ACAL 45 University of Kansas

The exceptional properties of the 1sg and reflexive object markers in Bantu: syntax, phonology, or both?

Michael R. Marlo University of Missouri

1. Introduction 1.2 OP combinatorics. This talk focuses on three phenomena in where the 1sg OP behaves differently from other OPs: (2) In Khayo, 1sg N‑ can co-occur with a second OP (Marlo 2009: 92). a. a‑ ‑m ‑ ‑ (/‑β ‑/) (i) Reduplication 3sg-3sg-1sg[see-appl-fv] (ii) OP combinatorics ‘he sees him for me’ (iii) Allomorphy of the final vowel in the imperative b. Other combinations of OPs are ungrammatical. 1.1 Reduplication. -βa ‑ [...... 3sg-3pl[...... 3sg-refl[... (1) In Saamia, CV- prefixes are not generally overcopied in the reduplica- tion of C-initial stems. (Botne et al. 2006, Marlo 2002, 2004) 1.3 Allomorphy of the final vowel in the imperative. a. [ ][ - ] βa[ ][ - - ] inf[red][chop-fv] 3pl[red][bite-rec-fv] (3) Rundi exhibits a common pattern of final vowel allomorphy in the im- ‘to cut into pieces’ ‘they are biting each other continually’ perative in which the 1sg behaves differently from other OPs (Ndayiragije 2003: 183-185). b. Prefixes, including OPs, are not copied with V-initial stems. a. Bare imperatives in end in ‑a. [ ][ - ] y[ ][ - ] [vun-a a a a ‘break the bone!’ inf[red][sing-fv] 3sg [red][warm-fv] ‘to sing repeatedly’ ‘he is warming himself repeatedly’ b. Verbs with an OP end in ‑e. tu[vun-e] ‘break us!’ Ø-mw - - - - (/-ixal-/) i[vun-e] ‘break yourself!’ 1sg-3pl[red][sing-appl-hab-fv] 3sg[red][sit-fv] ‘I am always singing for him’ ‘he is sitting down repeatedly’ c. However, verbs with the 1sg OP end in ‑a. m[peb-a] (/N[heb-a]/) ‘abandon me!’ c. However, the nasal 1sg OP is overcopied in reduplication. -ɲ[ ][ɲ- - - ] (/-imb-/) (4) There are different factors that could be responsible for the divergent 3sg-1sg[red][1sg-sing-appl-fv] behavior of the 1sg OP compared to other OPs. ‘he is singing for me continuously’ a. Phonological Most OPs have a CV- shape, but 1sg has monophone N-. -a -n[aa ][n- - ] (/-laang-/) 3sg-pst-1sg[red][1sg-call-fv] ‘he called me repeatedly’

April 17 – 19, 2014 1 / 15 b. Syntactic cl. 3 - cl. 4 - The 1sg OP is often required to surface closest to the verb stem cl. 5 - cl. 6 - (Polak 1983: 297) and may be in a different morpho-syntactic posi- cl. 7 - cl. 8 - tions from other OPs (Buell 2005, Muriungi 2008). cl. 9 - cl. 10 - cl. 11 - cl. 12 - In the rest of this talk, I present micro-typological surveys on overcopying of cl. 15 - cl. 14 - prefixes in reduplication, the number of OPs that are allowed, and the allomorphy of the final vowel in the imperative, and argue that both morpho- 1sg ‑ assimilates in place of articulation to the following consonant and syntactic and morpho-phonological factors are responsible for the unique triggers cross-linguistically common NC effects (Hyman 2001, 2003, Odden behavior of the 1sg OP. to appear).

2. Background on OPs Other OPs are generally not known to trigger similar effects, though reflexive i‑ does trigger mutations of the following stem-initial segment in OPs sit in a position immediately preceding the stem (or reduplicated stem). Tswana (Dickens 1984, Mmusi 1992: 34-35). The OPs and the (reduplicated) stem form the ‘macrostem’. 2.2. The number of OPs. Some Bantu languages in the northwest have no (5) Inflectional prefixes macrostem{OP [RED] stem[Root-Extensions-FV]} OPs, and a large number of other languages are reported to limit the number of OPs to just one. ‘OP’ is used here in a theory-neutral way here to indicate the object-indexing morphemes that precede the verb stem.1 I also set aside post-stem object (8) Only a single OP is allowed in Swahili, representing the beneficiary marking.2 (Marten & Kula 2012: 244). a. ni-li-m[p-a] 2.1. Phonological properties of OPs. OPs generally have an CV- shape in 1sg-pst-3sg[give-fv] Bantu verbs. ‘I gave him (it)’

(6) Two common exceptions b. It is ungrammatical to mark both objects with OPs. a. 1sg, reconstructed as ‑ (Meeussen 1967: 98, 1986: 373-374) *ni-li-i-m[p-a] *ni-li-m-i[p-a] b. reflexive ‑ (Polak 1983). 1sg-pst-9-3sg[give-fv] 1sg-pst-3sg-9[give-fv] Int.: ‘I gave him it’ Int.: ‘I gave him it’ (7) OPs of Nyaturu (Olson 1964, Schadeberg 1979) reflexive i- (9) Languages that allow only one OP 1sg N- 1pl - Bende (Abe 2006: 179-180) 2sg - 2pl - Chewa (Bresnan & Mchombo 1987: 743, Marten et al. 2007: 263) 3sg - 3pl - Herero (Marten et al. 2007: 263, hlig avari : -170) Kagulu (Petzell 2008: 102-103) 1 I set aside the question of whether these morphemes are affixes (‘infixes’ or ‘prefixes’), Kalanga (Rose Letsholo, p.c.) clitics, pronouns, or agreement markers (aka ‘concords’; see Adams 2010, Baker 2008, Kete (Kamba Muzenga 1980: 130-132) Bax & Diercks 2012, Bearth 2003, Bresnan & Mchombo 1987, Buell 2005, Byarushengo Kwezo (Forges 1983: 260-261) et al. 1976, Demuth & Johnson 1990, Diercks & Sikuku 2011, Duranti & Byarushengo Langi (Nurse 2007) 1977, Henderson 2006, Mchombo 2004, Riedel 2009, van der Spuy 1993, Woolford Beya Lega (Botne 2003: 440, 445) 1999, 2000, 2001, Zeller 2012, Zerbian 2006). Maore Comorian (Alnet 2009: 269) 2 See Beaudoin-Lietz et al. (2004) and Marlo (2013). 2 / 15 Mbala (Ndolo 1972: 25) 2.3. The order of OPs. In some Bantu languages, the order of OPs is deter- Mbonge (Friesen 2002: 107-117) mined by syntactic relations or thematic roles, often a mirror image of the Meru (Hodges 1977: 111) order of postverbal NPs (Bearth 2003), as in Tswana (Creissels 2002: 391). Mwani (Nurse 2007) Phuthi (Donnelly 2009: 352) (13) In Sambaa, the order of the OPs follows from their grammatical roles, Sotho (Morolong & Hyman 1977: 210) the indirect object (goal) being closest to the stem (Riedel 2009: 140). Swahili (Marten et al. 2007: 263) a. a-za-ku-ni[ony-esh-a] b. a-za-ni-ku[ony-esh-a] Swati (De Guzman 1987: 312) 3sg-perf-2sg-1sg[see-caus-fv] 3sg-perf-1sg-2sg[see-caus-fv] Tonga (Nkolola 2004: 181) ‘s/he pointed you out to me’ ‘s/he pointed me out to you’ Tura (Marlo 2008: 163) Zulu (Adams 2010: 24, Zeller 2012: 219-220) In languages that allow more than one OP, the 1sg and reflexive prefixes are often required to surface finally, in a position linearly adjacent to the stem Among languages that allow multiple OPs, there is considerable diversity in (Polak 1983: 297). the exact number of OPs that are allowed. (14) In Ruri, syntactic relations normally determine the order of multiple Some freely allow two, three, four, up to a maximum of five or even six OPs OPs, as in Sambaa. However, the 1sg OP must always appear closest to (Bearth 2003: 124-125, Kearns 2008, Marlo 2013, Marten et al. 2007, the stem, irrespective of its grammatical role (Massamba 1982: 52). Marten & Kula 2012). It has been claimed for some Bantu languages that the inner placement of the (10) Five OPs in Kuria (Rodrigo Ranero p.c.) 1sg OP (as well as the reflexive) reflects the influence of animacy-topicality n-aa-chi-mo-ge-ge-ba[haa-yey-e] and person-number hierarchies on the order of the OPs (Alsina 1994, 1sg-pst-10-1-9-9-2[give-caus.inst.ben-fv] Contini-Morava 1983, Duranti 1979, Rugemalira 1993). ‘I made them (the girls) give it (the money) to it (Pomona) using it (M-Pesa) on behalf of him (Mokami).’ 1sg vs. reflexive Meeussen (1967: 110) reconstructs the reflexive for Proto-Bantu as being the (11) Six OPs in (Kimenyi 2002: 20) innermost OP. a-ra-na-ha-ki-zi-ba-ku-n [som-eesh-eesh-er-er-ez-a] woman 3sg-pres-also-16-7-10-3pl-2sg-1sg The reflexive always “immediately precedes the stem, regardless of which [read-caus-caus-appl-appl-appl-fv] argument it saturates (Buell 2005: 41)”. ‘the woman is also making them read it7 (book) with them10 (glasses) to you

for me there16 (in the house)’ (15) 1sg-reflexive order a. Bukusu (Jastino Sikuku, p.c.) (12) Languages that allow four or more OPs Wekesa a-a-nj-i[siim-isy-a] Haya (Duranti 1979: 39) Wekesa 3sg-pst-1sg-refl[like-caus-fv] Kuria (Rodrigo Ranero p.c.) ‘Wekesa made me like myself’ Rwanda (Beaudoin-Lietz et al 2004: 183, Kimenyi 2002: 20) Vunjo dialect of Chaga (Moshi 1998: 144-145) b. Rundi (Juvenal Ndayiragije, p.c.) a- -ra-n-i[heb-ir-ye] Some languages allow only two or three OPs, sometimes only in restricted 3sg-tns-disj-1sg-refl[abandon-appl-fv] contexts. ‘he sacrificed himself for me’

3 / 15 c. Kerewe (David Odden, p.c.) d. Nyala-West (Ebarb et al. in prep) (18) Variable order of reflexive and 1sg OP in Tiriki (Marlo in prep.) e. Wanga (Marlo field notes) a. -l- -m - - ] - -nz- - - 3sg-fut-refl-1sg[count-appl-fv] 3sg-fut-1sg-refl[count-appl-fv] However, this is not universally true: in some languages, the 1sg OP follows ‘he will count himself for me’ ‘he will count himself for me’ the reflexive. b. -l- -m - - ] - -nz- - - (16) Reflexive-1sg order 3sg-fut-refl-1sg[shave-appl-fv] 3sg-fut-1sg-refl[shave-appl-fv] a. Kikuyu (Barlow 1951: 122) ‘he will shave himself for me’ ‘he will shave himself for me’ w- -n a - -e] 2sg-refl-1sg[turn-appl-fv] 3. Reduplication ‘turn (yourself) toward me’ Verb reduplication is a common phenomenon in Bantu languages, typically b. Tharaka (Muriungi 2008: 122) involving the doubling of the stem.3 OPs are generally not reduplicated. a- - -n[kum- r-i-a] 3sg-tns-refl-1sg[be.proud-appl-caus-fv] It is clear that the overcopying of 1sg N- in reduplication has a phonological ‘s/he has been proud to the detriment of me’ explanation.

c. Tswana (Cole 1955: 234, cf. Chebanne 1992) (19) In addition to the 1sg OP N-, the 1sg subject prefix N- is overcopied in go-i-m[pola-ɛ -a] Saamia when it immediately precedes the stem. inf-refl-1sg[kill-appl-fv] n[ ][n- - ] (/-lum-/) n[ ][n - ] (/-lom-/) ‘to kill himself for me’ 1sg[red][1sg-bite-fv] 1sg[red][1sg-talk-fv] ‘I am biting all over’ ‘I am talking’ Muriungi (2008: 121-122) takes the reflexive-1sg order in Tharaka as evidence that the 1sg OP occupies a lower structural position than the (20) In Hehe verb reduplication, prefixes and OPs normally do not copy in reflexive, which itself is lower than other OPs. reduplication (Odden & Odden 1985). a. No OP (17) Reflexive ‑, 1sg N‑, and other OPs in Tharaka (adapted from Muriungi a - -a] -ite] 2008: 121-122) inf[red][build-appl-fv] 1pl[red][buy-fv] Macrostem1 ‘to build for a bit’ ‘we shopped a bit’

b. CV- OP OP Macrostem 1 2 -fi a -a] inf-8[red][buy-fv] OP2 Macrostem3 ‘to buy a bit of them8’ ‑ OP3 Stem3 N‑

An additional variant is found in Tiriki, in which either order of 1sg and 3 See Downing 1994, 1999a, b, 2000, 2004, 2006, 2009, Hyman 2009, Hyman et al. reflexive is possible, without a concomitant change in meaning. 2008, Matondo 2003, Odden 1996, Odden & Odden 1985, Poletto 1998. 4 / 15 c. However, the 1sg OP ɲ‑ is copied. (23) One strategy to augment a monosyllabic stem to meet the disyllabic a- -n[doongaa][n- -a] minimality condition in Ndebele is to incorporate an OP (zi‑) into the 3pl-pres-1sg[red][1sg-nag-fv] stem and copy it. ‘they nag me’ (/-loong-/ ‘nag’) Non-reduplicated Reduplicated a. uku-zi[thum-a] uku-zi[thuma][thum-a] ‘to send them’ d. The 1sg SP ɲ‑ also overcopies when it immediately precedes the *uku-zi[thu][zi-thum-a] ‘to send them’ stem. b. uku-zi[dl-a] uku-zi[dla][zi-dl-a] ‘to eat them’ n n-dim- -e] 1sg[red][1sg-cultivate-appl-fv] (24) The reflexive can be overcopied in reduplication is Tharaka when it co- ‘cultivate for me!’ (/-lim-/ ‘cultivate’) occurs with a monosyllabic stem, satisfying a minimality requirement (Muriungi 2008: 110). e. Any CV-shaped prefix, including OPs and other types of [pa][ -p-e] inflectional prefixes, is overcopied in the reduplication of V-initial refl[red][refl-give-fv] verbs. ‘give yourself quickly’ kw[íita][kw-iít-a] kú-lw[iita][lw-iít-a] inf[red][inf-pour-fv] inf-11[red][11-pour-fv] The fact that the reflexive generally does not overcopy in reduplication is not ‘to pour a bit’ ‘to pour it a bit’ surprising since ‑ is syllabified in a syllable separate from the stem-initial consonant. (21) Downing (1998) a. The overcopying of nasal prefixes from outside the stem in 3.2. Cross-Bantu variation in overcopying. reduplication is the result of the shared place features across across the prefix-reduplicant boundary. The assimilated nasal is brought (25) There are differences among Bantu languages in the contexts in which into the reduplicant, and the nasal is backcopied to the base due to overcopying is found. base-reduplicant identity. a. Like Hehe, Kamba has overcopying in both CV+V and nasal prefix b. Principles of syllabification are responsible for the overcopying of contexts in verbal reduplication (Angelina Kioko, p.c.). CV- prefixes with V-initial stems, due to a requirement that stems b. Kerewe (Odden 1996) and Nande (Mutaka & Hyman 1990) also begin with an onset consonant. have reduplication CV+V and nasal prefix contexts, but in nominal, not verbal, reduplication. 3.1. Minimality-driven overcopying. Prefixes can also be overcopied in c. Sukuma has optional overcopying of cl. 9 N‑ in nominal and Bantu reduplication when a minimality requirement is imposed on the base adjectival reduplication (Matondo 2003). or reduplicant. d. Nzadi has obligatory reduplication of 1sg N‑ in verbal reduplication marking the future tense (Hyman 2011: 33-35). (22) In several Bantu languages, OPs can be recruited into an otherwise e. CV- prefixes overcopy in combination with a V-initial verb stem in subminimal reduplicative stem to meet the minimality requirement. Ndebele (Sibanda 2004: 310) and Sukuma, which has optional Bukusu (Mutonyi 2000: 111) overcopying in verbal, nominal, and adjectival reduplication Lungu (Bickmore 2007: 317) (Matondo 2003). Ndebele (Hyman et al. 2008, Sibanda 2004: 304-307) Nkore (Poletto 1998: 177) Sukuma (Matondo 2003) Tharaka (Muriungi 2008: 110)

5 / 15 (26) Lungu has the reverse pattern compared to Saamia. (Bickmore 2007: 4.1. Micro-typological patterns. 318) a. A CV- object prefix optionally overcopies in combination with a (29) In Nyaturu, two CV- OPs are generally not allowed (Hualde 1989: V-initial stem. 183-185, Olson 1964, Polak 1986, Schlindwein 1986). - -m[ ][m- - ] *w-a-va-k [t m- -aa] -vá- - - ~ - -m[ ][ - ] 3sg-pst-3pl-2sg[send-appl-fv] inf-3pl-3sg[send-appl-fv] 3pl-pres-3sg[red][3sg/Ø-beat-fv] ‘he sent them to you’ ‘to send him to them, ‘they are beating him repeatedly’ to send them to him’

b. However, the 1sg OP N‑ cannot overcopy. *n-a- -va[r ɣ- -aa] * -k - m - - -ɲ ʒ[ ][ - ] 1sg-pst-14-3pl[cook-appl-fv] inf-7-3sg[give-fv] * - - ʒ[ ][ɲ ʒ- - ] ‘I cooked it14 for them’ ‘to give it7 to him’ 3pl-pres-1sg[red][Ø/1sg-dig.up-fv] ‘they are digging me up repeatedly’ (30) However, two OPs are possible when one is the 1sg OP. w-a- -n[d ɣ- -aa] -vá-n[t m- - Bantu languages also differ in the types of prefixes that overcopy. 3sg-pst-14-1sg[cook-appl-fv] inf-3pl-1sg[send-appl-fv] ‘he cooked it14 for me’ ‘to send them to me, (27) Some languages allow any type of prefix to be overcopied. to send me to them’ a. Hehe: an OP, infinitival prefix, SP, etc. b. Saamia: 1sg SP N‑ or 1sg OP N‑. w- -ɲdʒ - -k - m - 2sg-14-1sg[lend-fv] 14.bow inf-7-1sg[give-fv] (28) Other languages restrict the prefixes that may overcopy to OPs only. ‘you should lend me your bow’ ‘to give it7 to me’ a. Only OPs overcopy with V-initial stems in Lungu. b. Only OPs overcopy with monosyllabic stems. (31) Or when one is the reflexive. Lungu (Bickmore 2007: 317-318) -mw- - Ndebele (Sibanda 2004) 3sg-3sg-refl[swear-fv] Nkore (Poletto 1998: 202-211) ‘he has sworn himself to him’ Tharaka (Muriungi 2008: 110) (32) Several other Bantu languages normally allow only a single OP but Muriungi (2008) takes the fact that only the OPs and reflexive can overcopy allow a total of two OPs just in case one is 1sg N‑ or reflexive i‑. as evidence that these morphemes occupy a different structural position from Fuliiru (Van Otterloo 2011: 40-41) other prefixes. Thus, although the motivations for overcopying in Kamba (Angelina Kioko, p.c.) reduplication are phonological, the effects are constrained by morpho- Marachi (Marlo 2007) syntactic considerations. Nyala-West (Marlo 2007) Nyaturu (Hualde 1989: 183-185, Olson 1964, Polak 1986, Schlindwein 4. OP combinations 1986) Shi (Polak-Bynon 1975: 210-212) Many Bantu languages restrict the number of OPs to just one or two (Bearth Wanga (Marlo field notes) 2003: 124-125, Marlo 2013, Marten & Kula 2012, Marten et al. 2007). There is quite a lot of variation across languages in the specific patterns that occur.

6 / 15 (33) Fuliiru allows two OPs with 1sg N‑ and reflexive yi‑ (Van Otterloo kw‑ii‑yii[kis‑ish] 2011: 40-41). inf-refl-refl[dirty-caus] a. - - -m - ‘to cause oneself to dirty oneself’ 3sg-pst-8-1sg[give-fv] ‘he’s the one who gave them8 to me’ In a number of other languages, either the 1sg OP or the reflexive licenses an additional OP, but we cannot at present be certain that the other prefix does b. - - - - not license a second OP. 3sg-1-pst rep 11-refl[throw-fv]18 ‘the one who threw himself in it11’ (36) Languages that allow two OPs only if one is 1sg N‑ Libinza (Polak 1986: 404) The Fuliiru reflexive yi‑ has a CV- shape, which is unlike many other lan- Luba-Kasai (Kuperus & Mpunga wa Ilunga 1990: 34) guages that have a V-shaped reflexive. Lulua (Polak 1986: 404) Ndengese (Polak 1986: 404) Sometimes the 1sg OP or the reflexive but not the other licenses an Punu (Polak 1986: 404) additional OP. Suku (Nurse 2007, Polak 1986: 404) Yaka (Kidima 1987: 186) (34) In Bukusu, two OPs are generally not allowed (Diercks & Sikuku 2011, Sikuku 2012). (37) Languages that allow two OPs only if one is the reflexive a. *Wamalwa a-a-mu-ba[siim-isy-a] Bakweri (Hawkinson 1986: 151-152) Wamalwa 3sg-pst-3sg-3pl[cook-caus-fv] Bukusu (Diercks & Sikuku 2011, Sikuku 2012) ‘Wamalwa made him like them’ Havu (Polak 1986: 404) Kikuyu (Barlow 1951: 266) b. Reflexive i‑ can co-occur with a second OP. Lozi (Polak 1986: 404) Wekesa a-a-mu-i[siim-isy-a] Tharaka (Muriungi 2008: 90, 121-122) Wekesa 3sg-pst-3pl-refl[like-caus-fv] ‘Wekesa made him like himself’ (38) Languages that allow three OPs only if one is the reflexive Rundi (Meeussen 1959: 102) c. 1sg n‑ cannot. Ruwund (Nash 1992: 571) *Wekesa a-a-si-m[b-a] Wekesa 3sg-pst-7-1sg[give-fv] There are a handful of cases where the 1sg OP N‑ licenses an additional OP ‘Wekesa gave me it7’ but limited additional combinations of OPs that make it clear that the correct generalizations concerning the combinatorial possibilities of OPs involve d. It is also possible for there to be two instances of the reflexive, non-phonological properties such as person and animacy. where each instance of the reflexive represents a distinct thematic role Bukusu (Sikuku 2012). (39) Lungu allows two OPs iff one is 1st person (Bickmore 2007: 30). Khalayi a-a-i-i[siim-isy-a] a. 1sg n‑ may co-occur with a second OP. Khalayi 3sg-pst-refl-refl[like-caus-fv] - - -n - - ‘ halayi made herself like herself’ 3pl-pres-3sg-1sg[grab-appl-fv] ‘they are grabbing him for me’ (35) Another language that allows multiple instances of the reflexive, each indexing a distinct thematic role, is Ruwund (Nash 1992: 571).

7 / 15 b. 1pl ‑ also licenses a second OP. c. 1sg N‑ can co-occur with an inanimate OP. y ------m - 3pl-pres-1pl-3sg[grab-appl-fv] 3sg-7-1sg[give-appl-fv] ‘they are bringing him for us’ ‘he should give it7 to me’

c. The 1st person OP must represent the indirect object d. 1sg N‑ can even co-occur with two OPs, if one is animate. (goal/beneficiary). - - - -n - - d. There are ordering restrictions: 1pl ‑ must be ordered first, and 2pl-fut-7-3sg-1sg[return-appl-fv]17 1sg n‑ must be ordered second. ‘you should return it7 to him/her for me’ e. Other combinations of OPs are “very marginal or completely ungrammatical”. 4.2. Towards a morphosyntactic analysis. If there is a general analysis of the restrictions on combinations of OPs, it is clear from the data in Lungu, (40) Two OPs are generally allowed in Nyambo; three are allowed iff one is Nyambo, and Bemba that morpho-syntactic properties such as person and 1st person (Rugemalira 1991: 203, 1993a: 229, 1997: 210, 2005: 89- number features are responsible for the surface patterns. 90). a. OP + OP + 1sg n‑ Following Buell (2005) on Zulu, Muriungi (2008) on Tharaka, and Sikuku a-ka-ga-mu-m -er-a] a-ka-ga- -n[siij-ir-a] (2012) on Bukusu, I propose that there are multiple structural positions for 3sg-pst-6-3sg-1sg[give-appl-fv] 3sg-pst-6-3sg-1sg[smear-appl-fv] OPs in these languages. ‘he gave it6 to him for me’ ‘she smeared it6 on him for me’ There is an inner morpho-syntactic position for an additional OP which is b. OP + OP + 1pl u‑ highly restricted in the features of the OPs that can appear in it. a-ka-ga-mu-tu - -ez-a] 3sg-pst-6-3sg-1pl[give-caus-appl-fv] (42) Restrictions on multiple OPs ‘he made her drink it6 for us’ a. Languages that normally allow one OP: OP2 is highly restricted

Macrostem1 c. The 1st person OP must be the rightmost OP.

(41) In Bemba, most combinations of OPs are disallowed (Marten & Kula OP1 Macrostem2 2012: 245) a. ------m - OP2 Stem 1sg-pst-3sg-6[give-fv] 1sg-pst-3sg-6[give-fv] ‘I gave him it6’ ‘I gave him it6’ Khayo restricts OP2 to 1sg only. Bukusu restricts OP2 to reflexive only. b. Two OPs are allowed if both are animate. Nyaturu restricts OP2 to 1sg or reflexive. - - - - - Lungu restricts OP2 to first person. 2pl-fut-3pl-3sg[tell-appl-fv]17 ‘you will tell them for him’

8 / 15 b. Languages that normally allow two OPs: OP3 is highly restricted Fuliiru (Van Otterloo 2011: 40-41) Ganda (Ashton et al. 1954: 29-31, 36, 65-66, 70, 93-96, 220-226) Macrostem1 Gweno (Marlo field notes)

Ha (Harjula 2004: 88-89) OP1 Macrostem2 Haya (Hyman & Byarushengo 1984) Ila (Yukawa 1987c: 242) OP2 Macrostem3 Jita (Downing 1996: 35) Kikuyu (Barlow 1951: 67, 70)

OP3 Stem3 Marachi (Marlo 2007) Mwimbi Nyaturu (Olson 1964: 178-180, Schadeberg 1979, Yukawa 1989d: Nyambo restricts OP3 to first person. 473-474) (43) Analytical challenges Nyiha (Yukawa 1989e: 507-509) a. Bemba’s restrictions are more complex and appear to require sim- Rundi (Meeussen 1959: 112, Ndayiragije 2003: 183-185) ultaneous reference to multiple OP slots. Shi (Polak-Bynon 1975: 222-223) b. In Tharaka, there is evidence for a three-way structural division of some varieties of Swahili (Haddon 1955: 81) OPs (Muriungi 2008: 121-122), but a maximum of two OPs, one Tiriki of which must be 1sg or reflexive, is allowed at any one time. Tsootso (Dalgish 1974: 85-88) Tswana (Cole 1955: 241, Hyman 2001: 157, 162, Mmusi 1992) 4.3. A morpho-phonological account? (46) In Kamba (Angelina Kioko, p.c.) and Mwera (Harries 1950: 87, cited (44) Polak (1986) in Devos & Van Olmen 2013) there is variation in the form of the im- perative verb with the 1sg OP : the FV is optionally or . a. In CV-N- OP combinations with 1sg ‑ and in CV-V- OP N‑ ‑a ‑e

combinations with reflexive ‑, the two OPs are together only a single syllable in size, just as when there is only a single CV- OP. (47) Some other languages have generalized the pattern of CV- OPs such b. When two CV- OPs co-occur, the OPs together are two syllables in that all OPs trigger final ‑e. size. a. 1sg ni‑ c. One could potentially implement this insight as a monosyllabic Mwenyi (Yukawa 1987b: 64) template on the OPs, which prevents CV-CV- combinations but Sukuma (Yukawa 1989b: 387-388) allows CV-N- and CV-V-. Swahili d. However, a phonological analysis along these lines could not handle the facts of Fuliiru, Lungu, Nyambo, and Bemba, since OP b. 1sg ngu‑ Luvale (Yukawa 1987a: 64) combinations like ‑lu‑yi‑, ‑ ‑ ‑, ‑ga‑ u‑ u‑, and ‑ ‑ ‑ are clearly larger than a single syllable. Makonde (Yukawa 1989f: 551-552)

5. Allomorphy of the FV in the imperative c. 1sg ndi‑ Chewa (Ndayiragije 2003: 183) (45) Many other modern day Bantu languages have the Rundi pattern of FV allomorphy in the imperative, which Meeussen (1962: 74, 1967: 112) d. 1sg ngi‑ reconstructs to Proto-Bantu. Zulu (Engelbrecht 1957) Bajuni (Nurse n.d.: 23) 9 / 15

e. 1sg n ‑ (51) Bukusu has a unique pattern of FV allomorphy in the imperative (Mau- Nilamba (Yukawa 1989c: 439-440) rice Sifuna, p.c., Justine Sikuku, p.c., Sikuku 2012). a. Bare stems end in ‑a. f. 1sg N‑ [bek-a] ‘shave!’ Khayo (Marlo 2009) Lungu (Bickmore 2007: 307) b. Forms with a CV- OP end in ‑e. eastern Tswana (Cole 1955: 241) mu[bek-e] ‘shave him!’ Tura (Marlo 2008) c. The 1sg OP n‑ also triggers final ‑e. (48) Imperative FV allomorphy in Khayo (Marlo 2009: 111) m[bek-e] ‘shave me!’ [liingaal- ] ‘watch!’ mu aa - ] ‘watch him!’ d. Strikingly, the reflexive takes the FV ‑a. n aa - ] ‘watch me!’ i[bek-a] ‘shave yourself!’ *i[bek-e] ‘shave yourself!’ There are a number of other languages that show interesting variations on the (Recall that reflexive i‑ is also unique in Bukusu in exceptionally Rundi pattern of FV allomorphy in the imperative. licensing a second OP.)

(49) Tharaka has the usual pattern of FV allomorphy (Muriungi 2008: 118- (52) Digo is another example of the familiar pattern of FV allomorphy in 123) the imperative (Kisseberth 1981: 95-97, 1984: 158). a. Bare stems condition the FV ‑a. a. Bare stems take the FV ‑a. [ring-a] ‘hit!’ [ ʃeke ʃer-a] ‘sift for!’ [tsukur-a] ‘carry!’ [ring-ith-i-a] John -thingo ‘coerce John to hit the wall!’ b. Forms with other OPs take the FV ‑e. b. 1sg N‑ triggers final ‑a. a[ ʃ ʃer-e] ‘sift for them!’ a -e] ‘carry them!’ n[tem-a] ‘cut me!’ c. Stems with the 1sg OP take the FV ‑a. Unlike the other examples c. Other OPs, including the reflexive, trigger final ‑e. we have seen, the 1sg OP ni‑ has a CV- shape. [ring-e] ‘hit it9!’ ni[ ʃeke ʃer-a] ‘sift for me!’ ni[tsukur-a] ‘carry me!’ [tem-e] ‘cut us!’ [tem-e] ‘cut him!’ (53) There are several other languages with the Digo pattern of FV allo- [ring-e] ‘hit yourself!’ morphy in the imperative, where 1sg ni‑ triggers the FV ‑a in the imperative. Tharaka possesses an additional type of imperative form with an ‘expletive’ Chimwiini (Kisseberth & Abasheikh 2004) prefix ‑, which “appears to be semantically vacuous” and occurs only in Giryama (Lax 1996: 267, 293-294, 301) imperatives (Muriungi 2008: 119). Forms with expletive ‑ end with the FV Mbaga Pare (Kagaya 1989: 142-145) ‑a, and expletive ‑ does not co-occur with an OP. tang’ata Swahili (Whiteley 5 : 32, 33, cited in Devos & Van Ol- men 2013: 20) (50) Expletive ‑ in Tharaka (Muriungi 2008: 119, 122) a. [tem-a] ‘cut!’ b. * -n[tem-a] ‘hit me!’ * -n[tem-e] ‘hit me!’ 10 / 15 (54) A similar pattern is found in Machame Chaga (Yukawa 1989a: 324- uriungi’s analysis straightforwardly extends to languages like Digo and 326). Machame Chaga, which have 1sg OPs with a CV- shape. a. Bare stems end in ‑a. [kab- ] ‘beat!’ - ] ‘look for!’ (57) 1sg below other OPs (Digo, Machame Chaga) Macrostem1 b. Most OPs including 3sg n ‑ and 3pl va‑ take the FV ‑e. [kab- ] ‘beat him!’ - ] ‘look for him!’ OP1 Macrostem2 va - ] ‘beat them!’ va - ] ‘look for them!’

OP Stem c. The 1sg OP ʃi- takes ‑a. 2 ʃi[kab- ] ‘beat me!’ ʃi - ] ‘look for me!’ ni‑ (Digo) ʃ ‑ (Machame Chaga) (55) A further important variation on the most common pattern of excep- tional allomorphy is found in Ndebele (Sibanda 2004: 29, 112-114). uriungi’s analysis also extends to Bukusu. In Bukusu the reflexive is the a. Bare stems end in ‑a. structurally lowest OP (Sifuna 2012). Two consequences of the lower posi- [that-a] ‘take!’ tion of the reflexive are that (i) the reflexive licenses a second OP and (ii) it [ph-an-a] ~ yi[ph-a] ‘give!’ triggers final ‑a in the imperative. Thus, the rule that inserts final ‑e in Bukusu applies when there are prefixes higher in the tree than the reflexive. b. Most OPs, including the reflexive, require the FV ‑e, ba[thath-e] ‘take them!’ (58) Reflexive below other OPs (Bukusu)

ba[ph-e] ‘give them!’ Macrostem1 zi[tshay-e] ‘hit yourself! ‘give yourself!’ zi[ph-e] OP Macrostem 1 2

d. 1st person OPs (sg. and pl.) optionally take final ‑a or ‑e. ngi[tshay-a] ~ ngi[tshay-e] ‘hit me!’ OP2 Stem si[ph-a] ~ si[ph-e] ‘give us!’ i‑

5.1. A morpho-syntactic analysis. I largely adopt Muriungi’s (2008) I propose that the 1st person OPs of Ndebele are structurally lower than other morpho-syntactic analysis of the imperative allomorphy patterns in Tharaka OPs. The rule inserting ‑e is variable in the level of the tree that it is sensitive and adapt it to account for a broader range of data across Bantu. to, varying in including or excluding the 1st person OPs.

(56) Muriungi (2008) (59) 1st person below other OPs (Ndebele)

a. 1sg N‑ and expletive ‑ in Tharaka trigger different FVs from other Macrostem1 OPs because these morphemes occupy a lower structural position than other OPs. OP1 Macrostem2 b. The FV ‑a is treated as the ‘default’, and final ‑e “is inserted in some specific environments ... [such as] when there are prefixes higher than (and including) reflexive in the clausal hierarchy in OP2 Stem imperatives (Muriungi 2008: 123).” ngi‑ (1sg) si‑ (1pl) 11 / 15 I have argued that the exceptional patterns of OP combinations and (60) Khayo allomorphy of the FV in the imperative are the result of the lower structural a. 1sg OP N‑ triggers the FV ‑e in the imperative. position of these morphemes, closer to the stem, as proposed by Buell b. Two OPs are allowed only with 1sg N‑. (2005), Muriungi (2008), and Sikuku (2012) for their analyses of individual c. 1sg N‑ may still be structurally lower than other OPs. Bantu languages. d. The rule that inserts final ‑e is not sensitive to the structural difference between 1sg and other OPs—any structure larger than Although more work is required to provide fully fleshed out formal analyses the verb stem yields final ‑e. of all relevant object marking data in all of the languages mentioned here, I hope to have helped brings us some distance forward in clarifying how the 5.2. Rundi. Rundi has the same general patterns of allomorphy as Tharaka, components of grammar interact to produce analytically challenging excep- but the surface order of 1sg and reflexive is reversed: reflexive i‑ follows 1sg tional behavior of the 1sg and reflexive OPs in Bantu languages. N‑ (see (15b)), by Baker’s ( 5) irror Principle, should therefore be struc- turally lower than 1sg. References

A direct translation of uriungi’s (2008) analysis to Rundi predicts that the Adams, Nikki. 2010. The Zulu ditransitive verb phrase. Chicago, IL: University of Chicago FV would be ‑a with the reflexive and with bare verb forms, but ‑e with 1sg Ph.D. dissertation. Alsina, Alex. 1994. Bantu multiple objects: analyses and fallacies. Linguistic Analysis 24. N‑ and other OPs. 153-174. One possibility is to propose that at a certain point in the syntactic derivation, Ashton, Ethel O., Enoch M. K. Mulira, E. G. M. Ndawula & Archibald Norman Tucker. 1954. Rundi has the same structure as Tharaka, with the 1sg OP lower than A Grammar. London, New York & Toronto: Longmans, Green & Co. Baker, Mark. 1985. The Mirror Principle and morpho-syntactic explanation. Linguistic Inquiry reflexive. It is this structure that regulates the choice of FV. After the choice 16: 373-415. of FV has been determined, a later reordering of these prefixes subsequently Baker, Mark. 2008. The syntax of agreement and concord. Cambridge: Cambridge University takes place, such that the reflexive is pronounced closer to the stem. Press. Barlow, A. Ruffell. 1960. Studies in Kikuyu grammar and idiom. Edinburgh: Blackwood. 5.3. A morpho-phonological analysis? Bax, Anna & Michael Diercks. 2012. Information structure constraints on object marking in Manyika. Southern African Linguistics and Applied Language Studies 30. 185-202. (61) Ndayiragije (2003): The allomorphy patterns of the final vowel are Bearth, homas. . Syntax. In erek urse rard Philippson (eds.), The Bantu Lan- phonologically determined. guages, 127-142. London: Routledge. Beaudoin-Lietz, Christa, Derek Nurse & Sarah Rose. 2004. Pronominal object marking in a. The final vowel is ‑e when there is a syllable before the stem, i.e. Bantu. In Akinbiyi Akinlabi & Oluseye Adesola (eds.), Proceedings of the 4th World with CV‑ OPs. Congress of African Linguistics, New Brunswick 2003, 5- . ln: Rüdiger Köppe. b. Else the final vowel is ‑a. Bickmore, Lee. 2007. Cilungu phonology. Stanford, CA: CSLI Publications. Botne, Robert, Hannington Ochwada & Michael Marlo. 2006. A grammatical sketch of the (62) It is clear from Bukusu, Digo, Machame Chaga, and Ndebele that there Lusaamia verb. ln: Rüdiger Köppe. is no phonologically based analysis that can account for all of the Ban- Bresnan, Joan & Sam Mchombo. 1987. Topic, pronoun, and agreement in hiche a. Lan- tu patterns. guage 63. 741-782. Buell, Leston. 2005. Issues in Zulu verbal morpho-syntax. Los Angeles, CA: University of 6. Conclusion California Ph.D. dissertation. Byarushengo, Ernest, Larry M. Hyman & Sarah Tenenbaum. 1976. Tone, accent and assertion Overcopying in reduplication, however, has a phonological motivation, as in Haya. In Larry M. Hyman (ed.), Studies in Bantu tonology, 183-205. Los Angeles: De- proposed in several papers by Laura Downing. partment of Linguistics, University of Southern California.

12 / 15 Chebanne, Andy. 1992. The Setswana i(n) verbal prefix. Paper presented at the Annual Con- Duranti, Alessandro. 1979. Object clitic pronouns in Bantu and the topicality hierarchy. Stud- ference of African Linguistics 23, Michigan State University. ies in African Linguistics 10. 31-45. Cole, Desmond T. 1955. An introduction to Tswana grammar. Cape Town: Longman. Duranti, Alessandro Ernest Byarushengo. . On the notion of ‘direct object’. In Ernest Contini-Morava, Ellen. 1983. Ranking of participants in : the limitations of arbi- Byarushengo, Alessandro Duranti & Larry M. Hyman (eds.), Haya Grammatical Struc- trariness in language. Anthropological Linguistics 25. 425-435. ture, 54-71. Los Angeles, CA: Department of Linguistics, USC. Creissels, Denis. 2002. Valence verbale et voix en Tswana. s- Ebarb, Kristopher, Michael R. Marlo, & Peter Otiato. in prep. Nyala-West tonology. Bloom- tique de Paris 97. 371-426. ington, IN, Columbia, MO & Lawrence, KS: Indiana University, University of Missouri Dalgish, Gerald. 1974. Arguments for a unified treatment of y-initial and vowel-initial roots in & University of Kansas unpublished ms. Olutsootso. Studies in the Linguistic Sciences 4. 76-90. Engelbrecht, Jan Anthonie. 1957. Notes on the imperative in Zulu. African Studies 16. 102- De Guzman, Videa. 1987. Indirect objects in SiSwati. Studies in African Linguistics 18. 309- 107. 325. Friesen, Lisa. 2002. Valence change and Oroko verb morphology (Mbonge dialect). Grand Demuth, Katherine & Mark Johnson. 1990. Interactions between discourse function and Forks: University of North Dakota M.A. thesis. agreement in Setawana. Journal of African Languages and Linguistics 20. 41-61. Haddon, Ernest. 1955. Note on the verbal -e stem in East African Bantu. Journal of the Inter- evos, aud ani l an Olmen. 2013. Describing and explaining the variation of Bantu national African Institute 25: 79-83. imperatives and prohibitives. Studies in Language 37. 1-57. Harjula, Lotta. 2004. The Ha language of . ologne: diger ppe. Dickens, Patrick. 1984. The history of the so-called strengthening in Tswana. Journal of Afri- Hawkinson, Ann Katherine. 1986. Bakweri verb morphology. Berkeley, CA: University of can Languages and Linguistics 6. 97-125. California Ph.D. dissertation. Diercks, Michael & Justine Sikuku. 2011. Object marking as pronominal incorporation in Henderson, Brent. 2006. The syntax and typology of Bantu relative clauses. Urbana, IL: Uni- Lubukusu. Paper presented at the 4th International Conference on Bantu languages, versity of Illinois Ph.D. dissertation. Humboldt University of Berlin. Hualde, Jos . . ouble object constructions in irimi. In obert Botne & Paul Newman Downing, Laura J. 1994. Siswati verbal reduplication and the theory of generalized alignment. (eds.), Current Approaches to African Linguistics 5, 179-189. Dordrecht: Foris Publica- Proceedings of the North East Linguistic Society 24. 81-95. tions. Downing, Laura J. 1996. The tonal phonology of Jita. Munich: LINCOM Europa. Hubbard, athleen. 5. ‘Prenasalized consonants’ and syllable timing: evidence from Downing, Laura J. 1998. Prosodic misalignment and reduplication. In Geert Booij & Jaap van Runyambo and Luganda. Phonology 12. 235-256. Marle (eds.) Yearbook of Morphology 1997, 83-120. Dordrecht: Kluwer Academic Pub- Hyman, Larry M. 1987. Prosodic domains in Kukuya. Natural Language and Linguistic Theo- lishers. ry 5. 311-333. Downing, Laura J. 1999a. Morphological constraints on Bantu reduplication. Linguistic Anal- Hyman, Larry M. 1992. Moraic mismatches in Bantu. Phonology 9. 255-265. ysis 29. 6-46. Hyman, Larry M. 2001. On the limits of phonetic determinism in phonology: *NC revisited. Downing, Laura J. 1999b. Prosodic stem ≠ prosodic word in Bantu. In . Alan Hall Ursula In Beth Hume & Keith Johnson (eds.), The role of speech perception in phonology, 141- Kleinhenz (eds.), Studies on the phonological word, 73-98. Amsterdam: John Benjamins. 185. New York: Academic Press. Downing, Laura J. 2000. Morphological and prosodic constraints on Kinande verbal redupli- Hyman, Larry M. 2003. Segmental phonology. In erek urse rard Philippson (eds.), cation. Phonology 17. 1-38. The Bantu languages, 42-58. London: Routledge. Downing, Laura J. 2003. Compounding and tonal non-transfer in Bantu languages. Phonology Hyman, Larry M. 2008. Directional asymmetries in the morphology and phonology of words, 20. 1-42. with special reference to Bantu. Linguistics 46. 309-350. Downing, Laura J. 2004. Bukusu reduplication. In Chege Githiora, Heather Littlefield & Vic- Hyman, Larry M. 2009. The natural history of verb-stem reduplication in Bantu. Morphology tor Manfredi (eds.), Trends in African Linguistics 5, 73-84. Lawrenceville, NJ: Africa 19: 177-206. World Press. Hyman, Larry M. 2011. The macro-Sudan belt and Niger-Congo reconstruction. Language Downing, Laura J. 2005. On the ambiguous segmental status of nasals in homorganic NC Dynamics and Linguistic Change 1. 3-49. sequences. In Marc van Oostendorp & Jeroen van de Weijer (eds.), The internal organi- Hyman, Larry M. & Ernest Byarushengo. 1984. A model of Haya tonology. In George N. zation of phonological segments, 183-216. Berlin: Mouton de Gruyter. Clements & John Goldsmith (eds.), Autosegmental studies in Bantu tone, 53-103. Dor- Downing, Laura J. 2006. Canonical forms in prosodic morphology. Oxford: Oxford Universi- drecht: Foris Publications. ty Press. Hyman, Larry M., Sharon Inkelas & Galen Sibanda. 2008. Morpho-syntactic correspondence Downing, Laura J. 2009. Linear disorder in Bantu reduplication. Paper presented at the Work- in Bantu reduplication. In Kristin Hanson & Sharon Inkelas (eds.), The nature of the shop on the Division of Labor between Morphology and Phonology and Fourth Network word: essays in honor of Paul Kiparsky, 273-310. Cambridge, MA: MIT Press. Meeting, Meertens Instituut, Amsterdam. 13 / 15 Kagaya, Ryohei. 1989. A study of the tonal system of the Mbaga dialect of the southern Pare Mmusi, Sheila. 1992. Obligatory contour principle effects and violations: the case of Setswa- language. In Ryohei Kagaya, Ruth Mfumbwa Besha & Yasutoshi Yukawa (eds.), Studies na verbal tone. Urbana, IL: University of Illinois Ph.D. dissertation. in Tanzanian Languages, 95-204. Tokyo: ILCAA. orolong, ‘ alillo Larry . Hyman. 1977. Animacy, objects and clitics in Sesotho. Stud- Kearns, Lucie. 2008. L s m rq rs ’ bj s s s b ntoues et romanes. Montreal: ies in African Linguistics 8. 199-218. Universit du u bec ontr al .A. thesis. Muriungi, Peter. 2008. Phrasal movement inside Bantu verbs: der s r r Kidima, Lukowa. 1987. Object agreement and topicality hierarchies in Kiyaka. Studies in r . Tromsø: University of Tromsø Ph.D. dissertation. African Linguistics 18. 175-209. Mutaka, Ngessimo & Larry M. Hyman. 1990. Syllables and morpheme integrity in Kinande Kimenyi, Alexandre. 2002. A tonal grammar of Kinyarwanda—an autosegmental and met- reduplication. Phonology 7. 73-119. rical analysis. Lewiston, NY: The Edwin Mellen Press. Mutonyi, Nasiombe. 2000. Aspects of Bukusu morphology and phonology. Columbus, OH: Kisseberth, Charles W. 1981. Displaced tones in Digo (Part 2). Studies in the Linguistic Sci- Ohio State University Ph.D. dissertation. ences 11. 73-120. Nash, Jay. 1992. Aspects of Ruwund grammar. Urbana, IL: University of Illinois Ph.D. disser- Kisseberth, Charles W. 1984. Digo tonology. In George N. Clements & John Goldsmith (eds.), tation. Autosegmental studies in Bantu tone, 105-182. Dordrecht: Foris Publications. dayiragije, Juvenal. . h ories linguistiques et r ciprocit en hichewa: la lec on du Kisseberth, Charles W. & Mohammed I. Abasheikh. 2004. The Chimwiini lexicon exemplified. . In Patrick Sauzet & Anne Zribi-Hertz (eds.), Typologie des lang s ’A r q Tokyo: Research Institute for Languages and Cultures of Asia and Africa. universaux de la grammaire, Volume 1, 169- . Paris: L’Harmattan. Kuperus, Julianna & Antoinette Mpunga wa Ilunga. 1990. Locative markers in Luba. Ter- Nurse, Derek. 2007. Bantu tense and aspect systems. http://www.ucs.mun.ca/~dnurse/tabantu vuren: us e royal de l’Afrique centrale. .html. Accessed 2 June 2013. Lax, Michelle. 1996. The verbal system of Kigiryama. Madison, WI: University of Wisconsin Nurse, Derek. n.d. Bajuni grammatical sketch. St. John’s: Memorial University of Newfound- Ph.D. dissertation. land unpublished ms. Marlo, Michael R. 2002. Reduplication in Lusaamia. Indiana University Working Papers in Odden, David. 1996. Patterns of reduplication in Kikerewe. Ohio State University Working Linguistics Online 2. Papers in Linguistics 48. 111-148. Marlo, Michael R. 2004. Prefixal reduplication in Lusaamia: evidence from morphology. In Odden, David. to appear. Bantu phonology. Oxford Handbooks Online. Oxford: Oxford Uni- Proceedings of the 4th World Congress on African Linguistics, Akinbiyi Akinlabi & versity Press. Oluseye Adesola (eds.). Köln: Rüdiger Köppe. Odden, David & Mary Odden. 1985. Ordered reduplication in Kihehe. Linguistic Inquiry 16. Marlo, Michael R. 2007. The verbal tonology of Lumarachi and Lunyala-West: two dialects of 497-503. Luluyia (Bantu, J.30, Kenya). Ann Arbor, MI: University of Michigan Ph.D. dissertation. Olson, Howard. 1964. The phonology and morphology of Rimi. Hartford, CT: Hartford Semi- Marlo, Michael R. 2008. Tura verbal tonology. Studies in African Linguistics 37. 153-243. nary Foundation. Marlo, Michael R. 2009. Khayo verbal tonology. Africana Linguistica XV. 77-129. Polak, Louise. . Le r fl chi en bantou. Africana Linguistica IX. 270-304. Marlo, Michael R. 2013. On the number of object markers in Bantu languages. Columbia, Polak, Louise. . Les infixes (“pr fixes objets”) du bantu et leur reconstruction. Africana MO: University of Missouri unpublished ms. Linguistica X. 365-421. Marlo, Michael R. 2014. Exceptional patterns of object marking in Bantu II. Columbia, MO: Polak-Bynon, Louise. 1975. A Shi grammar. ervuren, Belgium: us e oyal de l’Afrique University of Missouri unpublished ms. Centrale. Marten, Lutz & Nancy Kula. 2012. Object marking and morpho-syntactic variation in Bantu. Poletto, Robert. 1998. Topics in Runyankore phonology. Columbus, OH: Ohio State Universi- Southern African Linguistics and Applied Language Studies 30. 237-253. ty Ph.D. dissertation. Marten, Lutz, Nancy Kula & Nhlanhla Thwala. 2007. Parameters of morpho-syntactic varia- Riedel, Kristina. 2009. The syntax of object marking in Sambaa: a comparative Bantu per- tion in Bantu. Transactions of the Philological Society 105. 253-338. spective. Utrecht: LOT. Massamba, David. 1982. Aspects of accent and tone in Ci-Ruri. Bloomington, IN: Indiana Rugemalira, Josephat M. 1991. What is a symmetrical language? Multiple object constructions University Ph.D. Dissertation. in Bantu. In Proceedings of the Seventeenth Annual Meeting of the Berkeley Linguistics Matondo, Masangu. 2003. Tone and prosodic morphology in Kisukuma. Los Angeles, CA: Society: Special Session on African Language Structures, 200-209. University of California Ph.D. dissertation. Rugemalira, Josephat M. 1993. Bantu multiple “object” constructions. Linguistic Analysis 23: Mchombo, Sam. 2004. The syntax of Chichewa. Cambridge: Cambridge University Press. 226-252. Meeussen, A. E. 1959. Essai de grammaire rundi. ervuren: us e royal de l’Afrique central. Rugemalira, Josephat M. 1997. The upper limit constraint on argument structures. In Robert Meeussen, A. E. 1962. De tonen van subjunktief en imperatief in het Bantoe. Africana Lin- Herbert (ed.), African linguistics at the crossroads: papers from Kwaluseni, 207-219. guistica I. 57-74. Koln: Rudiger Koppe. Meeussen, A. E. 1967. Bantu grammatical reconstructions. Africana Linguistica III. 79-121. 14 / 15 Rugemalira, Josephat M. 2005. A grammar of Runyambo. Dar es Salaam: Languages of Tan- Yukawa, Yasutoshi. 1989f. A tonological study of Makonde verbs. In Ryohei Kagaya, Ruth zania Project. Mfumbwa Besha & Yasutoshi Yukawa (eds.), Studies in Tanzanian languages, 519-560. Schadeberg, Thilo. . ber die t ne der verbalen formen im imi. A r b rs 62: Tokyo: ILCAA. 288-313. Zeller, Jochen. 2012. Object marking in Zulu. Southern African Linguistics and Applied Lan- Schlindwein, Deborah. 1986. On the invisibility of the first person singular object marker in guage Studies 30. 219-235. KiRimi. Paper presented at Annual Conference on African Linguistics 17, Indiana Uni- versity. Sibanda, Galen. 2004. Verbal phonology and morphology of Ndebele. Berkeley, CA: Universi- ty of California Ph.D. dissertation. Sikuku, Justine M. 2012. Comparing reflexive and object marking in Lubukusu. Paper pre- sented at World Congress of African Linguistics 7. Buea: University of Buea. van der Spuy, Andrew. 1993. Dislocated noun phrases in Nguni. Lingua 90. 335-355. Van Otterloo, Roger. 2011. The Kifuliiru language, volume 2: a descriptive grammar. Dallas, TX: SIL International. Woolford, Ellen. 1999. Animacy hierarchy effects on object agreement. In Paul Kotey (ed.), New dimensions in African linguistics and languages, 203-216. Trenton, NJ: Africa World Press. Woolford, Ellen. 2000. Agreement in disguise. In Vicki Carstens & Frederick Parkinson (eds.), Advances in African linguistics, 103-117. Trenton, NJ: Africa World Press. Woolford, Ellen. 2001. Conditions on object agreement in Ruwund (Bantu). In Elena Benedic- to (ed.), Indigenous languages, 177-201. Amherst, MA: GLSA. Yukawa, Yasutoshi. 1987a. A tonal study of Luvale verbs. In Yasutoshi Yukawa, Simooya Jerome Hachipola & Ryohei Kagaya (eds.), Studies in Zambian languages, 1-34. Tokyo: ILCAA. Yukawa, Yasutoshi. 1987b. A tonal study of Mwenyi verbs. In Yasutoshi Yukawa, Simooya Jerome Hachipola & Ryohei Kagaya (eds.), Studies in Zambian languages, 35-72. To- kyo: ILCAA. Yukawa, Yasutoshi. 1987c. A tonal study of Ila verbs. In Yasutoshi Yukawa, Simooya Jerome Hachipola & Ryohei Kagaya (eds.), Studies in Zambian languages, 185-256. Tokyo: IL- CAA. Yukawa, Yasutoshi. 1989a. A tonological study of Machame verbs. In Ryohei Kagaya, Ruth Mfumbwa Besha & Yasutoshi Yukawa (eds.), Studies in Tanzanian languages, 223-338. Tokyo: ILCAA. Yukawa, Yasutoshi. 1989b. A tonological study of Sukuma verbs. In Ryohei Kagaya, Ruth Mfumbwa Besha & Yasutoshi Yukawa (eds.), Studies in Tanzanian languages, 339-404. Tokyo: ILCAA. Yukawa, Yasutoshi. 1989c. A tonological study of Nilamba verbs. In Ryohei Kagaya, Ruth Mfumbwa Besha & Yasutoshi Yukawa (eds.), Studies in Tanzanian languages, 405-450. Tokyo: ILCAA. Yukawa, Yasutoshi. 1989d. A tonological study of Nyaturu verbs. In Ryohei Kagaya, Ruth Mfumbwa Besha & Yasutoshi Yukawa (eds.), Studies in Tanzanian languages, 451-480. Tokyo: ILCAA. Yukawa, Yasutoshi. 1989e. A tonological study of Nyiha verbs. In Ryohei Kagaya, Ruth Mfumbwa Besha & Yasutoshi Yukawa (eds.), Studies in Tanzanian languages, 481-518. Tokyo: ILCAA. 15 / 15