Eunsuk Lee, University of North Carolina at Chapel Hill
Typologies of Events and Typologies of Morphemes
Linguistics has a long tradition of classification of predicates in terms of aspectual properties. Since Dowty (1979), Vendler’s (1967) aspectual classification has been integrated into research on typologies of languages. In Ritter & Rosen (2000), it is argued that different languages represent events in different ways. Starting from the idea that the typology of predicates serves as a useful typology of languages, Ritter & Rosen (2000) propose their “event-based language typology” partitions languages into two sets; I-languages and D-languages. Although the conceptualization of an event involves both an initial and a terminal bound, Ritter & Rosen argue that it takes only a single bound to identify an event. In I-languages, this bound is the initial bound, thus activities and accomplishments are treated as events. In D-languages, this bound is the terminal bound, thus achievements and accomplishments are treated as events. To support this typology, Ritter & Rosen offer evidences drawn from many languages such as the grammatical distinction between initiating (agentive) and non-initiating (non-agentive) subjects in I-languages and the grammatical distinction between delimiting and non-delimiting direct objects in D-languages.
In this paper, however, I argue against their claim that languages can be typed in terms of the way they choose to represent events. Contrary to Ritter & Rosen’s claim, the eventive interpretation of a sentence is not determined by whether the language is classed as a D- or I-language. The conceptual problem is that the properties of their two groups can be observed within the same language. To support this conclusion, I present some data concerning telicity in Chinese, Malagasy, and some data about agentivity from English.
Buying the typology of Ritter & Rosen, the event of D-languages such as English and Chinese is necessarily telic in the sense of having an endpoint, since they always require events to have the terminal bound, as shown in the English example of (1). However, the Chinese presents a serious problem that cannot be explained by Ritter & Rosen. The Chinese sentences corresponding to (1) are presented in (2). As shown in below, Chinese can express (1) in two different ways; with –le, and without –le. The sentence with –le makes it paralleled its English counterpart in (1) in being semantically anomalous. However, the same sentence can appear without –le, and in that context it becomes acceptable. When –le drops, (2b) has an atelic meaning in which the terminal bound of the event does not matter. Yet this is precisely the characteristic of that Ritter & Rosen attribute to I-languages, rather than to D-languages. The surprising upshot is that Chinese has properties of two language types in terms of Ritter & Rosen. Their claim that languages can be typed in terms of the way a language provides an eventive interpretation cannot be maintained. In fact, the typology is relevant to morphemes - allowing a language like Chinese to have morphemes of both types.
Chinese is not an isolated example. We also find I-languages that have properties of Ritter & Rosen’s D-languages as well as morphemes with properties of I-languages. Malagasy, an Austronesian language, is such an I-language. As shown in below, in Malagasy, the first event of (3) with the prefix (m)an- is not necessarily telic meaning which gives us the delimited interpretation of events. Following Ritter & Rosen (2000), (3) does not show the property of D-languages. Then, Malagasy has to be an I-language, since in I-languages, the terminal bound of events does not matter. However, in (4), the use of prefix (m)aha- makes us consider Malagasy as D-language like English. This is because in (4) the first event has the telic meaning which gives us the delimited interpretation of events. That is, the completion of the event cannot be denied in sentences with the prefix (m)aha-, whereas it can be denied in sentences with a different prefix such as (m)an-. In Malagasy, it is also the morpheme that the typology is relevant to. Then, we may ask ourselves a question whether any language is consistently D- or I-language. In fact, every language that we talk about has mixed properties.
In addition to this problem undercutting Ritter & Rosen’s typology, we can also find another problem from English. According to them, in D-languages such as English, achievements and accomplishments form a natural class. However, they do not form a natural class with respect to agentivity. In fact, activities can class with accomplishments even in English which is supposed to be not an I-language but a D-language, since in (5) only agentive predicates are appropriate in the complement to persuade.
Data
(1) *I bribed the mayor, but he refused to accept the bribe.
(2) a.*wo huilu-le shizhang, danshi ta jujue jieshou huilu.
I bribe-aspect mayor but he refuse accept bribe
b. wo huilu shizhang, danshi ta jujue jieshou huilu.
I bribe mayor but he refuse accept bribe
(3) a. namory ny ankizy ny mpampianatra
pst.an.meet the children the people
“The teachers gathered the children”
b. … nefa tsy nanana fotoana izy
… but neg pst.have time they
“… but they didn’t have time.” (Travis 2000: 172)
(4) a. nahavory ny ankizy ny mpampianatra
pst.a.ha.meet the children the people
“The teachers gathered the children”
b. *… nefa tsy nanana fotoana izy
… but neg pst.have time they
“… but they didn’t have time.” (Travis 2000:173)
(5) a.#Jones persuaded the couple to be happy. (states)
b. Jones persuaded Tina to walk in the park. (activities)
c.#Jones was persuaded to notice the marks on the wallpaper. (achievements)
d. Jones persuaded him to ea the pie. (accomplishments) (Kearns 2000: 211)
Reference
Dowty, David. 1979. Word Meaning in Montague Grammar. Dordrecht: Reidel.
Kearns, Kate. 2000. Semantics. New York, St. Martin’s Press.
Ritter, Elizabeth & Sara Rosen. 2000. Event Structure and Ergativity. In Carol Tenny & James Pustejovsky (ed.) Events as Grammatical Objects, 187-238. CSLI publications.
Tenny, Carol. 1994. Aspectual Roles and the Syntax-Semantics Interface. Dordrecht: Kluwer.
Travis, Lisa. 2000. Event structure in syntax. In Carol Tenny & James Pustejovsky (ed.) Events as grammatical objects, 145-185. CSLI publications.
Vendler, Zeno. 1967. Linguistics in Philosophy. Ithaca, N.Y.: Cornell University Press.
