Colloquium: Malte Zimmermann (Potsdam)

Prof. Dr. Malte Zimmermann from the University of Potsdam will give a colloquium talk on May 8th, at 4pm in SH 3.104.

Title: Find Construction Analyze: Making Sense of Serial Verb Constructions

Abstract: The objectives of the talk are threefold: First, to discuss a range of empirical
diagnostics (Adv-modification, A-quantification, cumulation, negation, reordering) for
elucidating the underlying semantic representation of serial verb constructions (SVCs) across
languages. Secondly, to apply the diagnostics to different types of SVCs in Igbo (BenueCongo,
Niger-Congo). We will show that there is a correlation between SVC-type and
semantic interpretation. Thirdly, we will propose formal semantic accounts of different SVCtypes
in terms of event extension, event conjunction, and event modification, respectively.
Background: SVCs are commonly defined on the basic of formal criteria alone. They involve
juxtaposed sequences of V(P)s with a shared subject, at most one shared grammatical object,
single tense/aspect specification (on V1), and single negation (Veenstra & Muysken 2018):
(1) Uche. gbu-ru o.ku.ko. si-e (= S V1 O V2) [Igbo]
Uche kill-pst chicken cook-sfx
‘Uche killed the chicken and cooked it.’
There is a general consensus that SVCs such as (1) involve some combination of subevents or
event specifications, and that SVCs may fall into several informal semantic types (Dechaine
1993, V&M18): (i.) Dative/Directional; (ii.) causative/instrumental; (iii.) resultative; and (iv.)
multi-event serial. In contrast, formal event semantic criteria have received little attention in
the analysis of SVCs (V&M18), with the notable exception of Dechaine (1993), Stewart
(1998) and Stewart & Baker (2001) for Yoruba, Igbo and Edo.
Types of Event Composition: Stewart (1998) analyses different types of SVCs in terms of
the three event-compositional processes in (2); see also Anderson (2017) on cumulation. To
these symmetric processes, we add Event Extension in (2d), combining two independent
subevents into a complex event with an asymmetric relational (eg temporal) substructure
(2)a. Event Modification (resultatives): ∃e [P1(e) ∧ P2(e)] (1 event)
b. Event Cumulation (multi-events, OBJ-sharing)
∃e [e=e1⊕e2 ∧ P1(e1) ∧ P2(e2)] (2 subevents)
c. Event Conjunction (multi-events, no OBJ-sharing):
∃e1 [P1(e1)] ∧ ∃e2 [P2(e2)] (2 events)
d. Event Extension: ∃e1e2 [P1(e1) ∧ e1<Re2 ∧ P2(e2)]
Whereas such event extension is standardly assumed for the interpretation of A-quantifiers
(von Fintel 1995), it has so far not been employed in the analysis of SVCs. Concretely, we
argue that OBJ-sharing multi-event SVCs like (1) do not involve event cumulation (2b) but
event extension (2d). To this end, we will show that the event semantics of OBJ-sharing
SVCs (1) differs from that of SVCs with no structural OBJ-sharing, cf. (3) with a pronoun:
(3) Uche gbu-ru o.ku.ko.1 si-e ya1 (= S V1 O V2 PRON/NP) [Igbo]
Uche kill-pst chicken cook-sfx 3sg
‘Uche killed the chicken and cooked it.’
Diagnostics for Event Composition: We will discuss three diagnostics that are sensitive to
the event semantic structure of SVCs: contradictory adverbs (4a), sub-quantification with Aquantifiers
(4b), and cumulation (4c): Cumulative SVCs should allow for plural subjects to
distribute over sub-events. Negation facts and reordering will serve as auxiliary diagnostics.
(4)a. S/he kill chicken quickly cook slowly. b. S/he kill chicken sometimes cook
c. Uche and Obi kill chicken cook ⇒?? Uche killed and Obi cooked the chicken
The three main diagnostics square with the four event-compositional processes as follows:

SVCs in Igbo: Applying the diagnostics, multi-event SVCs with and without OBJ-sharing
show a different behavior. This is shown for contradictory adverbs and A-quantifiers in (5)
and (6). (5b/6b) pattern with multiple SVCs where V1 and V2 select for distinct OBJ-NPs.
(5)a.*Uche gbu-ru o.ku.ko.i o.si.i.so. sie nwayo.o. nwayo.o. +OBJ share: *
b. Uche gbu-ru o.ku.ko.i o.si.i.so. sie yai nwayo.o. nwayo.o. – OBJ-share: 
Uche kill-pst chicken quickly cook ∅/3SG slowly
‘Uche killed the chicken quickly and cooked it slowly.’
(6)a.*Uche na-egbu o.ku.ko.i oge u.fo.du.sie +OBJ share: *
b. Uche na-egbu o.ku.ko.i oge u.fo.du. sie yai – OBJ-share: 
Uche HAB-kill chicken sometimes cook ∅/3SG
‘Uche regularly kills chicken, sometimes cooking them.’
Neither construction type allows for cumulative distribution over V1V2-subevents: (7) cannot
mean that Uche caught the fish and Obi cooked it, as incorrectly predicted on the cumulative
event construal in Stewart (1998), Anderson (2017).
(7) Uche na Obi ko.-ta-ra azu.i sie __ / yai
U & O catch-dir-pst fish cook
Not: ‘Uche caught the fish and Obi cooked it.’
We model the two SVC types in (1) and (3) in terms of event extension at the V-level
(Ramchand 2008) (8a) and covert event conjunction of VPs (8b), respectively. With minimal
event extension ‘</Ext’ (von Fintel 1995) in (8ab), agent and theme of e1 and e2 are the same:
(8)a. [[ (1)]] = ∃e1e2 [ kill(e1) ∧ AG(Uche,e1) ∧ TH(chicken,e1) ∧ e1<e2 ∧ cook(e2)]
LF: ∃ [ V1 kill chicken [Ext< [V2 cook ]]]
b. [[ (3)]]g = ∃e1 [ kill(e1) ∧ AG(Uche,e1) ∧ TH(chicken,e1)] ∧ ∃e2 [cook(e2) ∧ AG(Uche,e2)
∧ TH(g(1),e2)
LF: ∃ [VP3 [VP1 kill chickeni] [ & [VP2 eat PRONi ]]
Other SVCs-types: Instrumental SVCs (Uche hold knife cut meat) show the behavior of
semantic event conjunction (8b): two juxtaposed subevents are conjoined. In contrast, dative
and resultative SVCs (Uche buy book give Ada, Uche hit break plate) seem to involve
modification of single events. Resultatives reflect this syntactically in the form of Vcompounding.
Finally, directionals (Uche run leave town, Uche carry firewood go market) do
not form a unified structural or semantic class: directional concepts are either expressed as
V+V-resultatives or as multi-event SVCs without OBJ-sharing.
Outlook: This formal event-semantic analysis based on the diagnostics in (4) allows for a
better understanding of the underlying structures and the semantic composition of different
SVC-types in Igbo, as well as for cross-linguistic comparison with SVCs in other languages.
Refs: Anderson, C.J. (2017).The andative and venitive construction in San Lucas Quiaviní
Zapotec. TALK at SVC-workshop, HU Berlin. Déchaine, R. (1993). Predicates across
categories. PhD thesis, UMASS. Von Fintrel, K. (1995). A minimal theory of adverbial
quantification. Ms, UMASS. Ramchand, G. (2008). Verb meaning and the lexicon: A firstphase
syntax. Cambridge: CUP. Stewart, O.T. (1998). The SVC Parameter. PhD, McGill.
Veenstra, T. & P. Muysken (2018). Serial Verb Constructions. Blackwell Companion Syntax.