1 Introduction An applicative is a valency-changing morpheme that - - PDF document

Kyle Jerro ACAL 46 Presentation University of Texas at Austin University of Oregon jerrokyle@gmail.com March 27, 2015

Locative applicatives and the interaction of verb class

1 Introduction

• An applicative is a valency-changing morpheme that adds a new object to the argument

structure of the sentence.

• An example from Kinyarwanda (spoken by approximately 12 million people in Rwanda,

Burundi and DR Congo), with the applicative morpheme –ir:1 (1)

• a. Umu-gabo

1-man a-ra-ndik-a 1S-pres-write-imp in-kuru 9-story ku for mw-ana. 1-child ‘The man is writing the story for the child.’

• b. Umu-gabo

1-man a-ra-ndik-ir-a 1S-pres-write-appl-imp umw-ana 1-child in-kuru. 9-story ‘The man is writing the child the story.’

• Most of the literature has focused on the syntactic properties of applied object, espe-

cially “object symmetry,” which looks at whether the thematic object and the applied

• bject have equal access to objecthood operations like passivization and object incor-

poration.

• These approaches assume that the semantics of applicativization transparently adds a

new object participant with a specific thematic role to the verb’s argument structure.

• However, work on other languages has shown that the semantics of particular verb

classes affects argument realization patterns (Fillmore 1970, Levin 1993, Rappaport Hovav and Levin 2008, Beavers 2011, inter alia).

• I show that verb class affects the argument realization of the applicative morpheme,

introducing a four-way typology dependent on the nature of location encoded by the verb.

• I analyze applicative morphology as an operation that adds a monotonically stronger

lexical entailment to the meaning of the verb.

• Roadmap

– Previous approaches to applicative morphology

1All Kinyarwanda data in this paper come from elicitations conducted by the author. This applicative

form contrasts with the applicative described in Kimenyi (1980), i.e. –ho. For all the speakers I’ve consulted, –ir is the locative applicative, while –ho is one of a class of locational clitics (cf. §4).

1

– Variation in applicative uses – Four separate verb classes – Analysis: Semantic strengthening

2 Previous Approaches

• Previous work on applicative morphology has looked almost exclusively at the syntactic

nature of the applied object in relation to the thematic object (i.e. the object licensed by the verb).

• Crucially, the mainstay of research on applied objects has looked at their syntax,

analyzing applicativization as an operation that adds an object to the argument struc- ture of the verb (see Baker 1988, Bresnan and Moshi 1990, Alsina 1992, Alsina and Mchombo 1993, Marantz 1993, Pylkk¨ anen 2008, McGinnis 2001, McGinnis and Gerdts 2003, Baker and Collins 2006, Zeller 2006, Zeller and Ngoboka 2006, Marten et al. 2007, Jeong 2007, Baker et al. 2012, Jerro in press, inter alia).

• Some researchers have noticed that the wholesale addition of an object is not sufficient,

showing instances where the applicative affects the meaning of the thematic object instead of adding a new syntactic object (Marten 2003, Creissels 2004, Cann and Mabugu 2006, Bond 2009).

• For example, Marten (2003) notes a use of the applicative in Swahili that indicates a

pragmatically noteworthy property of the object encoded by the verb, as in (2). (2)

• a. Salma

Salma a-li-ka-a 1sgS-pst-sit-fv kiti-ni. chair-loc ‘Salma was sitting on the chair.’

• b. Salma

Salma a-li-kal-i-a 1sgS-pst-sit-appl-fv kiti chair cha gen uvivu. laziness ‘Salma was slouching/sitting in a comfortable chair.’ (Marten 2003:8,(12))

• The data in (2) show that there is a semantic/pragmatic use of applicatives that lies
• utside the standard analysis of object addition.
• Specifically, Marten argues for what he terms “concept strengthening,” where the ap-

plicative is used to make a claim about the object of the sentence than what is available with the non-applied form.

• Few studies have investigated the effect of verb class on the realization of applicative

morphology (though see Cann and Mabugu (2006) for some discussion of this).

• I explore the interaction of verb class and applicatives from the view of the semantic

meanings that are available with and without applicative morphology.

• In this talk, I use locative applicatives in Kinyarwanda as a case study for understand-

ing the interaction of applicative meaning and verb meaning. 2

3 Typology of Locative Meanings

• I lay out four kinds of locative meanings added to a verb, contingent upon the class of

the base predicate.

• 1. general location
• 2. goal
• 3. path
• 4. sub-location
• The first category are verbs where the applicative adds a general location, as in (3).

(3)

• a. Yohani

John a-ri 1S-be ku-vug-a. inf-talk-imp ‘John is talking.’

• b. Yohani

John a-ri 1S-be ku-vug-ir-a inf-talk-appl-imp mu in nzu. house ‘John is talking in the house.’

• In (3a), the verb ku-vuga ‘to talk’ is intransitive, while in (3b), there is a new locative

PP, licensed by the applicative morpheme, indicating the location of the talking event.

• The second category is where a goal is added to the meaning of the verb.

(4)

• a. Yohani

John a-ri 1S-be kw-iruk-a. inf-run-imp ‘John is running.’

• b. Yohani

John a-ri 1S-be kw-iruk-ir-a inf-run-appl-imp kw’ to isoko. market ‘John is running to the market.’

• In (4b), the new location licensed by the applicative is not a general description of

where the event took place, but rather the goal of the running event.

• Third, the applicative may add a path to the event, as in (24).

(5)

• a. N-di

1sg-cop kw-injir-a inf-enter-imp mu in nzu. house ‘I entered the house.’

• b. N-di

1sg-cop kw-injir-ir-a inf-enter-appl-imp mu in muryango. door ‘I entered through the door.’

• Here, the applied object describes the path through which the motion event occurs.

3

• Finally, the applied object may describe a “sub-location” — some small item that is

positioned beneath the subject. (6)

• a. N-icay-e

1sg-sit-perf mu in bitaro. hospital ‘I sat in the hospital.’

• b. N-icar-iy-e

1sg-sit-appl-perf terefoni. telephone ‘I sat on the telephone.’

3.1 Evidence for the Typology

3.1.1 Interpretive Differences

• One indication of the differences between the applied and non-applied variants is the

interpretive difference of the locational phrase.

• For example, consider the following context: John is cooking, and he’s talking about

needing to run back to the store to get some things he forgot. I leave the room, but when I get back, he’s gone. I ask someone where he went. (7) Question: Where did John go? (8)

• a. Y-iruk-iy-e

1-run-appl-imp kw’ to isoko. store ‘He ran to the store.’

• b. *Y-irutse-e

1-run-imp kw’ to isoko. store ‘He ran to the store.’

• Consider another scenario: I’m standing in front of my house, and I see a lion. I want

to run inside to get away from it. (9)

• a. N-iruk-iy-e

1-run-appl-perf mu in nzu. house ‘I ran into the house.’

• b. N-iruts-e

1-run-perf mu in nzu. house *‘I ran into the house.’ ‘I ran in the house.’

• Positional verbs also show interpretive differences between the applied and non-applied
• variants. Locations can be used with both variants, but the interpretations differ:

4

(10)

• a. N-icay-e

1sg-sit-perf ku in mazi. water ‘I sat in the water. (e.g. in a lake or a pool).’

• b. N-icar-iy-e

1sg-sit-appl-perf amazi. water ‘I sat in the water. (e.g. a puddle of water on a bench after it rained).’ 3.1.2 Locational Clitics

• Kinyarwanda has three locational clitics that serve as pronouns for locative PPs/NPs.

– =yo : goals – =mwo/mo : inside of something – =ho : at or on something

• For verbs that acquire a goal with the use of an applicative, it is acceptable to use the

=yo clitic with the applied variant, but not the non-applied variant. (11) kwiruka : to run

• a. *N-iruts-e=yo.

1sg-run-perf=loc ‘I ran (to) there.’ *Goal

• b. N-iruts-e=mwo.

1sg-run-perf=loc ‘I ran inside of somewhere (e.g. the house). General Location (12) kwiruk-ir-a : to run to

• a. N-iruk-iy-e=yo.

1sg-run-appl-perf=loc ‘I ran (to) there.’ Goal

• b. N-iruk-iy-e=mwo.

1sg-run-appl-perf=loc ‘I ran into there.’ Goal

• The clitic =yo is only available when there is a goal, which is absent in (11), meaning

that (11a) is ungrammatical. It is acceptable in (12a), where the locative applicative licenses a goal.

• The clitic =mwo is permissible for both applied and non-applied variants but with

different interpretations. With the non-applied verb, it describes that the running is happening inside of a single location (e.g. inside the gym).

• When the applicative is present, the =mwo clitic indicates that there is a change in
• location. In (12b), there is a change in location from outside to inside, cf. (9).

5

• (13)

kwinjira : to enter

• a. N-di

1sg-cop kw-injir-a=mwo. inf-enter-imp=loc ‘I am entering (e.g. the house).’

• b. N-di

1sg-cop kw-injir-a=yo. inf-enter-imp=loc ‘I am entering (e.g. a country).’ (14) kwinjir-ir-a : to enter through

• a. N-di

1sg-cop kw-injir-ir-a=mwo. inf-enter-appl-imp=loc ‘I am entering through somewhere (e.g. a window).’

• b. N-di

1sg-cop kw-injir-ir-a=yo. inf-enter-appl-imp=loc ‘I am entering through somewhere (e.g. Canada en route to America).’

• Positional verbs may appear with the locative clitics in their non-applied form. With

the applicative, the clitics are disallowed. (15) kw-icara : to sit

• a. N-di

1sg-cop kw-icar-a=yo. inf-sit-imp=loc ‘I am sitting there (e.g. in my house).’

• b. N-di

1sg-cop kw-icar-a=mwo. inf-sit-imp=loc ‘I am sitting there (e.g. inside my car).’

• c. N-di

1sg-cop kw-icar-a=ho. inf-sit-imp=loc ‘I am sitting there (e.g. on a chair).’ (16) kw-icar-ir-a : to sit on

• a. *N-di

1sg-cop kw-icar-ir-a=yo. inf-sit-imp=loc ‘I am sitting there (e.g. in my house).’

• b. *N-di

1sg-cop kw-icar-ir-a=mwo. inf-sit-imp=loc ‘I am sitting there (e.g. inside my car).’

• c. *N-di

1sg-cop kw-icar-ir-a=ho. inf-sit-imp=loc ‘I am sitting there (e.g. on a chair).’ 6

4 Verb Class Interactions

• The four interpretations of the locative applicative above do no apply arbitrarily, but

rather the interpretation of the applied element is contingent upon the type of verb.

• Specifically, the meaning is contingent upon the underlying locative meaning encoded

by the verb.

• 1. With verbs that encode no location in their meaning, the applicative licenses a

general location.

• 2. Evidence for this is the inability for non-applied ku-vuga ‘talk’ to appear with

locative clitics. (17)

• a. Yohani

John a-ri 1S-be ku-vug-a. inf-talk-imp ‘John is talking.’

• b. Yohani

John a-ri 1-cop ku-vug-a(=*ho/yo/mo). inf-talk-imp=loc ‘John is talking (there). – As seen in (3), repeated in (23a), the applied variant licenses a general loca- tion. (18) Yohani John a-ri 1S-be ku-vug-ir-a inf-talk-appl-imp mu in nzu. house ‘John is talking in the house.’

• 3. The goal interpretation is reserved for for verbs of manner of motion, where the

goal is implicit in the meaning of the verb, such as kw-iruka ‘run’, gu-tembera ‘to go about’, ku-jya ‘to go’, gu-simbuka ‘to jump’. – Evidence of the underlying goal in manner of motion verbs comes from the fact that the goal can be grammatically licensed by the verb without an applicative. – Certain manner of motion verbs are ambiguous between a static location reading and a change of location reading. (19) Yohani John y-a-simbuts-e 1-pst-jump-perf mu in mazi. water ‘John jumped while in the water.’ ‘John jumped into the water.’ – The applied variant, however, is restricted to the goal reading: (20) Yohani John y-a-simbuk-iy-e 1-pst-jump-appl-perf mu in mazi. water ‘John jumped into the water.’ #‘John jumped while in the water.’ 7

• 4. When the applicative licenses a path, it is with verbs of change of location, such

as kw-injira ‘to enter’, gu-sohoka ‘exit’, ku-manuka ‘descend’, kuzamuka ‘ascend’, ku-rira ‘to climb’.

• 5. The sub-location reading is restricted to positional verbs like kw-icara ‘to sit’,

ku-ryama ‘lie down’. Table 1 applied meaning verb type general location no location goal manner of motion path change of location sub-location positional

5 Analyzing Locative Applicatives

• The previous two sections have shown that the meaning of a locative applied object is

contingent upon the meaning of the applicative.

• Previous approaches that assume that the applicative adds a new locative thematic

role do not capture the nuanced interpretations of the applied object with different verb classes.

• I propose instead analyzing applicative morphology as an argument alternation, where

the applied variant of a verb requires a monotonically stronger truth condition on the arguments of the verbal predicate (cf. Beavers 2010).2

• I show that there are two ways that this semantic strengthening effect is found in

locative applicatives, dependent upon the nature of the meaning of the verb.

• 1. Add a wholesale new locative object to the meaning of the verb
• 2. Realize an argument that is semantically encoded by the verb but not realized

syntactically

• This analysis predicts that the applied variant should include all of the information in

the non-applied variant with the addition of the applied element.

• I assume a neo-Davidsonian style semantics, where each participant is linked to a the

event by a specific thematic role (ag, th, path, goal, etc.). – Due to time restrictions, I do not give a fully articulated formalization of the mapping between syntax and semantics. – Crucially for this talk, any entity linked to a thematic role is a syntactically realized argument, e.g. ag(john’, e).

2By “monotonic” I mean that a new meaning is added without removing any prior meaning in the base

predicate (Koontz-Garboden 2007, 2012).

8

• For verbs that encode no location in the non-applied variant, a general location is en-

coded in the applied variant, strengthening the truth-conditional content by specifying the location at which the event took place. – For a verb like ku-vuga ‘talk’ there is no location encoded in the semantics of the verb, as shown by the inability to use a locative clitic: (21) Yohani John a-ri 1-cop ku-vug-a(=*ho/yo/mo). inf-talk-imp=loc ‘John is talking (there). – The meaning of kuvuga ‘talk’ is as follows: (22) ∃e.[talk′(e) & ag(john′, e)] – The applied variant licenses a locative object. (23)

• a. Yohani

John a-ri 1S-be ku-vug-ir-a inf-talk-appl-imp mu in nzu. house ‘John is talking in the house.’

• b. ∃e.[talk′(e) & ag(john′, e) & loc(house′, e)]
• Verbs which encode implicit locations behave slightly differently.
• In these languages, the applicative is used to overtly realize the locational meaning

implicit in the verb, strengthening the truth-conditional content by denoting a specific location. – For example, the verb kw-injira ‘to enter’ encodes a path, though this path is not realized syntactically. – Here, I notate semantically implicit location information by introducing a location (l) with an existential quantifier. When an element is overt, it is bound off with a specified location. (24)

• a. Yohani

John a-ri 1-cop kw-injir-a inf-enter-imp mu in nzu. house ‘John entered the house.’

• b. ∃e.[enter′(house′, e) & ag(john′, e) & th(house′, e) & ∃l.path(l, e)]

– When the applicative is used on kw-injira ‘enter’, the implicit path argument becomes obligatorily encoded in the syntax. (25)

• a. Yohani

John a-ri 1-cop kw-injir-ir-a inf-enter-appl-imp mu in muryango door mu in nzu. house ‘John entered through the door.’

• b. ∃e.[enter′(house′, e) & ag(john′, e) & th(house′, e) & path(door′, e)]

9

5.1 A Further Prediction

• My analysis is that an applicative requires stronger truth-conditional content, which

I showed is satisfied by either adding an object or syntactically encoding a meaning found in the verb.

• The insight of this analysis is that the semantic strengthening required by the applica-

tive subsumes the object-adding analysis that has been the mainstay of research on the topic.

• This semantic analysis predicts that the applied variant need not always license an

additional object, but may change the semantic nature of an existing thematic object

• f the verb, provided there is a stricter semantic meaning in the applied variant.
• The thematic roles of goal and recipient are in such a relation; namely, recipients have

all the entailments of goals, but with the additional meaning of prospective change of possession (Beavers 2011). – Goal = a place to which motion is directed – Recipient = a place to which motion is directed + prospective change of possession

• This means that with verbs that encode a goal, it should be possible to use the ap-

plicative to change the goal to a recipient with an applicative morpheme — without modifying the argument structure of the verb.

• This prediction is borne out with the verb gu-tera ‘to throw’, which is ditransitive in

its non-applied form. (26) Kyle Kyle y-a-tey-e 1-pst-throw-perf i-buye 5-rock John. John ‘Kyle threw the rock at John.’

• The sentence in (26) means that Kyle is pelting a rock at John, possibly trying to

harm him, and, crucially, without the intention of giving John possession of the rock

• The meaning of this sentence is as follows:

(27) ∃e.[throw′(rock′, john′, e) & ag(kyle′, e) & th(rock′, e) & goal(john′, e)]

• My analysis, in conjunction with the relationship of goals and recipients just outlined,

predicts that in an applied sentence, John is the recipient of the throwing event.

• This is exactly the reading that is obtained:

(28) Kyle Kyle y-a-ter-ey-e 1-pst-throw-appl-perf i-buye 5-ball John. John ‘Kyle threw the rock to John.’ 10

• Here, the required reading is one where Kyle is attempting to give John possession of

the rock. (29) ∃e.[throw′(john′, kyle′, e) & ag(kyle′, e) & th(rock′, e) & rec(john′, e)]

• Evidence for the distinction comes from the fact that prospective catching can only be

modified when the object has been “upgraded” to a recipient, as in (30a).

• In (30b), where there is no applicative, the object is not a recipient; hence, it is infe-

licitous to modify any notion of John trying to catch the rock. (30)

• a. Kyle

Kyle y-a-ter-ey-e 1S-pst-throw-appl-perf i-buye 5-rock John, John, ariko but John John nti-y-a-ri-fash-e. neg-1S-pst-5O-catch-perf ‘Kyle threw the rock to John, but John didn’t catch it.’

• b. #Kyle

Kyle y-a-tey-e 1sg-pst-throw-perf i-buye 5-rock John, John ariko but John John nti-y-a-ri-fash-e. neg-1S-pst-5O-catch-perf ‘Kyle threw the rock at John, but John didn’t catch it.’

• This analysis predicts that if the applicative strengthening is being used to introduce
• r strengthen another element in the sentence, only the goal reading is possible with

gu-tera ‘throw’.

• Consider the introduction of a locative applicative. It is predicted that in an applied

locative sentence with the verb gu-tera ‘throw’, the DO cannot be a recipient, but rather a goal.

• This prediction is borne out:

(31) Kyle Kyle y-a-ter-ey-e 1S-pst-throw-appl-perf i-buye 5-rock John John mu in nzu. 9.house ‘Kyle threw the rock at John in the house.’

• Evidence that the applicative is being used to license the locative comes from the scope

interpretation.

• For example, in (32a), the locative PP scopes over the object NP, and optionally

the subject. When the applicative is used, as in (32b), however, the subject is also

• bligatorily in the location described by the PP.

11

(32)

• a. Umuyobozi

chief y-ubats-e 1S-build-perf inzu house mu in mujyi. town ‘The chief built the house in town, (and his location is not explicit).’

• b. Umuyobozi

chief y-ubak-iy-e 1S-build-appl inzu house mu in mujyi. town ‘The chief built the house in town, (and the chief is obligatorily in the town).’

• The scope differences in (32) arise due to the applicative licensing a new object.
• Evidence that the locative phrase in (31) is an object is that the only interpretation is
• ne where all participants are in the house, i.e. the scope is over all participants.
• The interpretation in (31) that all participants are in the house shows that the ap-

plicative is being used to license the locative, which correctly predicts that the object is obligatorily a goal and not a recipient.

5.2 A Potential Counterexample

• In Lubukusu (Bantu, spoken in Western Kenya), there is an alternation with the ap-

plied and non-applied variants of the verb where applied variant changes the direction- ality of the location. (33)

• a. ∅-a-rum-a

1S-pst-send-imp ebarwa letter mu in posta. post.office ‘S/he sent the letter to the post office.’

• b. ∅-a-rum-er-a

1S-pst-send-loc-imp ebarwa letter mu in posta. post.office ‘S/he sent the letter from the post office.’ (Lubukusu)

• Source and goal locatives (from and to) do not lie in the same relationship as recipients

and goals discussed above, suggesting that the semantic strengthening analysis may not be at play here.

• However, an alternative analysis is that the applied variant in (33b) is creating the

same scope over the subject as in (32b) above; the reading difference arises because the locative, when applied, has scope over the entire sentence.

• This analysis naturally captures the data in (33) by analyzing this situation as a change

in the scope of the locational PP. 12

6 Conclusion

• This talk argued that applicative morphology is an operation that strengthens the

truth conditions of a verb’s arguments.

• Furthermore, I showed that any analysis must take into account the interaction of verb

class with applicative morphology.

• Centrally, I have shown that careful attention to the semantic meanings of applied

sentences is crucial for capturing the syntactic and semantic properties of applicatives.

• As a final note, investigating the semantics of verbs may shed light on the complex

question of object symmetry mentioned in §2.

• Preliminary data from Lubukusu show that ingestive verbs pattern differently than
• ther verb classes with respect to symmetry.
• The general pattern appears to be that morphological causatives are asymmetrical,

where the thematic object cannot undergo objecthood tests such as passivization. (34)

• a. Omu-khangarani

1-warrior ∅-a-p-isi-bw-a 1S-pst-hit-caus-pass-imp li-sisi 5-wall ne by

• mw-ekesi.

1-teacher ‘The warrior was made to hit the wall by the teacher.’ b.*Li-sisi 5-wall ly-a-p-isi-bw-a 5S-pst-hit-caus-pass-imp

• mu-khangarani

1-warrior ne by

• mw-ekesi.

1-teacher ‘The wall was made to be hit by the warrior by the teacher.’

• However, with ingestives such as kunywa ‘drink’ the pattern is symmetrical:

(35)

• a. Kyle

Kyle ∅-a-nyw-esy-ebw-a 1S-pst-drink-caus-pass-imp kamalwa beer ne by Mama Mama Leo. Leo ‘Kyle was made to drink the beer by Mama Leo.’

• b. Kamalwa

beer k-a-nyw-esy-ebw-a 6S-pst-drink-caus-pass-imp Kyle Kyle ne by Mama Mama Leo. Leo ‘The beer was made to be drunk to Kyle by Mama Leo.’

• These data show that verb class can have an affect on the symmetry patterns within a

specific language, which may answer many of the unresolved questions in the literature

• n object symmetry.

References

Alsina, A., 1992. On the argument structure of causatives. Linguistic Inquiry 23, 517–555. Alsina, A., Mchombo, S., 1993. Object asymmetries and the Chicheˆ wa applicative construc-

• tion. In: Mchombo, S. (Ed.), Theoretical Aspects of Bantu Grammar. Stanford: CSLI

Publications, pp. 17–45. 13

Baker, M., 1988. Theta theory and the syntax of applicatives in Chichewa. Natural Language and Linguistic Theory 6, 353–389. Baker, M., Collins, C., 2006. Linkers and the internal structure of vP. Natural Language and Linguistic Theory 24, 307–354. Baker, M., Safir, K., Sikuku, J., 2012. Sources of (a)symmetry in Bantu double object

• constructions. In: Arnett, N., Bennett, R. (Eds.), Proceedings of the 30th West Coast

Conference on Formal Linguistics. Somerville, MA: Cascadilla Proceedings Project, pp. 54–64. Beavers, J., 2010. The structure of lexical meaning: Why semantics really matters. Language 86, 821–864. Beavers, J., 2011. An aspectual analysis of ditransitive verbs of caused possession in English. Journal of Semantics 28, 1–54. Bond, O., 2009. The locative applicative in Eleme. Transactions of the Philological Society 107, 1–30. Bresnan, J., Moshi, L., 1990. Object asymmetries in comparative Bantu syntax. Linguistic Inquiry 21, 147–185. Cann, R., Mabugu, P., 2006. Constructional polysemy: the applicative construction in chiS- hona. In: The Cognitive Basis of Polysemy, Metalinguistica, vol. 19. Peter Lang, pp. 221–245. Creissels, D., 2004. Non-canonical applicatives and focalization in Tswana. Unpublished ms. Fillmore, C. J., 1970. The grammar of Hitting and Breaking. In: Jacobs, R., Rosenbaum, P. S. (Eds.), Readings in English Transformational Grammar. Ginn, Waltham, pp. 120–133. Jeong, Y., 2007. Applicatives: Structure and Interpretation from a Minimalist Perspective. Amsterdam:John Benjamins. Jerro, K., in press. Revisiting object symmetry in Bantu. In: The Selected Proceedings of the 44th Annual Conference on African Linguistics. Somerville, MA: Cascadilla Proceedings Project. Kimenyi, A., 1980. A Relational Grammar of Kinyarwanda. University of California Press. Koontz-Garboden, A., 2007. States, Changes of State, and the Monotonicity Hypothesis. PhD Dissertation, Stanford University. Koontz-Garboden, A., 2012. The monotonicity hypothesis. In: McNally, L., Demonte, V. (Eds.), Telicity Change and State: A Cross-categorial View of Event Structure. Oxford: Oxford University Press, pp. 139–161. Levin, B., 1993. English Verb Classes and Alternations: A Preliminary Investigation. Chicago: University of Chicago Press. Marantz, A., 1993. Implications of asymmetries in double object constructions. In: Theo- retical aspects of Bantu Grammar. Stanford: CSLI Publications, pp. 113–148. Marten, L., 2003. The dynamics of Bantu applied verbs: An analysis at the syntax-pragmatics

• interface. In: L´

ebikaza, K. K. (Ed.), Acts du 3e Congr` es mondial de Linguistique Africaine Lom´ e 2000. K¨

• ppe, K¨
• ln, pp. 207–221.

Marten, L., Kula, N., Thwhala, N., 2007. Parameters of morphosyntactic variation in Bantu. Transactions of the Philological Society 105, 253–338. McGinnis, M., 2001. Variation in the phase structure of applicatives. Linguistic Variation Yearbook 1, 105–146. McGinnis, M., Gerdts, D., 2003. A phase-theoretic analysis of Kinyarwanda multiple ap- 14

• plicatives. In: Proceedings of the 2003 Canadian Linguistic Association Annual Conference

Department of Linguistics. Universit´ e du Qu´ ebec ` a Montr´ eal, pp. 154–165. Pylkk¨ anen, L., 2008. Introducing Arguments. Cambridge: MIT Press. Rappaport Hovav, M., Levin, B., 2008. The English dative alternation: A case for verb

• sensitivity. Journal of Linguistics 44, 129–167.

Zeller, J., 2006. Agreement and the EPP in Kinyarwanda applicatives. In: Brandt, P., Fuß,

• E. (Eds.), Form, Structure, and Grammar. Berlin: Akademie-Verlag, pp. 275–295.

Zeller, J., Ngoboka, J. P., 2006. Kinyarwanda locative applicatives and the Minimal Link

• Condition. Southern African Linguistics and Applied Language Studies 24, 101–124.

15