigomo [SM17-graze CL10-cow] ‘There are cattle grazing’

They have the sense of a presentational FOCUS (Zerbian 2006a.367).64

63 Note that the gloss ‘the uncle’, and not ‘an uncle’, appears to divorce this usage from any necessary connection with a ‘new’versus ‘given’ dimension. There is no comment on the semantic contrast between the questions of (20a) and (20b) as formed and the alternative in which the questions would begin with something like Ké eng ... and Ké mang ...

64 That these Agents are not TOPICS is additionally confirmed by the absence of the 40 SYNTAX & SEMANTICS

In sum, Northern Sotho, like other languages that have the grammar of in situ FOCUS, has a second implementation of FOCUS that is also more INTENSE than the in situ one. The difference of Northern Sotho lies in using a form of paradigmatic FOCUS to form the more INTENSE FOCUS and not an extension of the syntagmatic morphosyntax that embodies the in situ FOCUS.

3.2 Hausa65

Hausa is by far the most widely spoken of the Chadic languages. These languages are spoken in the vicinity of Lake Chad, a lake with adjoining borders to Nigeria, Chad, Niger, and Cameroon. They belong to the Afro-Asiatic language family. Hausa is spoken by more than thirty-five million speakers. It is the first language of the ehtnic Hausas in northern Nigeria as well as in the south of Niger. Hausa is also used as a lingua franca in many northern regions of Nigeria where it is establishing itself as a mother tongue in many cases. (Hartmann & Zimmermann 2007b.366)

Unlike Modern Greek and Northern Sotho, Hausa will show that the semantics of ASSERTION is implicated with FOCUS, but that the two are orthogonal. That is, while the two intersect, they are mutually independent. Hausa shares this configuration with Modern Standard Arabic, but Hausa is much more thorough in the implementation. All those who have worked on Hausa agree that the language has a neutral SVO order (Newman 2000.718, Jagger 2001.415, Hartmann 2006.581, Green 2007.8, and Hartmann & Zimmermann 2007b.366).66 Consider these (Newman 2000.718):67 appropriate noun class agreement with the verb, which TOPICS have. “... the logical subject does not agree with the verb in noun class features. The verb shows class 17 agreement instead. Class 17 refers to a class reserved for locatives, and occurs as ‘default’ agreement in Sotho languages in all cases where no subject agrees with the verb” (Zerbian 2006a.362). Cf. also the discussion of Kinyarwanda in Chapter 3.

65 Abdoulaye 1997 & 2007, Green 1997 & 2007, Green & Jagger 2003, Hartmann 2006, Hartmann & Zimmermann 2007a & 2007b, Jagger 2001 & 2006, Junaidu 1987 & 1989, Newman 2000, Schachter 1973, and Tuller 1986 & 1989.

66 Junaidu (1987) writes an entire dissertation on topicalization in Hausa without explicitly mentioning that Hausa may be an SVO language, neutral, basic or otherwise. Likewise, Tuller (1989), writing on Hausa FOCUS, says nothing explicitly about Hausa word order.

67 Hartmann (2006.602) observes Hausa is a tone language. It differentiates three tones. It has a high tone, which is FOCUS: Nowhere 41

(21) [Maryam PAC68 cook food] ‘Maryam customarily cooks food’ (22) [children PAC come home with ] ‘The children came home yesterday morning’

(23) [Gambo PAC tell us news] ‘Gambo will tell us the news’

The forms which precede the verb in (21) - (23) will play an important role in the expression of FOCUS. In these examples:

Inflection ... occurs as an independent lexical item from the verb. In Hausa, the verb carries no information for person, number, gender or tense/aspect ... All this information is carried by ... [a preverbal element] which forms a separate lexical item and is always overt. (Green 1997.13)

... the inflectional categories of subject-agreement plus tense, aspect and mood (TAM) are represented in an obligatory second position string of affixes and clitics; here referred to as the “person-aspect-complex” .... (Jagger 2001.148)

Temporal and aspectual information as well as person agreement are encoded in a separate morpheme, which I will refer to as “the auxiliary”. The auxiliary is usually left adjacent to the verb. However, it can be separated from the verb by certain emphatic particles and adverbs .... (Hartmann 2006.581)

Ex situ questions in Hausa are formed by placing the wh-word initially and by replacing the PAC with a complementary “relative” or “focus” form: “In tensed sentences, a Q-word or an NP containing a Q-word normally moves to the front of the sentence into focus position. The general TAMs are obligatorily replaced by the corresponding Rel forms ...” (Newman 2000.493). Here are a

not marked in the examples, a low tone (à), and a falling tone. Falling tones ( ) appear only in heavy, bimoric syllables, A circumflex on an open vowel, which is always long, indicates tone and length ( ‘liking’ is represented as ). The language has no rising tone. Some write long vowels as geminates, and some use a macron. I have not generalized either notation.

68 PAC = “person-aspect-complex”. Cf. also Jagger 2001.148. Some will call this TAM or Auxiliary, and others, INFL. When I refer to it here, I shall use PAC. 42 SYNTAX & SEMANTICS few examples (Green 2007.22-23, 84):69

(24) [who children 3PL.FOC.PF see] ‘Whom did the children see?’

(25) (a) [3pl-who 3PL.FOC.PF go America] ‘Who went to America?’

(b) [3pl-who FM.PL 3PL.FOC.PF go America] ‘Who went to America?’

(26) [to who.MS 2MS.FOC.PF return with book] ‘To whom did you return the book?’

(27) (a) [where 3PL.FOC.PF go] ‘Where have they gone?’

(b) [where FM.MS 3PL.FOC.PF go] ‘Where have they gone?’

(28) (a) [when 3PL.FOC.PF come] ‘When is it they are coming?’

(b) [when FM.MS 3PL.FOC.PF come]

69 The focused element in these questions appears sentence-initially, and it “is optionally followed by a focus-sensitive particle nee (masc./pl.) or cee (fem.)” (Hartmann & Zimmermann 2007b.367). Green (2007.60) calls it “the non-verbal copula ..., which shows polar tone.” We will discuss further below.

70 It is not clear to me why the PAC form is . That shape does not appear in Green’s own paradigms (2007.9) nor in Jagger’s (2001.161) paradigm of the Relative Perfective PAC. And Green does not comment on it. FOCUS: Nowhere 43

‘When is it they are coming?’

“Question words are by nature focused and typically appear fronted in focus position” (Newman 2000.188). The essential morphosyntax of ex situ of wh- question FOCUS is sentence-initial position plus the presence of “Rel” form of the PAC.71 Figure 5 (Abdoulaye 2007.236) illustrates the contrast between the General/Regular forms and the Relative forms for the Imperfective and the Perfective aspects.

General Relative General Relative Imperfective Imperfective Perfective Perfective others/west others/west

1s inàa nikèe/nikà naa na/niC 2ms kanàa kakèe/kakà kaa ka/kaC 2fs kinàa kikèe/kikà kin kikà/kinkà=kiC 3ms yanàa=shinàa yakèe=shikèe/shikà yaa ya/yaC 3fs tanàa takèe/takà taa ta/taC 1p munàa mukèe/mukà mun mukà/munkà 2p kunàa kukèe/kukà kun kukà/kunkà 3p sunàa akèe/akà sun sukà/sunkà one anàa akèe/akà an akà/ankà

Figure 5: General & Relative Paradigms of the Hausa PAC.

These are some questions paired with their answers (Jagger 2001.494, Hartmann & Zimmermann 2007a.243):

(29) (a) Wà kukà à kàsuwa? [who 2PL.FOC.PF see in market] ‘Whom did you see in the market?’

71 The non-“Rel” paradigm has been called “the general category” (Newman 2000.567). Others call it the “regular” form (Abdoulaye 2006.236). Hartmann (2006.582) calls it the “absolute,” and the form itself is an “auxiliary” (581). There is some additional disagreement about terminology. Newman (2000.567) writes, “In this grammar, I have followed the long- established convention in Hausa studies of using the label relative for this category [that is part of the grammar of FOCUS, PWD]. Jagger ... [(2001]) has recently suggested switching to focus, an innovation that in my opinion does not provide sufficient analytical or mnemonic advantages to justify the change.” Hartmann (2006) and Hartmann & Zimmermann (2007a & 2007b) continue to use “rel.” Green (1997) does not distinguish between the two in her grammatical glosses. Green (2007) uses “FOC” as do Green & Jagger (2003). Because there seems to be, as yet, no consensus, I will use whatever gloss the author(s) use(s). 44 SYNTAX & SEMANTICS

(b) mukà [boy.of.2MS 1PL.FOC.PF see] ‘It was your boy we saw.

(30) (a) [who 3SG.REL.CONT call-her] ‘Who is calling her?’

(b) [Dauda 3SG.REL.CONT call-her] ‘DAUDA is calling her’

While there seem not to be a lot of paired examples such as these in the literature, we are told “Hausa can focus-front on a wide range of syntactic functions” (Jagger 2001.499)72 These are a few additional examples (Jagger 1001.500-501 and Newman 2000.192-193):

(31) [with speed 1PL.FOC.PF return] ‘We returned quickly’

(32) [via London 1PL.FOC.PF come] ‘It’s via London we came’

(33) [with knife 3SG.REL.PF stab him] ‘It was with a knife that he stabbed him’

(34) [paying taxes Tanko 3SG.REL.PF do] ‘It is paying the taxes that Tanko did’

Like the wh-questions, responses may use the same syntax: sentence-initial

72 Green & Jagger (2003.196): “Hausa freely allows the focus fronting of any constituent, including VP ....” Hartmann (2006.584): “... nominal arguments ..., prepositional arguments and adjuncts ..., adverbials ..., and even clauses ... can be focused by fronting them ....”

73 I shall use underlining to note implosives. FOCUS: Nowhere 45 position and the Rel-form of the PAC. This grammar can generate opposed pairs such as these (Hartmann & Zimmermann 2007b.366-267):74

(35) (a) Kànde taa dafà kiifii [Kande 3SG.PF cook fish] ‘Kande cooked fish’

(b) Kànde ta-kèe dafà kiifii [Kande 3SG.REL.CONT cook fish] ‘KANDE is cooking fish’

Assuming that Hausa is SVO, when the Agent is focused, the only grammatical mark of that fact is the choice of the Rel-form of the PAC.75 The expression of FOCUS sentence-initially has uses in addition to serving as responses to wh-questions. We will discuss them as they contrast with the in situ expression of FOCUS. Hausa allows FOCUS to appear in situ in both wh-question forms and elsewhere. Its presence in the grammar of Hausa is a recent discovery:

In all previous accounts, focus in Hausa has been analyzed as exclusively syntactic, i.e. entailing movement [sic] ... There are some recently discovered facts, howeever, which demonstrate that in situ constituent focus is possible as an alternative to movement ... in situ focus is in general less common than preverbal stuctural focus. (Jagger 2001.496)

Jagger (2001:496-498) has recently demonstrated that, contrary to received wisdom, and consistent with the facts of a number of other (west) Chadic languages, there is evidence for focus in-situ in Hausa ... As Jagger points out, focus fronting is strongly preferred over focus in situ, which perhaps explains why these facts have been overlooked in the literature to date. (Green & Jagger 2003.196)

74 I have omitted cèe, the FOCUS MARKER, from (35b) in order to emphasize the contrast between (35a) and (35b). Sentence (35b) also uses the Continuative aspect, while (35a) uses the Perfective; but that is not relevant to the point here.

75 Denying their ears and eyes, some who work on Hausa insist on seeing “movement” where it does not exist. Hartmann (2006.599) refers to “The fact that subjects have to move if they are focused.” Newman (2000.189) writes, “even with subjects it is important to recognize that movement into the focus position has taken place.” In a formalism that injects “movement” as a part of the description, “subject wh-questions” are sometimes described as sentence-initial “as a result of (vacuous) movement” (Jagger 2001.496). “Vacuous” because in a language that is neutrally SVO, they are sentence-initial to begin with. And Green (2007.24) concurs: “subject focus always co-occurs with the focus form of INFL [which, in Hausa, is the PAC, PWD], suggesting vacuous displacement.” 46 SYNTAX & SEMANTICS

In situ focus is a common and frequent focus strategy in Hausa. (Hartmann 2006.597)

Although the ex situ WH- and focus constructions ... represent the norm in Hausa, there is much more to the story. Both WH-elements ... and focus elements ... also pattern together in terms of an alternative, pragmatically equivalent, information- packaging strategy — they can (and often do) occur in situ ....” (Jagger 2006.50)

Newman (2000.494) recognizes the possibility of in situ wh-questions, but (apparently) no other in situ FOCUS. The grammar of in situ FOCUS does not differ from that of a declarative SVO utterance. “... in situ focus is unmarked not only syntactically and morphologically, but also prosodically ... If a constituent is focused in its base position, it does not move, and the auxiliary does not appear in the relative form ... in situ focus is not marked at all in Hausa ...” (Hartmann & Zimmermann 2007b.367-368, 392). The grammatical organization of in situ FOCUS in Hausa is then similar to that in Northern Sotho (and some other languages); that is, the presence of FOCUS is vague and indeterminate in its morphosyntax, and it is the context of usage that identifies some portion of the utterance as distinct by the presence of FOCUS.76 The following are some examples of usage in which in situ FOCUS is probably present:

(36) (a) Wànè kaayaa kikà ? (Jagger [which things 2FS.FOC-PF forget] 2006.64) ‘Which things did you forget?’

(b) Naa jàkaataa dà hùulaataa [1S.PF forget bag.of.1S and hat.of.1S] ‘I forgot my bag and my hat’

(37) (a) Mèe sukà kaawoo? (Jagger 2006.65) [what 3PL.FOC.PF bring] ‘What did they bring?’

76 The circumstance in which there is no overt grammatical mark of focus requires that one pay attention to usage and context, otherwise FOCUS will go undetected. “... because overt syntactic movement with a morphological reflex on the TAM is much more visible than the in situ strategy, this is the only one documented in standard descriptions (with the marginal exception of WH-constructions ...). The result is that, prior to the in situ focus facts reported in Jagger (2001:496-98) and Green & Jagger (2003), the non-canonical in situ versions were hidden ‘below our radar’ ...” FOCUS: Nowhere 47

(b) Sun kaawoo rìigaa [3PL.PF bring gown] ‘They brought a gown’

(38) (a) kòfi? (Jagger 2001.497) [where coffee] ‘Where’s the coffee?’

(b) Yanà can cikin kwabà [3M.IMPF there in cupboard] ‘It’s there in the cupboard’

(39) (a) Mèè ya fààru gà ‘yan-taawaayèn?77 [what 3SG.REL.PF happen to rebels] (Hartmann ‘What happened to the rebels?’ 2006.597)

(b) An daurèè su [4PL.PERF imprison them] ‘They IMPRISIONED them’

(40) Àlbishi inkà, bàk sun [good.news.of.2M guests 3PL.PF arrive] 2003.211) ‘Guess what? The guests have arrived!’

In each of these pairs (36) - (39), the first utterance is a wh-question with the wh-word in the ex situ position and accompanied by a Rel-form of the PAC. In (40), the entire utterance is in in situ FOCUS following ‘’Guess What?’ The responses in (36) - (39) differ from the earlier ones in (29) and (30) in the placement of the answering content in the neutral position in an VSO utterance and in the absence of the Rel-form of the PAC. Agents do not participate in this usage (Jagger 2006.67):

(41) (a) ? [3PL-who 3PL.FOC.PF go America] ‘Who went to America?’

(b) %Su Audu dà Muusaa sun tàfi [3PL Audu and Musa 3PL.PF go]

77 Hartmann attributes this example to the prepublication version of Jagger 2006. The example does not, however, appear in the published paper. 48 SYNTAX & SEMANTICS

‘Audu and Musa went’

The normal declarative utterance of (41b) fails to respond to the wh-question of (41a). Only (41c) with the Rel-form of the PAC answers (41a). This is the basis of the conclusion that in situ FOCUS is not compatible with Agents: “subjects must be focused ex situ” (Hartmann & Zimmermann 2007b.371).78 The following two questions and their answers will provide some indication of the nature of the semantic contrast between ex situ and in situ FOCUS (Hartmann & Zimmermann 2007b.385 & 397):

(42) (a) Mèeneenèe ya fàaru? [what 3SG.REL.PF happen] ‘What happened?’

(b) Dabboobi-n jeejìi nee mutàanee su-kà kaamàa [animals-of bush PRT men 3PL-REL-PF catch] ‘The men caught WILD ANIMALS’

(43) (a) Mèe Audu ya yii [what Audu 3SG.REL.PF do] ‘What did Audu do?’

(b) % [fridge PRT 3SG.REL.PF buy] ‘He bought A FRIDGE’

Of course the questions differ, but not enough to explain why the FOCUS initial response of (42b) succeeds, while the FOCUS initial answer of (43b) is inappropriate. The contrast appears to lie in the interest, surprise, or startle effect of ‘wild animals’ and the mundaneness of ‘a fridge’: “both our informants indicated that the object [of (42b)] provided the interesting or surprising part of the utterance” (Hartmann & Zimmermann 2007b.385).

78 Cf. also Jagger 2006.67 and Green 2007.105, 108. FOCUS: Nowhere 49

Analogous to (39) and (40), Green (2007.64, 104) has these:79

(44) (a) Mè ya ? [what 3SG.FOC.PF happen] ‘What happened?’

(b) [thieves FM.PL 3PL.FOC.PF do IO.1S theft] ‘Thieves have stolen from me!’

(45) (a) À [good.news.of.2MS guests.DD 3PL.PF arrive] ‘Guess what? The guests have arrived!’

(b) % À [good.news.of.2MS guests.DD 3PL.FOC.PF arrive] ‘Guess what? The guests have arrived!’

The ex situ FOCUS of Agent ‘thieves’ in (44b) contrasts with its absence — and incorrectness — in (45). As in (42) and (43), the unexpected occurrence of being robbed by thieves is in opposition to the uneventfulness of guests arriving. In commenting on this example:

(46) (a) [naira twenty FUT 2SG.SUBJ pay if 3SG.PF do makà for.you] ‘It is twenty Naira that you will pay once he has done it for you’

(b)

‘No, I will pay fifteen’

79 The “DD” is “Definite Determiner”. Example (44) is also discussed in Hartmann & Zimmermann 2007b.385. Example (45b) is not actually provided by Green, but she notes “Hausa requires in situ subject focus in presentational focus contexts” and then gives (45a) as an example. The implication is that (45b) is not acceptable. Green’s comment also suggests that Agents can in fact function as FOCUS in situ. But given the example, it may be that the entire utterance is focused. 50 SYNTAX & SEMANTICS

Hartmann & Zimmermann (2007b.379, 390) remark, “Speaker B corrects A’s previous statement, but he does so without fronting the focus constituent. In the cultural context of Hausa speakers, this kind of negotiation is the norm ... In the cultural setting of Northern Nigeria, nobody would be surprised by B’s intention to pay less than he was asked for.” Wh-questions in the in situ expression are often not actually solicitations of information, but echo questions: “A Q-word can be left in situ, but if so it is generally interpreted as a semantically marked echo question” (Newman 2000.494). Green (2007.101) notes “considerable inter-speaker variation in Hausa with respect to the acceptability of wh-in situ80 ... Since these [echo questions] are not morphosyntactically distinct from genuine wh-in situ questions, context determines the echo versus the genuine wh-interpretations.” Ex situ FOCUS syntax never expresses the sense of an echo question. They are the more intense formulation of FOCUS. Such examples, uses, and comments upon them suggest that the semantic contrast between ex situ and in situ FOCUS turns in part on the degree of intensity, which may appear in the stronger ex situ FOCUS as surprise, unexpectedness, and the like. Hartmann & Zimmerman (2007b.389) conclude in this manner:81

... we would like to argue that focus fronting is always pragmatically induced. A focus constituent is fronted if it serves to guide the hearer’s attention to

80 Jagger (2006) discusses the variation in wh- in situ questions in greater detail.

81 There is, however, a bit of descriptive schizophrenia in Hartmann & Zimmermann’s paper. For example. Sentence (47b) below is a response to (47a) and illustrates “corrective focus” (Hartmann & Zimmermann 2007b.379). Hartmann & Zimmerman also see “corrective focus” in (46b), and conclude “that corrective focus can also appear ex situ or in situ.” However, the fact that the ex situ (47b) and (47c) are responses to (47a) show that the morphosyntax of (48a) - (48c) has nothing to do with “corrective focus” per se. Ex situ focus appears “corrective” in one and “affirming” in the other. Had the response to ‘Was it his mother who died?’ been ‘Yes, it was his mother who died,’ ex situ focus would have still been used even though no “correction” was present. “Corrective” is therefore not what the ex situ FOCUS of (47) turns on. Hartmann & Zimmermann import a priori notions such as “corrective focus,” “contrastive focus,” “exhaustive focus,” “selective focus” (380), etc. They then discover, as illustrated above, that Hausa does not align its grammar to the imported, foreign-to-Hausa concepts. The conclusion is not that the importations are a mistake, but that Hausa itself is inconsistent. The language is at fault (Hartmann & Zimmermann 392-393):

... it appears that focus in Hausa is marked inconsistently and even if it is marked there is no simple way to indicate the precise internal focus structure of the focus constituent ... focus does exist as a grammatical category in Hausa even though it is not consistently marked. FOCUS: Nowhere 51

unexpected discourse moves ... In our view, a focus constituent, or part of it, appears ex situ to mark its content or discourse function as unexpected or surprising in a given discourse situation.

Overall in situ FOCUS is the weaker of the two implementations of Hausa FOCUS, and ex situ FOCUS is the stronger.82 A second aspect to the contrast between ex situ and in situ FOCUS lies in the pointedness of the choice, content that is set off and distinguished from the remainder of the proposition.83 When an utterance is questioned yes/no and contains ex situ FOCUS, then the focused content is the point of primary interest.84 When in situ FOCUS is present, it is the occurrence of the proposition as a whole that is globally questioned (Hartmann & Zimmermann 2007b.379):85

(47) (a) [mother-of-his PRT 3SG.REL.PF die] ‘Was it his mother who died?’

(b) [no wife-of-3M PRT 3SG-REL-PF die] ‘No, it was his wife who died’

(d) [yes mother-of-his 3SG.PF die]

82 Noting that “All the above in situ locative WH- and focus elements could (and often would) be fronted to left periphery, of course, producing the following (near) synonymous ex situ variants with their focus-imperfective TAMs,” Jagger (2006.58) reports “some, but not all speakers, consider the preposed alternatives to be slightly more emphatic.”

83 This will be the dimension of FOCUS semantics, which in Figures 1 (Modern Standard Arabic), 3 (Modern Greek) & 4 (Northern Sotho), is summarized as GLOBAL — PARTICULAR.

84 This is also the pattern of Modern Greek (cf. [19] in Chapter 9.).

85 “Yes/No ... questions preserve normal word order ... The question, as opposed to the declarative statement, is overtly marked in one or more of four ways: (1) by addition of the q- morpheme; (2) by question intonation; (3) by a sentence-final interrogative tag; and (4) by a sentence-initial interrogative word” (Newman 2000.496-497). Cf. also Jaggar 2001.523-526. The examples here use intonation. Sentence (44c) is mine, and I have yet to confirm it with a speaker of Hausa. 52 SYNTAX & SEMANTICS

‘Yes, his mother died’

Because “Subjects represent a special category and cannot receive in situ focus” (Jagger 2006.67), the in situ (47d) is not a response to (47a). Jagger’s example is this:

(48) (a) ’yan-sàndaa sun gaanoo gaawar mamàc [police 3PL.PF discover body.of dead man.the] ‘Did the police discover the dead man’s body?’

(b) Aa’aa, ’yaa’yan mamàc sukà [no children.of dead.man.the 3PL.FOC.PF gaanoo shì discover 3MS] ‘No, the dead man’s children discovered him’

(c) %Aa’aa, ’yaa’yan mamàc sun [no children.of dead.man.the 3PL..PF gaanoo shì discover 3MS] ‘No, the dead man’s children discovered him’

Sentence (48a), however, is an in situ yes/no question and exhibits the vagueness of in situ FOCUS. It does not specifically identify the Agent as its FOCUS, and a morphosyntactically similar question (Jagger 2006.65-66) may just as well take some other function as its FOCUS. In (49b), it is the Patient that is focused in situ:

(49) (a) Kaa gayàa wà Augù làabaar_ìn? [2MS.PF tell to Audu news.the] ‘Did you tell (to) Audu the news?’

(b) Aa’aa, naa dai gayàa wà Kànde [no 1SG.PF actually tell to Kande] ‘No, I actually told Kande’

A yes/no question with ex situ FOCUS does not have this latitude. The play of GLOBAL versus PARTICULAR may also be at work in this pair (Hartmann & Zimmermann 2007b.380): FOCUS: Nowhere 53

(50) (a) Dàalìbai sun sàyi lìttàttàafai kawài [students 3PL.PF buy books only] ‘The students bought only books’

(b) Lìttàttàafai kawài dàalìbai su-kà sàyi [books only students 3PL.REL.PF buy] ‘The students bought only books’

The first, in situ expression is appropriate to a context in which the students ran out of money in the store. The activity failed to reach its intended outcome. The ex situ expression would be appropriate if the students completed their planned purchase and the purchase happened to include only books. The in situ form identifies with some imprecision what was purchased.86 The target of the intended purchase is incomplete and ill-defined, while in the latter, the purchase is well-defined: books and nothing more.87 In combination with the Perfective Aspect, the pointedness of ex situ FOCUS (Jagger 2001.162) is appropriate to:

express specific, punctual events ... [and] it is also the main marker of foregrounded material in historical narrative sequences, where it tracks single- occurrence, chronologically-sequenced events in the past ...

Jagger’s example is (Jagger 2001.163):88

(51) (a) mukà búdè , [1PL.FOC.PF open door car ‘We opened the car door and

(b) mukà yí , [1PL.FOC.PF move quickly and we moved quickly

86 I have not yet been able to consult a speaker of Hausa about this interpretation.

87 Hartmann & Zimmermann (2007b.380) conclude that both (50a) and (50b) demonstrate “exhaustive” FOCUS and together, they contribute to the conclusion that “focus in Hausa is marked inconsistently” (Hartmann & Zimmermann 2007b.392).

88 This is taken from a “brush-with-death narrative,” but the context is not provided. 54 SYNTAX & SEMANTICS

then some man, he opened it for me

(d) mukà [1PL.FOC.PF went we went

(e) mukà [1PL.FOC.PF open we opened it’

The staccato, one-event-upon-another is compatible with the precise pointing of ex situ FOCUS, in contrast with the more fluid, unformed global in situ FOCUS. Ex situ FOCUS has another interaction with Hausa Aspect that reveals additional information about its content. Hausa has a complex system of at least eight tense-aspect-modal forms:

(52) Perfective: “The canonical use of the Perfective is to report anterior, completed past-time events in the non-complex, often monoclausal, affirmative statements, where the deictic notion ‘past’ is relative to a given time-point” (Jagger 2001.156).

(53) Imperfective: “When occurring with predicates containing a verbal or dynamic noun, the Imperfective encompasses the ternse-aspect dimensions of both durativity (action-in- progress) and habituality, and is used to highlight the internal time-structuring of the situation relative to a given time- point” (Jagger 2001.168-169)

(54) Habitual: “The Habitual is used tyo encode activities/events, present and past, with a habitual customary time- reference—and so overlaps partially with the Imperfec- tive—but can also express sporadic actions” (Jagger 2001.182). FOCUS: Nowhere 55

(55) Subjunctive: “The Subjunctive TAM is basically a modal, often non-factual, category which expresses a wide range of context-sensitive, and sometimes overlapping illocutionary acts, including commands, prohibitions, permission, intentions, instructions, proposals, suggestions, obligations, responsibility, requests wishes, etc.” (Jagger 2001.184).89

(56) Future: “The Future, affirmative and negative, is used to express both tense (future-time) and modal (attidutinal) distinctions, in both general and focus contexts” (Jagger 2001.194).

(57) Allative: “The Allative typically expresses actions which are future with respect to the moment of speaking, corresponding to an English imminent/ingressive ‘be going to, be off, be on the way to, etc.’ construction with locative goal complements” (Jagger 2001.198-199).

(58) Potential: “The Potential is best analysed, like the Subjunctive, as a modal category, and is less assertive than the Future. It expresses a range of attitudes, including uncertainty, doubt, indefiniteness, probability, vagueness, etc. as to the future realization of an action/event, and so is usually glossed as ‘will probably/likely ...’ or ‘may/might ...’ in English” (Jagger 2001 301).90

(59) Rhetorical: “The Rhetorical TAM — also known, inter alia, as the ‘Rhetorical Future’ and ‘Future Relative’ — basically conveys the futurative notion of possibility, eventuality, etc. and is regularly used to express self-answering rhetorical questions or statements, including fixed proverbs and compounds” (Jagger 2001.205).91

89 “The subjunctive expresses wishes, desires, purpose, obligation, etc. In the second person, it serves as a somewhat softer alternative to the imperative for expressing commands” (Newman 2000.591).

90 “... the ‘potential’, which is short for potential-future, indicates an action that will possibly take place in the future (God willing)” (Newman 2000.587).

91 “The rhetorical is an infrequently encountered TAM that has not been adequately studied and about which we therefore know very little. It is used primarily in set expressions, idioms, epithets, proverbs, compounds, and such, but not exclusively so ... It occurs in the affirmative only ... The rhetorical implies doubt or even a dare with respect to the possibility of 56 SYNTAX & SEMANTICS

The reason that I have cited lengthy descriptions of the specific tense-aspect- modalities is that ex situ FOCUS is not possible with the Subjunctive and the Potential, and it is required with the Rhetorical.92 Semantically, the Subjective and the Potential appear to represent an extreme of uncertainty (Recall Newman’s “God willing”.), while the Rhetorical, on the other hand represents the opposing extreme of certainty. The remaining tense-aspect-modalities lie somewhere between the extremes.93 The highest degree of certainty (actualized or not) is manifest in the assuredness of self-answering questions of the Rhetorical and in the inevitability of proverbs, even though they are not specific (Newman 2000.131):

(60) [whatever strength man someone 3SG.MS.REL.PF

exceed he] ‘Whatever a man’s strength, there is someone stronger’

Jagger (2001.504) remarks of the incompatibility between the Subjunctive and the Potential and ex situ FOCUS:94

This prohibition is probably due to the semantic incompatibility between a non- achieving some action. It is often best translated with such English modals as ‘should’ or ‘could’. As the name indicates, it is commonly used in rhetorical questions or statements” (Newman 2000.589).

92 “For many speakers, there is a restriction against using the quasi-modal Potential and Subjunctive TAMs in focus constructions ...” (Jagger 2001.504). “The subjunctive does not occur in Rel [i.e. FOCUS] environments” (Newman 2000.591). “The potential does not occur in Rel environments” (Newman 2000.588). “Syntactically it [the rhetorical] is restricted to Rel environments” (Newman 2000.589). “The key feature of its [the Rhetorical’s] syntactic distribution in SH is that it occurs almost exclusively in affirmative focus environments” (Jagger 2001.205). It would be useful to investigate the qualifications “for many speakers” and “occurs almost exclusively.”

93 Because of the Imperfective, the Future and the Allative, the dimension is not ‘realized’ or not, ‘actual’ or not, nor ‘past’ or not.

94 Jagger does not, that I can find, comment on the Rhetorical and ex situ FOCUS. Tuller (1986.71) writes similarly:

One of the fundamental properties of the subjunctive mood is that it expresses uncertainty/indefiniteness. Suppose that the property of the subjunctive may be grammaticized by a subjunctive INFL ... Now the asymmetry in occurrence becomes a mechanical phenomenon of form. FOCUS: Nowhere 57

specific modal category and the type of high specific narrow focus entailed by focus ... constructions.

‘Non-specific” would seem to be a problematic choice here, since proverbs, which contain “non-specific” references, occur in the Rhetorical Aspect and require ex situ FOCUS. ‘Certainty’ in its extremes —absence or presence — appears to be closer to capturing the exclusion of ex situ FOCUS from the Subjunctive and the Potential, and, at the same time, it permits the complementary behavior of the Rhetorical Aspect, which constantly associates with ex situ FOCUS.95 An example of the interaction can be seen in (58) (Jagger 2001.504):

(61) [Musa FM 3PL.FOC.PF agree COMP 3SG.SUBJ

become director] ‘It’s Musa they agreed should become director’

Had Musa occurred in the clause ‘Musa should become director’, it could not have shown the ex situ grammar that it in fact has in its occurrence in the first clause, ‘They agreed on Musa, [that he should become director]’. While not itself a component of the semantics of FOCUS, the scale of ‘uncertain’ — ‘certain’ is certainly a semantics to which FOCUS is sensitive and a semantics to which FOCUS aligns itself. But the dimension of ‘uncertain’ — ‘certain’ is not a semantics that we have previously found in a relationship with FOCUS ... in that form. One interpretation of ‘uncertain’ — ‘certain’ is this. We have previously discovered the semantics of ASSERTION to be

95 In a discussion of semantic explanations of the intersection between FOCUS and Aspect, Green (1997.62) cites “a challenge to the proposed semantic account,” which turns on “the fact that the focus perfective [kikà in (i)] can occur in open conditional constructions, which also express epistemic modality.” The example is this: (i) In kin/kikà mù tàfi [if/when 2FS.Pf/2FS.FOC.PF be.ready FUT 1PL go] ‘If/when you’re ready, we’ll go’ Jagger (2001.608) says that in “is often used to initiate a chain of sequential events where its meaning is closer to ‘when’ ...,” which certainly is not close to the ‘uncertain’ extreme of the Subjunctive and the Potential, which Jagger goes on to note “are incompatible with open conditional statements” such as (i). It seems that (i) that “open conditional constructions” do not show an extreme of ‘uncertainty’ that would prohibit ex situ FOCUS, and indeed, they [i.e. in ‘if/when’] are so ‘certain’ that the Subjunctive and the Potential are incompatible with them. If anything, (i) supports a semantic explanation of the interaction between FOCUS and Aspect. 58 SYNTAX & SEMANTICS

involved in FOCUS. In our experience to this point, Wolof is perhaps the most extreme example of this. It expresses its degrees of FOCUS INTENSITY in terms of degrees of increasingly narrowed ASSERTION (Chapter 5). Modern Standard Arabic expresses two degrees of GLOBAL INTENSITY of FOCUS by morphological means: lahal for the INTENSE value and its absence for the opposed LAX value. It might be that the extreme ‘uncertainty’ of the Subjunctive and the Potential is actually a weak degree of ASSERTION, and the Rhetorical represents the strongest degree of ASSERTION. One indication of this is the occurrence of nee/cee, a morphological mark of ASSERTION. While it occurs with the Rhetorical, it is not compatible with the Subjunctive or the Potential (Jagger 2001.185, 201):96

(62) (a) Mù tàfi yànzu [1PL.SJN go now] ‘Let’s go now’

(b) *Mù tàfi yànzu nee [1PL.SJN go now ASSERTION] ‘Let’s go now’

(63) (a) dà gòbe iyà Hausa [today and tomorrow 2F.POT be.able Hausa] ‘In time you’ll probably master Hausa’

(b) dà gòbe iyà Hausa nee [today and tomorrow 2F.POT be.able Hausa ASSERTION] ‘In time you’ll probably master Hausa’

The contradiction between the ASSERTION of nee and the ‘uncertainty’ of the Subjunctive and the Potential confirms the absence of ASSERTION in ‘uncertainty’. The mark of ASSERTION is nee/cee.97 In one of its simplest applications,

96 Sentences (62b) and (63b) are mine, and I have not yet been able to confirm them with a native speaker of hausa.

97 The particle “takes the form when agreeing with items that are masculine singular or plural and when agreeing with items that are feminine singular” (Newman 2000.545). The particle has many glosses and labels. Newman (2000.545 ff.) generally calls it “stabilizer.” Because it occurs as in (61) and appears to be verbal, some call it a “copula” (e.g., Jagger 2001.508 and Abdoulaye 2007.263), “however this is not intended to imply that FOCUS: Nowhere 59 when nee/cee follows a Noun, it provides the ASSERTION necessary to create a PROPOSITION (Green 1997.186): (64) Yanzu likita ne [now doctor ASSERTION] ‘Now he’s a doctor’

When it follows two Nouns or a Noun and an Adjective, it ASSERTS/ predicates the second of the first (Green 1997.135-136, Hartmann & Zimmermann 2007a.261):

(65) (a) Audu dalibi ne [Audu student ASSERTION] ‘Audu is a student’

(b) Audu dogo ne [Audu tall.M ASSERTION] ‘Audu is tall’

“... the time-reference/aspect of these constructions is interpreted with respect to context ...” Green 1997.186). The unacceptability of (65c) shows that the ASSERTION is required.98 In addition to appearing in verbless expressions such as (65a) and (65b), nee/cee also occurs following Verbs (Jaggar 2001.509, Green & Jagger 2003.198):

(66) (a) Bài cikà zuwà ba nè [NEG 3M.PF often come NEG ASSERTION] ‘He doesn’t come here too often’

ne/ce is accorded verbal status” (Green 1997.135). And because it interacts with FOCUS, e.g. in (25b), (27b), and (28b) above, it is also called a ‘focus marker’ (e.g. Newman 2000.546). Green (1997) glosses it as FOC. In previous examples above, it has been glossed grammatically as FM. Henceforth, I shall use ASSERTION.

98 “nee/cee is always obligatory in predicative constructions” (Hartmann & Zimmermann 2007a.261). For these authors, the presence of nee/cee “can be derived from the fact that predicates necessarily involve focus ....” 60 SYNTAX & SEMANTICS

(b) [FUT.1S buy car ASSERTION] ‘I’m going to buy a car’

The masculine form of ASSERTION in (66b) does not accord with the feminine ‘car’, and the disagreement prompts Green & Jagger to conclude that the particle “is not agreeing with the constituents immediately left-adjacent to it, but with the whole sentence.”99 Jagger distinguishes between the function of the sentence-final nee and the function of the unmarked in situ FOCUS. In (67b) (Jagger 2006.62-63):

(67) (a) Ìnaa zuwàa? [where going] ‘Where are you going?’

(b) tàfi gidaa nèe [1S.FUT go home ASSERTION] ‘I’m going home’

“the (optional) occurrence of the default masculine focus marker nèe as used in sentence-final position in ... [(67b)] correlates with sentence-focus, i.e. where the focus extends its (wide) scope over the entire proposition” (Jagger 2006.63). There is then a second, complementary FOCUS on gidaa ‘home’: “the (in situ) focus is unambiguously assigned to the goal locative noun gidaa ‘home’” (Jagger 2006.63).100 Jagger (2006.63) interprets nèe as an expression

99 This slightly alters Jagger’s (2001.497) earlier description.

100 Abdoulaye (2007.234 et passim) confirms Jagger’s 2006 description, “distinguishing the contribution of the fronting and the contribution of the copula nee/cee to the semantics of focus constructions.” Hartmann & Zimmermann (2007a) have a different interpretation of sentence-final nee (Hartmann & Zimmermann 2007a.248): (i) (a) Wàacee cèe ka ganii à ? [who-FEM PRT 2SG.REL.PF see at school] ‘Whom did you see at school?’ (b) Naa gaa dèelu à nè [1SG.ABS.PERF see Delu at school PRT] ‘I saw DELU at school’ Here, the masculine nee not only does not agree with the gender of the feminine in situ FOCUS Delu, it does not directly follow the content to which it makes its semantic FOCUS: Nowhere 61 of FOCUS. In an utterance such as (66b), “the default masculine singular focus marker nèe ... in sentence-final position ... correlates with sentence-focus, i.e., where the focus extends its (wide) scope over the entire sentence.” If the entire utterance of (66b) is focused by nee, then how does (66b) differ from (40), repeated here as (68), in which the entire utterance similarly appears to be focused:

(68) Àlbishi inkà, bàk sun [good.news.of.2M guests 3PL.PF arrive] 2003.211) ‘Guess what? The guests have arrived!’

Either there is a difference of kind, or the contrast is a matter of degree. The former seems correct; (66b) shares GLOBAL with (68), but (66b) has the increment of ASSERTION provided by nèe. Abdoulaye (2007.245) provides a pair, based on an example from Jagger (2001.510), which illustrate the semantic effect of a sentence-final nee:101

(69) Bàa màataa-taa taa yi yaajìi ba nèe [NEG wife-1SG 3FS.PF do run NEG ASSERTION] ‘It is not the case my wife has left me’

(70) Bàa màataa-taa taa yi yaajìi ba [NEG wife-1SG 3FS.PF do run NEG] ‘I am not talking/do not want to talk about my wife’s leaving me’

In (69), nee seems clearly to take the content ‘My wife has not left me’ and then ASSERT of it ‘It is the case’ ... ‘My wife has not left me’.102 Newman (2000.546) writes:

The stabilizer also functions as a sentence-level or clause-level reinforcing element. (In this function, it always appears as because sentences and clauses are intrinsically masculine.) The STAB sometimes occurs at the very end of the contribution. Hartmann & Zimmermann (2007a.248) comment, “nee/cee can associate with the focus at a distance ....” Jagger’s interpretation seems preferable.

101 Jagger’s example is (i) Bàa wai màtata yi ba nè [NEG wife 3FS.PF do run NEG ] ‘It’s not the case that my wife has left in a huff’

102 Abdoulaye is correct in saying that (68) “is not a more emphatic version of” (69). The contrast lies elsewhere. 62 SYNTAX & SEMANTICS

sentence, and sometimes at the end of the core sentence but before adverbial adjuncts or complements. The exact semantic contribution of the STAB is not clear. Sometimes it appears to add a degree of insistence [i.e., ASSERTION, PWD] to the truth value of the clause or sentence, e.g. ‘I do know (it) (cf. ‘I know (it)’. Other times, it appears to demarcate the core sentence from subsequent modifiers or adjunct. Finally, in some instances, it seems to be little more than a pragmatic pause filler comparable to ‘you know’ in colloquial English.

If we hew closely to our heuristic that FOCUS can always be, at its core, recognized by the grammar used to respond to a wh-question, then nee/cee is clearly not FOCUS. As we first saw in (25b), (27b), and (28b), nee/cee may also follow the ex situ FOCUS. It can be present when a wh-question is answered, but it does not create the response. It modulates the response. (Jaggar 2001.495):

(71) (a) Nawà kikà shì? [how.much ASSERTION 2F.FOC.PF give 3M] ‘How much did you give him?’

(b) Naira nè na shì [naira thousand ASSERTION 1SG.FOC.PF give 3M] ‘It was a thousand naira I gave him’

Again, “The presence/absence of FM does not alter the type of focus, only its ‘impact’; native speakers describe it as ‘adding emphasis’” (Green & Jagger FOCUS: Nowhere 63

2003.196).103 Abdoulaye (2007.250) has two instructive examples:104

(72) (a) Kaa san Bàmàko? [2MS.PF know Bamako] ‘Do you know Bamako?’

(b) Nii na san Bàmàko kùwa ... [I 1SG.REL.PF know Bamako indeed ...] ‘I indeed know Bamako ...’

(73) (a) Wàa-ee nèe a-kà cée yaa [who-MS ASSERTION one-REL.PF say 3MS.REL.PF san Bàmàko cikin-kà? know Bamako inside-2PL] ‘Who was it said knows Bamako among you?’

(b) Nii nèe na san Bàmàko [I ASSERTION 1SG.REL.PF know Bamako] ‘I am the one who knows Bamako’

The yes/no question in (72) allows nii ‘I’ to assume the ex situ FOCUS using

103 Abdoulaye (2007.344-345), following Newman (2000) finds three functions for nee/cee in sentence-final position: (i) “clause or sentence-level reinforcement,” (ii) “demarcative,” or (iii) “pause filler functions.” When nee/cee follows an ex situ FOCUS, it “expresses the deictic identification focus, the specific reference focus, and the exhaustive listing focus” (Abdoulaye 2007.348). When nee/cee appears in wh-questions, Abdoulaye (2007.262) finds another three functions: (i) “the deictic identification focus with visible wh-referents,” (ii) “the non-empty set or specific referent focus” for remote referents, and (iii) “the hearer knowledge presupposition focus, i.e., the presupposition that the hearer knows the referent of the wh-word.” Except for useing the term “copula” to label nee/cee, there is no suggestion how all these nine functions hang together. Hartmann & Zimmermann (2007a) see “exhaustivity” in nee: “Nee/cee exhibits typical exhaustivity effects ... Nee/cee triggers an exhaustivity effect by meants of a conventional implicature...” (Hartmann & Zimmermann 2007a.258). Nee is not a mark of FOCUS: “... the hypothesis that the particle nee/cee is a focus marker appears to be unwarranted” (Hartmann & Zimmermann 2007a.246). Rather, nee is a “focus-sensitive particle”: “... if nee/cee occurs, a focus must occur to its left ... “ (Hartmann & Zimmermann 2007a.231). This, of course, leaves aside uses such as (65a), (65b), and (68).

104 I have replaced Abdoulaye’s ‘cop’ with ASSERTION. 64 SYNTAX & SEMANTICS

the FOCUS PAC form na in (72b), but the ASSERTION which nee adds to nii ‘I’ in (72c) is not appropriate in response to an inquiry about the entire PROPOSITION.105 Nee is suitable, however, in response to the question of (73a), which solicits information specifically about the identity of the one who knows.106 Although there is no unanimity in the literature on whether nee/cee represents FOCUS, no one questions that nee/cee is somehow intertwined with FOCUS. There is disagreement on what nee/cee contributes semantically and on how nee/cee relates to the other expression of FOCUS. Because nee/cee interacts with FOCUS all along the dimension of GLOBAL — PARTICULAR, from the absence of any specific FOCUS to in situ FOCUS to ex situ FOCUS, it appears that the ASSERTION that nee/cee signals is orthogonal with the FOCUS signaled by the in situ — ex situ dimension. That is, for each value of FOCUS along the dimension of GLOBAL — PARTICULAR, there is a second, crosscutting added value of LAX — INTENSE. Figure 6 attempts to depict the relation.

105 However, sentence (i) with a sentence-final nee might succeed as a response to the yes/no question: (i) Nii na san Bàmàko nee ... [I 1SG.REL.PF know Bamako ASSERTION ...] ‘I know Bamako ...’ Hartmann & Zimmermann (2007a.250) provide this example in response to ‘Who came?’: (ii) *Audù ya zoo nè [Audu 3SG.REL.PF come ASSERTION] ‘AUDU came’ They comment that (ii) is “only grammatical as [a] yes-no question[...] where the final particle functions as a question tag. A declarative reading of these sentences is not available. At present, the source of this additional restriction is mysterious to us.” They do not consider the contrast between (ii) as a yes/no question and a yes/no version of (ii) without nee, but with the appropriate intonation.

106 Abdoulaye (2007.251) sees “exhaustive” listing in the use of nee in (71b). “... it is strongly implied,” but it is not clear that it may not still be somehow avoidable. Failing to use nee in (73b) “would make the reply sound more modest,” but, nevertheless, apparently still “exhaustive.” If so, then “exhaustive” is not the contribution of nee in (73b). Nee is the contrast between the absence modesty and its presence. One could see the ‘pushiness’ of ASSERTION making that difference. FOCUS: Nowhere 65

INTENSE LAX FOCUS FOCUS With nee GLOBAL FOCUS

SVO with no contextually With nee identified in situ FOCUS

SVO with a With nee contextually Without nee identified in situ FOCUS

SVO with ex situ FOCUS Without nee

Without nee PARTICULAR FOCUS

Figure 6: FOCUS in Hausa.

Hausa appears to support the FOCUS in situ Conjecture of section 1. The configuration of FOCUS in Hausa contrasts dramatically with Wolof, in which the two dimensions are not orthogonal, but aligned and congruent. Hausa also contrasts sharply with Mupun, in which the two dimensions are again merged, but without multiple distinctions along the dimension of GLOBAL — PARTICULAR, which are characteristic of Wolof (and Hausa).

3.3 Mupun Mupun is a member of the “Angas subgroup of West Chadic” (Frajzyngier 1993.iv):

The language is spoken in the Plateau State of Nigeria, in an area with difficult access lying some sixty miles southeast of Jos ... The 1963 census gave the number of speakers of Mupun as 11,016 .... (Frajzyngier 1993.i)

Green (1997.193) writes that, “Mupun appears to be a wh-in situ language across the board: there is no syntactic reordering of subjects or non-subjects in 66 SYNTAX & SEMANTICS

wh-questions.” This is an immediate challenge to the FOCUS in situ Conjecture, which requires a FOCUS in situ language to have a second expression of FOCUS that is more INTENSE than the in situ one. The key to understanding Mupun follows soon after: “... it is striking that wh-expressions are preceded by the copula (COP) a” Green 2007.193). Like Northern Sotho and Hausa, Mupun is an SVO language (Frajzyngier 1993.viii, 178).107 For example (Frajzyngier 1993.181, 186, 194, 252, 359):108

(74) [mother DEF spread shirt DEF] ‘Mother spread the shirt’

(75) N-den da n-duk bin [1SG-put calabash DEF PREP-on bench] ‘I put the calabash on a bench’

(76) Wu ji a keen fen [3MS come PREP before 1SG] ‘He came before me’

(77) (a) War a mat kamkam [3FM ASSERTION woman teacher] ‘She is a teacher’

(b) [goat ASSERTION riches 3MS] ‘A goat is all he has’

The form a in (77) provides the semantic ASSERTION to create a PROPOSITION

107 The literature on Mupun is limited. The primary sources are a grammar, Frajsyingier 1993, and a dictionary, Frajzyngier 1991. Mupun also figures in a 1997 dissertation (Heusing 1999).

108 Underlining is used to mark implosives. To emphasize the continuity with the preceding section on Hausa, I have replaced Frajzyngier’s COP gloss of a in (77) and elsewhere with ASSERTION. FOCUS: Nowhere 67 as nee/cee did in Hausa, but unlike Hausa, Mupun adjectives are sufficiently EVENT-like in their ASSERTION that a is not needed to mark the fact of PROPOSITION (Frajzyngier 1993.254-255):

(78) (a) Nwo mo a de bis ASSERTION ‘All snakes are bad [things]’

(b) Nwo mo bis

‘All snakes are bad’

The difference between (78a) and (78b) lies first in the presence of de in the former which renders bis ‘bad’ as a Noun, de bis ‘bad thing’. The second difference is that (78a) has a asserting de bis of nwo mo ‘snakes’. “An inherently verbal form may occur in an equational sentence [i.e., one with a, PWD] only after it has been nominalized, which in most cases means after it is preceded by the relative marker de ...” (Frajzyngier 1993.254). Yes/no questions in Mupun are marked by “an extra high tone” sentence- finally (Frajzyngier 1993.359), but that tone falls on one of three morphological elements that qualify the intent of the yes/no question: “the primary means of indicating questions in Mupun are clause final vowel markers, which occur in all types of questions” (Frajzyngier 1993.368). For example (Frajzyngier 1993.362):

(79) Yo a sin [REC.PAST 2MS give goat DEF n-Nalep-a? PREP-Nalep-ASSERTION] ‘You gave the goat to Nalep, didn’t you?’

(81) A Nalep-a? [ASSERTION Nalep-ASSERTION] ‘This is Nalep isn’t it?’

Sentence-final a “indicates that the speaker actually believes in the truth of the proposition and is looking for confirmation of his beliefs” (Frajzyngier 1993.362). That is, it is the issue of ASSERTION that is being questioned. Sentence (81) has a twice, once to assert Nalep and again to assert (i.e., questioningly) the PROPOSITION in its entirety. Frajzyngier (1993.363) also 68 SYNTAX & SEMANTICS suggests that “It is likely that the clause final marker -a is the same as the copula a.”109 There are two additional choices with Mupun yes/no questions. One can also question a PROPOSITION with no ASSERTION as in (82) (Frajzyngier 1993.360):

(82) Wu naa mun-e? [3MS see 3PL-NO.ASSERTION] ‘Did he see us?’ or with “surprise, disbelief” (Frajzyngier 1993.363) as in (Frajzyngier 1993.364):

(83) A ha siar Darap-o? [ASSERTION 2MS friend Darap-SURPRISE.ASSERTION] ‘Are you a friend of Darap?’

In (82), there is no expectation positively or negatively. The speaker simply wants to know. Question (83) suggests that “friends don’t behave like that, I think you are not a friend of D.” (Frajzyngier 1993.364). Wh-questions similarly require a (Frajzyngier 1993.370-372):110

(84) A mi ya jos -i? [ASSERTION what caught rat DEF- ] ‘What caught the rat?’

(85) Yo a sin a mi n-Nalep-i? [REC.PAST 2MS give ASSERTION what PREP-Nalep-] ‘What did you give to Nalep?’

109 This recalls the use of Hausa ASSERTION in sentence-final position, the difference being that in Mupun final ASSERTION always constitutes a question, while in Hausa it expresses a yes/no question only when there also is an ex situ FOCUS. Recall (Hartmann & Zimmermann 2007a.250):

(i) Audù ya zoo nè? [Audu 3SG.REL.PF come ASSERTION] ‘Did AUDU come?’ which exists in Hausa only as a yes/no question. Frajzyngier’s gloss for the sentence-final use is INTERR.

110 The sentence-final vowel for wh-questions is i (morphophonemically absent following i). FOCUS: Nowhere 69

(86) Wu sin a n-wi? [3MS give goat DEF ASSERTION PREP-who] ‘To whom did he give the goat?’

The targets of wh-questions appear in the linear positions in which unquestioned contents filling the same function would appear.111 That is, wh- questions are in situ. Responses to wh-questions will be marked with ASSERTION in the way the wh-question itself is (Frajzyngier 1993.367):

(87) (a) A wi? [ASSERTION who] ‘Who is it?’

(b) A an [ASSERTION 1SG] ‘It is me’

Oddly, this example is as close as we come to discovering how to answer Mupun wh-questions. Frajzyngier (1993.397-425) discusses FOCUS in terms of “contrastive focus” (as illustrated in [88] - [92]), but it seems clear enough that those same expressions would also respond to wh-questions (Frajzyngier 1993.400, 402, 404, 405):

(88) (a) A an tal ba a [ASSERTION 1SG ask tobacco NEG ASSERTION war kas 3FM NEG] ‘It is me who asked for tobacco, not she’

(b) A war a mat fen [ASSERTION 3FM ASSERTION wife 1SG] ‘It is she who is my wife’

(89) (a) War cet a lua ba a [3FM cook ASSERTION meat NEG ASSERTION pupwap kas fish NEG]

111 “... there is no movement of the interrogative complex ...” (Frajzyngier 1993.367). 70 SYNTAX & SEMANTICS

‘She cooked meat, not fish’

(b) Yi baal se a

‘Are you about to start eating your inlawship?’

(90) Wu gap a si sep [3MS cut branch tree DEF ASSERTION PREP ax ba a cu kas NEG ASSERTION knife NEG] ‘He cut a branch of tree with an ax not with a knife’

(91) Mo cet a cet lua ba mo [3PL boil ASSERTION cook meat DEF NEG 3PL sur a sur kas112 fry ASSERTION fry NEG] ‘They boiled the meat, they didn’t fry it’

(92) (a) Wur a [3MS ASSERTION tall] ‘He is tall (rather than short)’

(b) Wur [3MS tall] ‘He is tall’

Frajzyngier (1993.398) concludes about (88) - (92), “There is little doubt that is the same morpheme as the copula.” The first thing to notice about these sentences is that the focused constituent appears in the same linear order that its unfocused congener would occupy. That is, Mupun does appear to be a FOCUS in situ language. Now the question must be, “Just how FOCUS in situ is Mupun?” Is Mupun like Modern Standard Arabic, Modern Greek, Northern Sotho, and Hausa? The answer is, No, it is not. It is more like Bella Coola. Both Bella Coola and Mupun rely upon the semantics of ASSERTION to effect FOCUS. Bella Coola signals ASSERTION with sentence-initial position and in the neutral case

112 “When a verb is to be marked for contrastive focus, the appropriate construction has the form: V + a + V, i.e., the verb is repeated and the copula comes between the two stems” (Frajzyngier 1993.405). FOCUS: Nowhere 71 associates the EVENT with ASSERTION. Neutrally, FOCUS co-occurs with the EVENT, and the language is Verb-initial. If content other than the EVENT is to be focused, it has to assume the sentence-initial position, for that is where ASSERTION and hence FOCUS lies. Mupun, like Bella Coola, renders FOCUS with ASSERTION, and like Bella Coola, Mupun associates ASSERTION neutrally with the EVENT/Verb. But unlike Bella Coola, when some content other than the EVENT is focused, that content does not invoke the position which is neutrally occupied by the EVENT/Verb. In Mupun, FOCUS is delivered by means of a preceding the constituent to be focused.113 Neither language has contrasting degrees of FOCUS.114 In sum, Mupun is not a FOCUS in situ language, and it seems to support the FOCUS in situ Conjecture.

3.4 Miya

Miya is a member of the Chadic language family ... Miya is spoken in the town of Miya (Miya) and a few nearby farming hamlets in Bauchi State, Nigeria, approximately 70 miles (110km) north of Bauchi ... The people refer to their language as v míy “mouth of miy” ... there are approximately 5,000 speakers of Miya.” (Schuh 1998.1-2).

Miya is of interest here because it shows semantic and morphosyntactic

113 The difference between Bella Coola and Mupun is the difference between having to go to the restaurant for supper and calling for take out. Bella Coola forces one to make the trip, and Mupun allows the food to be delivered to the desired location.

114 This is clearly so of Bella Coola. With respect to one function, the Agent, Mupun has this contrast (Frajzyngier 1993.399-400): (i) Yo a mo sat [REC ASSERTION 3PL say PREP DEM] ‘It was they who said so’ (ii) Yo a mo mo sat [REC ASSERTION 3PL 3PL say PREP DEM] ‘It was they who said so’ The glosses for the two are identical, and there is no comment on distinct usages. It appears, however, that (ii) may actually be composed of two clauses: (iii) A mun mu sat [ASSERTION 1PL 1PL say PREP DEM] ‘It was us who said so’

In (iii) the pronoun mun is “the one that follows the copula,” and mu is “the subject form” (349). This suggests one clause A mun ‘It is us’ and another, Mu sat 72 SYNTAX & SEMANTICS properties found in other Chadic languages, but in a configuration that is somewhat different from those we have so far examined: Hausa, Mupun, Kanakuru, and Pero.115 SVO is a frequent order in Miya utterances (Schuh 1998.282, 295):116

(93) Ndùwya a . [Nduya PF plant millet] ‘Nduya planted millet’

(94) áa baya zúw nuwun [I IMPF take sorghum old.person male my] ‘I will take sorghum to my father’

(95) tsà-yá suw mìr-áy [I.PF give-him TOTALITY1 money-TOTALITY2] ‘I gave him money’

(96) a vár s-áy [horse PF run TOTALITY-TOTALITY] ‘The horse ran away’

“When both direct object and indirect object are present and both are nominal the direct object precedes the indirect object [as in (94)] ... Other nominal constituents such as instruments and locatives follow objects” (Schuh 1998.280). A “neutral (i.e. no questioned or focused nominal or adverbial constituents) declarative statement” will also will also have one of a system of tense-aspect-mood markers (Schuh 1998.281). The TAM markers will appear before the verb as in Hausa, or they are semantically combined with the pronominal expression of the Agent (Schuh 1998.187-188). Sentences (95) and (96) illustrate a Miya option of signalling what Schuh (1998.170-175 et passim) calls “Totality,” which adds “a sense of finality.”117 Totality does not co-occur with Negation, FOCUS (of any constituent function, Agent or non-

115 The grammatical data on Miya are taken entirely from Schuh 1998. There is a list of other work on Miya in Schuh 1998 (pp. 5-6). I have discovered no work on Miya published after 1998.

116 There is a competing order to which we shall return at the end of this section.

117 When “used with a noun alone or a nominalized construction ... [it] give[s] the sense of ‘just’, ‘nothing but’, ‘fina’” (Schuh 1998.175). FOCUS: Nowhere 73

Agent), or yes/no questions. Miya differs from Mupun and Hausa in that two nouns adjacent to each other suffice to express a proposition without another mark of ASSERTION (e.g., the a of Mupun or the nee/cee of Hausa) (Schuh 1998.316, 317):

(97) Ndùwya miy- [Nduya Miya-man] ‘Nduya is a Miya-man’

(98) [sheep] ‘It’s a sheep’

The following illustrate possible Miya wh-questions and their answers (Schuh 1998.333, 334, 335, 336):

(99) (a) W-áa diy-ùws-a?118 [who- follow-him-QUESTION] ‘Who followed him?’

(b) Ndùwya diy- [Nduya FOC.PF follow-him] ‘NDUYA followed him’

(100) (a) Níy wàa jíy tlíwìy-a? [PL who FOC.IMPF buy meat-QUESTION] ‘Who will buy meat?’

(b) Mìy jíy tlíwìy [we FOC.IMPF buy meat] ‘WE will buy meat’

(101) (a) À ? [PF abuse who] ‘Who did he abuse?’

(b) À ? [PF abuse Nduya]

118 “Constituent questions require the sentence final question marker à or wà” (Schuh 1998.331). 74 SYNTAX & SEMANTICS

‘He abused NDUYA’

(102) (a) Fà tsa-yá màa? [you give-him what] ‘What did you give him?’

(b) tsà-yá mìr [I give-him money] ‘I gave him MONEY’

(103) (a) Fà tsaa mír wèe? [you give money who] ‘Who did you give money to?’

(b) tsàa yásu-wan [I give brother-my] ‘I gave (it) to MY BROTHER’

(104) (a) Fà naf aa màa? [you bind mat our with what] ‘What did you bind your mat with?’

(b) aa [I bind with palm.fronds] ‘I bound (it) WITH PALM FRONDS’

When the Miya Agent is questioned, or when the Agent responds as an answer to a wh-question, it holds the sentence-initial position, but a tense-aspect- modality appropriate to signal FOCUS replaces the non-FOCUS TAM. “When subjects are questioned or focused, the only possible TAM’s are Perfective and Imperfective.” And “Perfective and Imperfective have special auxiliaries when subjects are questioned or focused” (Schuh 1998.331).119 In sentences such as (105a) and (105b), a form jíy is used (Schuh 1998.229):

(105) (a) Wàa jíy malv-áa? [who FOCUS chief-QUESTION]

119 Also “Questioned and focused subjects require special TAM’s. It is the TAM which shows that a subject is focused — there are no word order differences or other overt markers distinguishing neutral sentences from those with a focused subject” (Schuh 1998.332-333). FOCUS: Nowhere 75

‘Who was chief?’

(b) Ròoya jíy

As expected, FOCUS has uses beyond responding to wh-questions (Schuh 1998.332, 346):

(106) (a) Ná dúw, “ faarà [this.one QUOTATION1 FOC.PF be.first buwáhìy-uwun!” coming-ICP]120 ‘This one says, “I came first!”

(b) Ná dúw, “ faarà [this.one QUOTATION1 FOC.PF be.first buwáhìy-uwun!” coming-ICP] ‘[Then] This one says, “I came first!”

(107) ... pay t-aaMángìla, [ now pond of-Mangila yeah of Gwarama tá aaDáya sabòoda Gwarama of people of Daya because Gwarama bàa-za FOC.PF dig.out-it] ‘... now the pond of Mangila, yeah, it (belongs to) Gwarama, that is, to the people of Daya, because GWARAMA dug it out’

Schuh (1998.346) describes the context which has supported FOCUS on the second occurrence of Gwarama:

This [sentence (107)] is from a passage describing a depression high on Mangila, one of the two main inselbergs at Miya town, which was cleared out in order to form a natural cistern at the time when the people took refuge from slave raids on top of the inselbergs. Gwàrama, one of the first settlers at Miya and the founder of the Daya section of Miya town, has contrastive focus because he first cleared the pond even though Mangila inselberg itself is named after one of the other

120 ICP = “Intransitive Copy Pronoun.” 76 SYNTAX & SEMANTICS

founders of Miya.

When a constituent whose function is signaled by a post-verbal position is questioned or when it functions as an answer, that constituent, too, like the Agent holds its position; but it is recognized as focused by the absence of TOTALITY. Compare (Schuh 1998.334, 336):

(108) (a) À már suw zhàak-áy [he get TOTALITY donkey-TOTALITY] ‘He got a donkey’

(b) À már [he get donkey] ‘He got a DONKEY’

(109) (a) Míy zà-ma s-áakan hà [we ented-ICP TOTALITY-house of Róoy-ay Roya-TOTALITY] ‘We entered Roya’s house’

(b) Míy zà-m(a) áakan hà Róoyà [we ented-ICP house of Roya] ‘We entered ROYA’S HOUSE’

“Nominal objects in focus have no overt marking. Rather it is the absence of the Totality construction ... which shows an object to be in focus ... As with objects absence of the Totality construction ... indicates focus of a locative” (Schuh 1998.334, 336). “When a constituent other than the subject is questioned or focused in the Perfective, the questioned/focused constituent is in situ and subject pronouns, auxiliaries, and verbs are the same as ... [In affirmative clauses without questioned or focused constituents], but the Totality extension is absent” (Schuh 1998.123). “The Totality construction never appears when an object is questioned or when the context of a sentence presented for translation shows that the object is focused” (Schuh 1998.334). If we return to (93) and (94) above, the absence of TOTALITY suggests that FOCUS is present in them in some post-verbal position. Closer examination of (94), shows that the application of FOCUS must be vague. Is the FOCUS associated with zúw FOCUS: Nowhere 77

‘sorghum or with nuwun ‘my father’, or with both ... or neither, i.e., the entire clause?121 Lacking the context, we cannot tell.122 The use of “special TAM’s” to mark FOCUS in combination with an Agent is the Hausa ex situ pattern (section 3.2). Miya lacks a comparable ex situ pattern of FOCUS for the other functions, and relies instead on context to determine FOCUS. FOCUS of non-Agents then recalls Hausa in situ FOCUS, which similarly depended upon context for its recognition. There is a second implementation of FOCUS that relies on the syntax of ASSERTION in (97) and (105b).123 In this formation, the focused content is placed initially and followed by the FOCUS element jíy. The remainder of the propositional content is formed into a Relative Clause to express content of which is ASSERTED (Schuh 1998.343, 344):

121 Schuh (1998.282) comments on these, “Perfective clauses with overt nominal complements to the verb are generally volunteered without the Totality construction because such complements tend to be in pragmatic focus.”

122 “Note that in sentences which also contain objects, it is impossible to tell whether the object or the locative is focused without context” (Schuh 1998.336). Rather than admit to the vagueness inherent in the expression in SVO utterances without S FOCUS and without TOTALITY, Schuh (1998.282, 334) suggests that FOCUS is always present: Perfective clauses with overt nominal complements to the verb are generally volunteered without the Totality constructions because such complements tend to be in pragmatic focus. In elicitation, speakers usually volunteer sentences without the Totality construction when they contain nominal objects ..., especially in the Perfective. This is the case even when nothing in the elicitation context suggests object focus. I interpret this as being a result of the universal tendency to introduce new discourse referents as objects ... Because of this tendency in discourse, speakers sense that nominal objects have inherent pragmatic focus ... In short, an object in a sentence without the Totality construction seems always to bear some sort of focus, which may range from a tendency to draw attention to nominal objects as new referents to strong contrastive focus. Such a conclusion immediately causes a problem with utterances which have postverbal Agents. Furthermore, reaching the conclusion in this way, reasoning through “new discourse referents”, depends upon the chancey association/identification of the semantics of participants introduced to the context (say, “new”) with the semantics of FOCUS. In the context of current descriptive practice in Hausa, I think one would simply admit now that you cannot always tell what the speaker intended. The morphosyntax of FOCUS can be indeterminate.

123 If it is correct that FOCUS is initial in (110) - (113), and if it is correct to see miy- miy- 78 SYNTAX & SEMANTICS

(110) jíy aa wàshsham [I FOCUS one.who PF exceed-them with years] ‘I am the one who has spent longer than (any of) them’

(111) T jíy ba faarà zahiya-yá [he FOCUS one.who be.first make.into-him ta Miyà leadership of Miya] ‘HE was the one who was first made the leader of the Miya’

(112) T jíy ba na d-àa bíy [he FOCUS one.who AUX -IMPF PRT tsa-yà give-him leadership] ‘HE was the one to whom they gave the leadership’

(113) Bá faarà bíyak àabíy jiy Gwàrama [one.who be.first dig.out water FOCUS Gwarama] ‘THE ONE WHO FIRST DUG OUT WATER was Gwarama’

Writing about the semantic contrasts between the bipartite FOCUS in (110) - (111) and the previous sentential FOCUS, Schuh (1998.344):

... there is a fuzzy pragmatic division in the contexts for their use. Only the subject focus construction would be pragmatically available in response to a constituent question involving the subject ..., whereas pseudo-cleft [i.e. the paradigmatic] seems like the only felicitous structure in a case like [(110) above] ....

Miya FOCUS appears to have two configurations. One is constituted within the frame of PROPOSITIONS identifying which, if any, of the components of the PROPOSITION is to be singled out in the context of its occurrence. The second FOCUS is constituted in terms of selecting pieces of content with the same propositional function that might have been alternatively selected had this one not been. The first is the syntagmatic FOCUS (from Chapter 10 and above); the second is the pointed paradigmatic FOCUS. Paradigmatic FOCUS is the more “focused” FOCUS, and hence, in terms of Figure 7, the more PARTICULAR and more INTENSE FOCUS. FOCUS: Nowhere 79

INTENSE LAX FOCUS FOCUS

GLOBAL FOCUS SVO without FOCUS I

SVO with FOCUS II PARTICULAR FOCUS Paradigmataic FOCUS III

Figure 7: An Organization of FOCUS in Miya.

“Focused constituents in Miya are in situ” (Schuh 1998.345).124 Although Miya exhibits properties of a FOCUS in situ language, but it is not the case that FOCUS is expressed entirely in situ. For at least two reasons. First, if we compare the focused Agent in Miya with the focused Agent in Hausa, the grammatical mechanisms they use are identical. Both occupy the sentence- initial position and both are identified as focused by means of the choice of TAM that accompanies them. In Hausa, that is recognized as an ex situ FOCUS even though no alternate position (i.e., “movement”) is invoked. The consensus is that there is no in situ FOCUS of Hausa Agents. What is good for Hausa should be good for Miya. Second, Miya possesses an additional FOCUS which exploits a paradigmatic syntax for which the issue of in situ expression is irrelevant. Miya underscores the importance of recognizing that FOCUS in situ does not turn on simply noting the absence of order in signalling FOCUS. The initial FOCUS in situ Conjecture proposed (i) that a FOCUS in situ language would contain at least one additional non-in situ expression of FOCUS and (ii) that the semantics of ASSERTION would not be part of the composite that made up FOCUS. While Miya seems to exhibit in situ FOCUS with respect to the non- Agent functions, the Agent exhibits an explicit grammatical mark (the choice of TAM) that distinguishes its combination with FOCUS from the non-Agents’ combination with FOCUS. The Miya Agent then parallels the Agent in Hausa in being distinguished by an appropriate FOCUS choice of the TAM. Thus, only the other postverbal material is left to show a true in situ expression of FOCUS.125

124 “Miya forms constituent questions and constituent focus in situ, i.e. the questioned or focused constituent occupies the linear position that the corresponding constituent would occupy in a declarative utterance” (Schuh 1998.331).

125 We have ignored, to this point, a large portion of Miya morphosyntax. Schuh’s 80 SYNTAX & SEMANTICS

(1998.280) reluctance to call Miya an SVO language stems from the presence of utterances in which the S is postverbal and which are paired with utterances with preverbal S’s (Schuh 1998.282): (i) (a) [Nd kay PF grind pepper] ‘Nk kay ground pepper’

(b) [PF grind pepper PS Nd kay] ‘Nk kay ground pepper’ (c) À [PF grind TOTALITY pepper-TOTALITY PS Nd kay] ‘Nk kay ground pepper’ (ii) (a) Ghàduw s-ay [wood dry TOTALITY-TOTALITY] ‘The wood dried up’ (b) s-a ghàduw [PF dry PS-TOTALITY wood] ‘The wood dried up’

“A postverbal subject is obligatorily preceded by a postverbal subject marker (PS), àà.” (Schuh 1998.281). Postverbal Agents do not occur in utterances with questioned or focused content (Schuh 1998.287, 288, 291, 333). This observation appears accurate, but it contradicts Schuh’s earlier claim (1998.334) that “an object in a sentence without the Totality construction seems always to bear some sort of focus.” Sentence (ib) has no Totality, but because of the postposed Agent, it must also have no FOCUS. We have to conclude that the absence of Totality does not guarantee the presence of FOCUS. “In some clause types, both orderings are possible with no clear difference in meaning; in some clause types, only postverbal order is possible; and in some clause types, only preverbal order is possible ...In my original fieldwork, I did not check alternative orders in main clauses, but in a brief visit to Miya in 1996 ... I looked specifically at this issue with a variety of sentences with nominal subjects. Though Vaziya [Schuh’s Miya language consultant] first volunteered SVX order, he readily accepted as quite normal the alternatives with VXS order in the TAM’s that I checked, affirmative and negative, with and without Totality marking” (Schuh 1998.281). Schuh (281) proposes a tentative interpretation that turns on TOPIC-COMMENT but does not “pursue the issue.” As an outsider to Chadic, I look at this and think “switch reference”. Switch reference would obvously explain why only the Agent function is involved in whatever is happening here. It would explain the ease with which a native speaker accepts both SVO & VXS in the isolated context of elicitation. It could explain why VXS conflicts with FOCUS. It might also explain those places in which postverbal Agents must appear (Schuh 1998.286-287). The semantics of switch reference is known to encompass more than just switching referents of Subjects. Although there is much Miya textual material (Schuh 1998.9-11), to date no texts have been published. While it remains uncertain precisely what semantic contrast is signaled by choice beween preverbal and postverbal Agent, it does seem clear that it does not directly impact FOCUS. And for that reason, I have set it aside as another issue. FOCUS: Nowhere 81

4. In Situ Languages that are SOV

[Version: April 20, 2009]