Adverbs of Change, Aspect, and Anaphoricity
Todor Koev, University of Düsseldorf
The puzzle A DVERBS OF CHANGE like quickly or slowly (also called “adverbs of space and
time” in Cresswell 1978) are typically classified as MANNER adverbs that characterize the rate at
which the described action evolves (Jackendoff 1972; Bellert 1977; Parsons 1990; Ernst 2004). For
example, (1) characterizes as quick the way Selena moved her body.
(1)
Selena ran quickly.
However, adverbs of change are known to give rise to a number of non-manner interpretations
as well (Cresswell 1978; Travis 1988; Pustejovski 1991; Shaer 1998; Tenny 2000; Schäfer 2002;
Ernst 2004; Thompson 2006; Kearns 2007; Eszes 2009; Rawlins 2013). D URATION readings modify the temporal extent of the whole event (2) and ANAPHORIC readings measure the temporal
distance between (some part of) the described event and some previously mentioned event (3).
(2)
Selena finished the assignment quickly.
(3)
The professor walked in. Selena quickly noticed him.
There are systematic and revealing correlations between available interpretations and lexical aspect
that have not been fully analyzed in the literature but give important insight into event modification
and aspectual structure. When modifying activity predicates, adverbs of change can only have manner readings (1). With accomplishment predicates, such adverbs are ambiguous between manner
vs. duration readings (4), and can also take on anaphoric readings (5). When modifying achievement predicates, adverbs of change are read anaphorically (3), and they are generally incompatible
with stative predicates (6). Table 1 on p.3 summarizes the full pattern.
(4)
Jack built the house quickly.
a. manner: Jack moved fast while he was building the house.
b. duration: The event of Jack’s building the house took a short period of time.
(5)
The tiger walked into the room. Lenard quickly moved to the window.
(6)
?She quickly liked her job.
Previous work Cresswell (1978) proposes that quickly modifies motion predicates and measures
the ratio between the distance traveled and time passed. He attributes manner/duration ambiguities
to a structural ambiguity whereby quickly attaches to the verb alone or the entire VP (respectively),
but fails to explain anaphoric readings. Rawlins (2013) derives the manner/duration contrast by
letting adverbs of change distribute over atomic events relative to an unspecified event property,
which is contextually resolved to the verb or the VP. He argues that when read anaphorically,
adverbs of change modify so-called “narrative events” (eventive counterparts of Reichenbachian
“reference times”), yet ignores the role of aspect. But if narrative events were real, there would be
no reason why adverbs of change with activities or states cannot have anaphoric interpretations.
Proposal A dynamic semantic account can overcome the main stumbling block for previous
analyses of adverbs of change, i.e. anaphoric readings. I develop a compositional update semantics
that represents events in a neo-Davidsonian fashion (e.g. Parsons 1990) and facilitates discourse
1
anaphoric dependencies (cf. Muskens 1996; Bittner 2011). The derived representation for (7) is as
shown (e is an event discourse referent introduced through existential closure; x and P are referents
introduced by the subject and the verb, respectively; ∧ is read dynamically or left-to-right).
(7)
Kurt ran ↝ ∃e ∧ ∃x ∧ x = kurt ∧ ag(e) = x ∧ running(e) ∧ ∃P ∧ P = λe′ . running(e′ )
I propose that adverbs of change modify the temporal distance between an ANCHOR and a TARGET
event. The target is the culmination of the described event e (marked as cul(e)) while the anchor
(marked as a) is resolved anaphorically and is temporally anterior to the target. Specifically, quickly
measures the temporal distance (δ) between two events and compares it to the arithmetic mean (µ)
of alternative distances, given different values for a and e provided by assignments in the information state (8). To derive manner readings, I assume that quickly freely combines with a covert
distributivity operator DX (9), where X is an unbound event predicate variable and atom(e, X)
stands for the set of atoms in e that still fall under X. The result of composing DX with quickly
is shown in (10). (11)-(12) schematize the modified VP, either directly by quickly or through the
intermediary of DX . (It is assumed that in (12) X resolves P , the variable introduced by the VP.)
(8)
(9)
quicklya ↝ λe. δ(a, cul(e)) < µ(a, cul(e))
DX ↝ λP λe. ∀e′ (e′ ∈ atom(e, X), P (e′ ))
(10) DX quicklya ↝ λe. ∀e′ (e′ ∈ atom(e, X), δ(a, cul(e′ )) < µ(a, cul(e′ )))
(11)
(12)
VPP quicklya ↝ λe. ∃P ∧ V P (e) ∧ P = λe′ . V P (e′ ) ∧ δ(a, cul(e)) < σ(a, cul(e))
VPP [DX quicklya ] ↝ λe. V P (e) ∧ ∃P ∧ P = λe′ . V P (e′ ) ∧
∀e′ (e′ ∈ atom(e, P ), δ(a, cul(e′ )) < σ(a, cul(e′ )))
The main claim of this paper is that the attested interpretations for adverbs like quickly naturally fall
out from the temporal ontology of events with different aspectual properties (cf. Moens & Steedman 1988; Kamp & Reyle 1993; Smith 1997). This is modeled as part-whole structures (Bach
1986; Krifka 1992; Link 1998) and is graphically outlined in Figure 1 on p.3, where black dots
mark culminations. First, quickly cannot compose with stative VPs (with or without a distributivity operator) because states are homogenous events that do not culminate, so cul is undefined.
Next, achievements are instantaneous and (11)-(12) produce equivalent interpretations, as in (12)
the distribution is over a singleton set (of the described event). The only predicted reading is thus
an anaphoric one. Finally, activities represent processes and accomplishments add a culmination
point. (Processes are chains of events that have the structure of accomplishments.) (11) is ruled
out for activity VPs because no unique culmination exists (cul is undefined) while (12) derives a
manner reading, provided that a is resolved to the beginning of each e′ . For accomplishment VPs,
(11) produces an anaphoric/duration reading, depending on whether a is resolved to some previous
event or the beginning of the described event (respectively). (12) makes available a manner reading, if X is resolved to the verb alone (not represented above). If X is resolved to P (the entire
VP), we once again get a duration/anaphoric interpretation, depending on how a is resolved.
Extensions Since in manner readings the distributivity operator variable needs to be bound to an
activity main verb, the account correctly predicts that preverbal occurrences of adverbs of change
cannot naturally describe manners (cf. ?Quickly, he ran / ?He quickly ran). Also, the account offers enough flexibility to accommodate the fact that adverbs of change like slowly seem to lack
anaphoric readings (cf. The professor walked in. ?Justin slowly noticed him).
2
activities
accomplishments
achievements
states
manner
3
3
7
7
duration
7
3
7
7
anaphoric
7
3
3
7
Table 1: Readings for adverbs of change with different aspectual classes
states
achievements
activities
accomplishments
●
ppppp ppppp●
Figure 1: Graphical representation of lexical aspect distinctions
Bach, E. (1986). The algebra of events. Linguistics and Philosophy 9: 5–16.
Bellert, I. (1977). On semantic and distributional properties of sentential adverbs. Linguistic Inquiry 8(2):
337–351.
Bittner, M. (2011). Time and modality without tenses or modals. In R. Musan & M. Rathert (eds.). Tense
across Languages, 147–188. Niemeyer.
Cresswell, M.J. (1978). Adverbs of space and time. In F. Guenthner & S. Schmidt (eds.). Formal Semantics
and Pragmatics for Natural Languages, 171–199. Reidel.
Ernst, T.(2004). The Syntax of Adjuncts. Cambridge University Press.
Eszes, B. (2009). Aspect and adverb interpretation – the case of quickly. In K. Kiss (ed.). Adverbs and
Adverbial Adjuncts at the Interface, 269–294. De Gruyter.
Hunter, J. (2013). Presuppositional indexicals. Journal of Semantics 30: 381–421.
Jackendoff, R. (1972). Semantic Interpretation in Generative Grammar. MIT Press.
Kamp, H. & U. Reyle (1993). From Discourse to Logic. Kluwer.
Kearns, K. (2007). Telic senses of deadjectival verbs. Lingua 117: 26–66.
Krifka, M. (1992). Thematic relations as links between nominal reference and temporal constitution. In I.A.
Sag & A. Szabolcsi (eds.). Lexical Matters, 28–53. Stanford.
Link, G. (1998). Algebraic Semantics in Language and Philosophy. CSLI.
McConnell-Ginet, S. (1982). Adverbs and logical form: A linguistically realistic theory. Language 58(1):
144–184.
Muskens, R. (1996). Combining Montague Semantics and Discourse Representation. Linguistics and Philosophy 19:143–186.
Parsons, T. (1990). Events in the Semantics of English. MIT Press.
Pustejovski, J. (1991). The syntax of event structure. Cognition 41: 47–81.
Rawlins, K. (2013). On adverbs of space and time. In B. Arsenijevic et al. (eds.). Studies in the Composition
and Decomposition of Event Predicates, 153–193. Springer.
Reichenbach, H. (1947). Elements of Symbolic Logic. Macmilian.
Schäfer, M. (2002). Pure manner adverbs revisited. Proceedings of Sinn und Bedeutung 6: 311–323.
Shaer, B. (1998). Adverbials, functional structure, and restrictiveness. Proceedings of NELS 28: 391–407.
Smith, C. (1997). The Parameter of Aspect. Kluwer, Second edition.
Tenny, C.L. (2000). Core events and adverbial modification. In C. Tenny & J. Pustejovsky (eds.). Events as
Grammatical Objects, 285–334. CSLI.
Thompson, E. (2006). The structure of bounded events. Linguistic Inquiry 37: 211–228.
Travis, L. (1988). The syntax of adverbs. WPL: Special Issue on Comparative Germanic Syntax. McGill.
3