On the External and Internal Syntax of Menominee Negation
MEREDITH JOHNSON AND BRYAN ROSEN University of Wisconsin–Madison
INTRODUCTION1
The purpose of this paper is to account for the syntax and morphosyntax of Menominee negation. Menominee is unique among the Algonquian lan- guages because of the presence of a separate NEGATIVE order, exempli ed in (1) below. The negative order is characterized both by the negative particle kan and the negative suf x -an.2
(1) a. Kan nekÓs-nëminan Kan nae-kÓs-nÓmi-n-an NEG 1-PST-dance.AI-AN-NEG ‘I didn’t dance.’ (Macaulay 2003:220)
1. We are grateful to Marie Floring (MF), the late Margaret King (MK), Larry Tomow (LT), and the late Lavina Shawano (LS) for providing the data used in this paper. We would also like to thank Monica Macaulay and two anonymous PAC reviewers for valu- able feedback and comments on this paper; and the Menominee Language and Culture Commission for their guidance. 2. The following abbreviations are used: 1, 2, 3 = rst, second, third person; AI = animate intransitive; AN = animate; AOR = aorist; CONJ = conjunct; CTR = contrast; EPIS = epistemic; EX = exclusive; FUT = future; HAB = habitual; II = inanimate intransitive; IN = inclusive; INAN = inanimate; INC = incompletive; LCL = local; NEG = negation; OBV = obviative; PL = plural; POSS = possessive; PROX = proximate; PST = past; TA = transitive animate; TH = theme sign; TI = transitive inanimate. ߧEߨ represents a morphophoneme used by Bloom eld (1962) that usually surfaces as [e]. The following source codes represent stories from Bloom eld 1928: BOM = Birth of Maeqnapos; BOY = A Boy is Blessed by Mosquitos; CRCW = The Childhood of Red Cloud Woman; HIC = How to Stop Hiccoughs; MIF = The Menomini Indian Fair; MW = How a Menominee Woman Earns Money; NTM = Nehtsiwihtuk as a Trencherman; RS = Red Swan. BM is the code for Bead Man, which is a story found in Bloom eld’s notes in the National Anthropological Archives at the Smithsonian Institution, translated and analyzed by Monica Macaulay and Marianne Milligan.
174 ON THE EXTERNAL AND INTERNAL SYNTAX OF MENOMINEE NEGATION 175
b. WÁpanow kan okÓs-aw n nan ot hsaeh WÁpanow kan o-kÓs-awÁN-Á-n-an o-ëhsaehs Waubano NEG 3-PST-take.away.with.TA-TH-AN-NEG 3-dog ‘Waubano didn’t walk his dog.’ (Macaulay 2003:221)
Based on both syntactic and morphosyntactic evidence, we argue that nega- tion in Menominee is situated between CP and IP/TP. An important conse- quence is that the Menominee verb does not sit in C0, contra what others have previously argued for in Algonquian languages. This paper is structured as follows: the next section provides an overview of Menominee negation, based on Macaulay 2003. In the section after that, we provide evidence for our claim that negation sits between CP and IP/TP. Then, we argue that the morphosyntax of the negative order in Menominee supports our analysis, and discuss the relationship between negation and agreement. The last section concludes the paper.
DESCRIPTION OF MENOMINEE NEGATION
Before introducing our analysis, we rst provide an overview of the nega- tion system in Menominee. As shown in (1) above, Menominee has a unique negative order marked by the suf x -an. Macaulay (2003) shows that the negative order exhibits three major differences in in ection from the other orders. First, the third person pre x o- is unique to the negative order (see example (1b) above). In the independent order, the third person pre x is null, as seen in (2):
(2) n nၤw n n-ၤ-w fetch.TA-TH-3 ‘He fetches him.’ (Bloom eld 1962:152)
Second, the person agreement suf xes are impoverished: -w is used to index the subject of II verbs (as in (3)), while -n marks the rst, second and third person animate arguments of AI, TA and TI verbs (as in (4)–(6)).3 As we
3. Example (5) contains the preverb aw- which surfaces as naw- when preceded by a person pre x (Bloom eld 1962:216). 176 MEREDITH JOHNSON AND BRYAN ROSEN discuss in more detail later in the paper, independent and conjunct order agreement suf xes distinguish both gender and number.
(3) II verb ahkanom wenah kan omas k s-tak n ahkanom wenah kan omas k s-takw-w-an long.ago it NEG here PST-exist.II-INAN-NEG ‘It wasn’t here, long ago.’ (MW; Macaulay’s (12))
(4) AI verb kan kၤkčh ukiw-iskihsinenan kan kၤkčh o-kew-aeskihsenၤ-n-an NEG something 3-HAB-be.left.over.AI-AN-NEG ‘He never has anything left.’ (MIF 4; Macaulay’s (22))
(5) TA verb nah w, kan taeh kina-Oၤwinan! nah w, kan taeh kae-aw-Oၤw-e-n-an well.then NEG and 2-INC-see.TA-TH-AN-NEG ‘Well, then, you won’t see me!’ (RS 167; Macaulay’s (14))
(6) TI verb kan nepëtčnan kan nae-p t- -n-an NEG 1-bring.TI-TH-AN-NEG ‘I do not bring it.’ (Bloom eld 1962:173; Macaulay’s (15))
Third, the obviative marker in the negative order (7a) is distinct from both independent order (7b) and conjunct order (7c). Consider the examples below with the obviative marker in bold:
(7) a. kan opčsenenÓnan kan o-pčse-n-enÓn-an NEG 3-embark.AI-AN-OBV-NEG ‘S/he (obv) does not embark’ (Bloom eld 1962:168; Macaulay’s (16))
b. pčsewan pčse-w-an embark.AI-3-OBV ‘S/he (obv) embarks’ (Bloom eld 1962:150; Macaulay’s (17)) ON THE EXTERNAL AND INTERNAL SYNTAX OF MENOMINEE NEGATION 177
c. pčsenet pčse-aene-t embark.AI-OBV-3.CONJ ‘S/he (obv) embarks’ (Bloom eld 1962:224; Macaulay’s (18))
Lastly, there are two negators in Menominee: kan and kat. Kan is used in both negative order and conjunct order, while kat is used only in conjunct order. Neither negator can be used with independent order. Macaulay (2003) points out that the two negators have distinct uses: kan is the “unmarked” negator, and kat is restricted to mirative contexts. The examples in (8) illustrate the uses of kan and kat, respectively.
(8) a. eneq ၤc kan wenah cew-kÓs-esÓqtat eneq aese-t kan wenah cew-kÓs-aeseqta-t it.is.that IC.say.AI-3.CONJ NEG she EPIS-PST-do.so.AI-3.CONJ ‘She said that she didn’t do it.’ (Macaulay 2003:227)
b. kat akekoh as Áwetuaq këqsÓhsak kat akekoh as Áwe-t-wÁw këqsÓhs-ak NEG these.AN AOR be.AI-3-3PL.CONJ girl-AN.PL ‘(Then all at once he knew) it was not those girls.’ (BM 84; Macaulay’s (37))
In sum, we have seen that Menominee exhibits a unique negative order and has two negators, kan and kat. In the next sections, we provide a formal analysis of these data.
EXTERNAL SYNTAX
Position of Negator
We analyze kan and kat as negators that occupy Neg0 of NegP. We argue that NegP is between CP and IP/TP. This is consistent with previous analyses of the syntax of negation (see, e.g., Belletti 1990; Laka 1990; Rivero 1994; Haegeman and Zanuttini 1996; and Wiltschko 2002). We argue that NegP is below CP because kan and kat are below elements that are in the CP domain. For example, negation is preceded by the complementizer kÓspen ‘if’ as shown in (9a) for kat and (9b) for kan: 178 MEREDITH JOHNSON AND BRYAN ROSEN
(9) a. kÓspen kat m c yan kÓspen kat m cy -an if NEG leave.AI-LCL.CONJ ‘If you don’t leave’ (Macaulay 2003:226)
b. kÓspen kan mÓset nekiah kÓspen kan mÓN-e-t nae-kyÁhs if NEG give.TA-TH-3.CONJ 1-mother ‘If my mother does not give me them’ (Macaulay 2003:229)
KÓspen has both the syntactic and semantic properties associated with com- plementizers crosslinguistically: it always appears in clause-initial position, and it appears to have the same semantic function as English if (McCloskey 1996; Branigan and MacKenzie 2002). The possibility of the co-occurrence of a complementizer-verb sequence is crosslinguistically used as a diagnos- tic for the position of the verb. For example, den Besten (1983) argues that V2 word order in German is derived by V-to-C movement since the verb does not raise in embedded clauses due to the presence of a complementizer. Furthermore, Diesing (1990) and Thráinsson (1985) argue for Yiddish and Icelandic, respectively, that the verb does not move to C, as V2 is found in both matrix and embedded clauses. A CP-recursion analysis is also not desirable for Menominee: Iatridou (1991) and Iatridou and Kroch (1992) argue that CP-recursion is not possible when the verb takes an irrealis complement. Clauses embedded by the complementizer if certainly seem to qualify as irrealis. We adopt the proposal of Johnson et al. (2015), who argue that the left periphery in Menominee consists of two topic phrases and a focus phrase. This is schematized in (10), below. According to Johnson et al., the highest topic (the “external topic”) can be separated from the rest of the clause with a pause, and it is not necessarily an argument of the verb. As for focus, Johnson et al. follow Aissen (1992), who says “The semantics of the focus construction has two essential parts, a presupposition and an asser- tion . . . What the focus construction asserts is that the focused constituent denotes an entity which satis es the variable of the presupposition” (p. 50).
(10) [ExtTopP … [FocP … [IntTopP …]]]
In Menominee, external topics and focused elements also precede negation. The noun phrase metÁtah mÓqsemenak ‘ten apples’ in (11) exem- ON THE EXTERNAL AND INTERNAL SYNTAX OF MENOMINEE NEGATION 179 pli es an external topic. As de ned by Johnson et al., external topics do not have to correspond to an argument of the verb.4
(11) MetÁtah mÓqsemenak, (nëw) kan (nëw) awၤtokÓhsewak. MaetÁtah mÓqsemen-ak (nëw) kan (nëw) awၤtokÓhsÓwe-w-ak ten apple-AN.PL (four) NEG (four) be.wormy.AI-3-3PL ‘Of the ten apples, four aren’t wormy.’ (MF/MK/LT 9/23/12)
The following examples all show that negation follows focused items. In (12) and (13), both kan and kat follow wh-words, which we assume to be in a focus position (Horvath 1986; Rizzi 1997).
(12) 8ၤkiq kan kew-mÓcek? Wၤkiq kan kew-mÓc-E-k what NEG HAB-eat.TI-TH-3.CONJ ‘What doesn’t he eat?’ (MF/LS 04/15/05)
(13) Wၤkiq kat kew-mÓcek? Wၤkiq kat kew-mÓc-E-k what NEG HAB-eat.TI-TH-3.CONJ ‘What doesn’t he eat?’ (MF/LS 04/15/05)
Negation follows restrictive NPs in (14) and inde nites in (15), which Johnson et al. claim also sit in focus position in Menominee:
(14) Nenah ap new kan kÓs-nëminan. Nenah ap new kan kÓs-nÓmi-n-an. I also CTR NEG PST-dance.AI-AN-NEG ‘I also didn’t dance last night.’ (MF/MK/LT 9/23/12)
(15) Weyak kan okÓs-menၤnan okahpÓm. Weyak kan o-kÓs-maenae-n-an o-kahpÓw-m someone NEG 3-PST-drink.AI-AN-NEG 3-coffee-POSS ‘Someone didn’t drink their coffee.’ (MF/MK/LT 9/23/12)
Lastly, negation follows eneq (also argued by Johnson et al. to be a focus element), as in (16):
4. As (11) shows, occasionally the modern-day speakers use independent order with kan. 180 MEREDITH JOHNSON AND BRYAN ROSEN
(16) Eneq-peh taeh Monica kan okÓs-pianan. Eneq-peh taeh Monica kan o-kÓs-pyÁ-n-an then-time and Monica NEG 3-PST-come.AI-AN-NEG ‘And then Monica didn’t come.’ (MF/MK/LT 9/23/12)
We propose that NegP is higher than IP/TP since negation obligatorily precedes the verb, and thus the preverbs that encode tense, aspect and mood. (17a) shows an example with the habitual preverb kew-, while (17b) and (c) illustrate the position of the future and past tense preverbs, respectively.
(17) a. kan ’s kew-m t kamiwၤhneh kan ’s kew-m t kam wၤhnaen-k NEG AOR HAB-splash.with.sound.of.moving.water.II-3.CONJ ‘The water wasn’t making any sound.’ (BOM 036)
b. kan nekčtoh kena-kënonaehÁnan kemëcehswan kan naekotoh kae-aw-k nonၤh-Á-n-an kae-mÓcehswan NEG ever 2-FUT-seek.in.vain.TI-TH-AN-NEG 2-food ‘You will never lack for food.’ (BOY 061)
c. tepၤh ၤhkwah kan ok s-nep nan tepၤh ahkuat-k kan o-kÓs-naepÁ-n-an at.night IC.all.long.II-3.CONJ NEG 3-PST-sleep.AI-AN-NEG ‘He had not slept all night.’ (NTM 118)
Our analysis is consistent with Lochbihler and Mathieu’s (2008/2009) analysis of the verb position in Ojibwe: they propose that preverbs sit in T0. They further note that the fact that tense precedes the verb cannot be accounted for if the verb moves to C0, as proposed in some accounts of Algonquian syntax (Halle and Marantz 1993; Campana 1996; Brittain 2001; Richards 2004). Our analysis departs from previous generativist accounts of Algon- quian negation that have proposed that negation is base-generated in the CP domain. Brittain (1996) proposes an analysis to account for the two negators in Innu-aimun: apû and ekâ. Both negators in Innu-aimun require the verb to bear conjunct morphology. Brittain places both negators in the CP domain: she argues that apû sits in Spec,CP and ekâ in the upper C0 of ON THE EXTERNAL AND INTERNAL SYNTAX OF MENOMINEE NEGATION 181 a double CP structure. This type of analysis is necessary under the assump- tion that the verb moves to C0 in conjunct order. Brittain’s analysis cannot straightforwardly be extended to Menomi- nee data. If both negation and the verb occupied C0, it would be necessary to posit three CPs to account for the examples in (9). This is illustrated in (18).
(18)
Furthermore, arguments can appear between the complementizer and the negator, as in (19). This is unexpected under Brittain’s account, as she is explicit that the embedded CP has no speci er position. We assume that weyak ‘someone’ sits in one of the left-peripheral positions proposed by Johnson et al.
(19) Këspen weyak kat ’s pīnëtah ’s onuawၤt. KÓspen weyak kat ’s pīnët-am-k ’s if someone NEG AOR stop.making.vocal.noise.TI-TH-3.CONJ AOR
onuawၤ-t hiccup-3.CONJ ‘If someone does not stop hiccoughing.’ (HIC 001)
As mentioned above, CP-recursion does not occur with irrealis com- plements, which would rule out a structure like (18). Instead, we propose a single CP in which the complementizer is in C0, the negators kan and kat are in Neg0 and the verb is in T0. Our analysis of example (9a) is schematized in (20). 182 MEREDITH JOHNSON AND BRYAN ROSEN
(20)
An anonymous reviewer suggests an alternative analysis of Menomi- nee negation. Consider again the example in (6) (repeated in (21)).
(21) kan nepëtčnan kan nae-pët-č-n-an NEG 1-bring.TI-TH-AN-NEG ‘I do not bring it.’ (Bloom eld 1962:173; Macaulay’s (15))
They propose that the verbal morphology -an represents Neg0, while kan/kat is merged in Spec,NegP. Under this analysis, the verb moves into Neg0 to pick up -an in the negative order and a null allomorph in conjunct order. This analysis is presented in (22), ignoring theme signs and other agreement morphemes.
(22) a. Underlying structure of (21)
[NegP kan [-an] [VP OBFQëUčO]] NEG NEG I bring it
b. Movement of V0 to Neg0 [ kan [ [ ] -an] [ t ]] NegP Neg V OBFQëUčO VP OBFQëUčO NEG I bring it NEG
However, this alternative analysis cannot account for many of the word order facts described above. Many elements can intervene between kan/kat and the verb, including argument NPs ((4), (8a)), adjuncts ((3), (17b)) and demonstratives (8b). This is unexpected if kan/kat is the speci er to the head that contains the verb: it is not clear where such elements could sit. ON THE EXTERNAL AND INTERNAL SYNTAX OF MENOMINEE NEGATION 183
In contrast, under our analysis, kan and kat sit in the Neg0 head while the verb does not move past T0; thus, there are positions between the negator and the verb for these intervening elements. In conclusion, our analysis can account for the following generaliza- tions about the syntactic positioning of negation in Menominee: (i) nega- tion is below CP, as the negator follows complementizers and elements in topic or focus position; (ii) negation is above TP, as it precedes preverbs that encode tense; and (iii) a CP-recursion analysis is not consistent with the semantics of the clause. In the next few sections, we show how the structure proposed in (20) applies to other aspects of Menominee negation.
Conjunct Order and Negation Here we address the connections between negation, conjunct order, and mirativity. In order to derive the difference between kan and kat with con- junct order, we appeal to the operation Agree as in Pesetsky and Torrego (P&T; 2004). P&T argue feature interpretability should be separated from feature values. This is represented in (23) with the four feature types:
(23) a. iF val interpretable, valued b. uF [ ] uninterpretable, unvalued c. iF [ ] interpretable, unvalued d. uF val uninterpretable, valued
P&T’s version of Agree is as follows:
(24) Agree (Feature sharing version) (P&T 2004)
i. An unvalued feature F (a probe) on a head H at syntactic location (F ) scans its c-command domain for another instance of F (a goal) at location