1.1 SPACE ACROSS LANGUAGES

It is important as well as interesting to study, first of all, the domain of space from a typological perspective.Given that human beings share a universal biological heritage in spatial understanding, it is reasonable to speculate that the language of space will “closely mirror the contours of nonlinguistic spatial understanding” （Bowerman 1999: 387）.However, it is found that languages actually differ considerably in how they describe space.To give an example, human beings, irrespective of language and culture, equally possess the ability to spatially locate an object in relation to another, but the interpretation of these spatial relationships varies strikingly from language to language.The spatial relationship between a man and a car, for example, can be encoded as ‘a man stands in frontof the car’, ‘a man stands to the leftof the car’; or ‘a man stands to the westof the car’, depending on which frame of reference （i.e.intrinsic, relative or absolute） a particular language adopts for locating a target.It is on such a basis that Levinson （2003） suggests that languages can be classified into five major groups according to which coordinate system, or which combination of different coordinate systems, the language systematically employs in spatial description （see Levinson 2003: 93 for details）.

In the specific domain of motion, it is found that a more or less universal set of semantic information is expressed in all languages, which includes manner of motion （e.g. to skip,to climb , to crawl ）, directionality of motion （e.g.up , down , across , into ）, cause of motion （e.g.to push , to pull,to drag ） and deixis （e.g.to come ）.At the same time, however, the degree of prominence attached to a given type of semantic information varies greatly across languages.For example, the description of a spatial event involving a change of location in English, such as A man runs across the roadwill be more typically represented in French as C'est un homme qui traverse la route en courant（‘It's a man who crosses a road by running’）.More specifically, in languages such as English, the manner of motion is foregrounded and expressed in the verb, whereas in languages such as French, the path of motion is rendered more semantically salient in the verb, with the manner information expressed （if at all） in the periphery.Based on the different ways in which the semantic components for motion are expressed across an utterance-particularly with regard to the expression of Path-Talmy （1991, 2000） proposes a general typological framework within which languages mainly fall into two broad categories: satellite-framed and verb-framed.In the satellite-framed language, the verb typically conflates Motion and Manner of motion and/or Cause of motion, while Path is encoded outside the verb in a satellite （e.g.through the use of particles and affixes）.By contrast, in the verb-framed language the verb typically conflates Motion and Path, while Manner and Cause are expressed separately in an adverbial or a gerund （when expressed at all）.Examples （1） and （2）, below, illustrate the difference.

（1）Satellite-framed （e.g.English and other Germanic languages）

a.The ball rolled into the hole .

b.The boy pushed the ball into the hole .

（2）Verb-framed （e.g.French and other Romance languages）

La balle est entrée dans le trou en roulant.

the ball has entered in the hole by rolling

‘The ball entered the hole by rolling.’

In example （1a） above, the verb roll combines the fact of the ball's movement with the way in which it moves, namely, to move forward along a surface by repeatedly turning over.In a similar vein, in （1b）, the verb pushdenotes not only the fact that the ball moves but also the reason for this movement, namely, an outward pressure or force is being exerted against the ball for the purpose of altering its location.In contrast, in the French example （2）, the verb est entrée（‘entered’） indicates that the motion occurs along a path and reaches an end boundary, but it does not indicate in which manner this motion takes place.The latter information is actually given separately in the gerund en roulant（‘by rolling’）.

This typological framework as proposed by Talmy fits particularly well with Indo-European languages （e.g.English, German, French, Spanish） where a main verb can be morphologically distinguished from its supporting elements such as particles.However, when we look at some other languages like Chinese, the shortcoming of this bipartite typology is obvious.In Chinese a motion event is typically encoded in a resultative verb compound （RVC） in which two or three constituents represent different semantic aspects of motion （e.g.manner, path, deixis） in a sequential way as demonstrated in example （3） below.

（3）Qiu 2 gun 3-jin 4 leshan 1dong 4.

ball roll-enter ASP cave

‘The ball rolled into the hole.’

In the above example, the first constituent in the RVC conflates the fact of motion and the manner of motion （i.e.gun 3 ‘roll’）, and the second one denotes the path of motion （i.e.jin 4 ‘enter’）.Since grammatical elements in Chinese are not in any way morphologically marked, it is hard to determine which constituent in an RVC is a main verb and which a satellite, thereby giving rise to the debate concerning the exact status of Chinese in motion event typology.Talmy （1991, 2000） treated path constituents（s） in an RVC as a satellite and patterned Chinese with English as being satellite-framed.

However, Slobin （2004） in his data regarding the Frog story^[1] found that it is very difficult to fit Chinese into this two-way classification system.He noticed that a particular motion scene in Chinese could be encoded by a Manner verb + Path verb combination （e.g.fei 1-chu1-lai2 ‘fly-exit-come’） or by path verbs alone （e.g.chu 1-lai2 ‘exit-come’）, with the two options occurring in a similar frequency.Notably, the latter means of expression reflects the defining property of verb-framed languages, that is, the conflation of Motion and Path in the verb.Slobin （2004） therefore concluded that Chinese possessed defining features of both satellite-framed and verb-framed families while not fully patterning with either.He therefore proposed a third type of languages with regard to motion events, which he termed an “equipollently-framed” language, thereby expanding Talmy's dichotomized typology into a trichotomized one.The crucial property of these equipollently-framed languages （e.g.Chinese, Thai and other serial verb languages） is that “both manner and path are expressed...by elements that are equal in formal linguistic terms, and appear to be equal in force or significance” （Slobin 2004: 228）.

As illustrated above,the universality of spatial cognitive concepts and the diversity of means by which these concepts are realized in varied languages make the typological study of languages of space an important topic.Specifically, despite the shared set of semantic components for motion, how do languages systematically vary in the means by which they select and encode these components in different grammatical devices across the utterance? The present study focuses on the Chinese language and compares it to English with the aim of ascertaining the exact status of Chinese in motion event typology: is it more satellite-framed, as classified by Talmy, or more equipollently-framed as proposed by Slobin? To this end, we will investigate and compare the description of voluntary and caused motion events by adult speakers in Chinese and in English.Specifically, we will investigate what semantic components for motion they choose to report, where they tend to place them and how they organize them at the discourse level.

The typological study of motion events not only helps shed fresh light on motion event typology per se but also has important implications for related topics such as the ‘rhetorical style’ of a narrative discourse involving motion scenes and the online conceptualization of a motion event.As regards the former, Slobin （1996b, 2004） made a detailed exploration of literary texts in verb-framed versus satellite-framed languages.He found that speakers of a satellite-framed language tended to use a greater amount and a greater variety of manner verbs than speakers of a verb-framed language in their narrative discourse, and that the former group of speakers tended to elaborate more frequently on path information and expressed more dynamic motion events.By contrast, discourses in a verb-framed language tended to exhibit a discursive narrative style in which speakers either implied location whilst omitting manner or represented a story via descriptions of static physical settings.These findings clearly indicate that the linguistic typology affects how speakers organize narrative discourse.Therefore, if Chinese is an equipollently-framed language, it will be interesting to explore the rhetoric stylethat the discourse of such a language like Chinese can demonstrate.

Regarding the conceptualization of a motion event for linguistic encoding, Slobin （1996a） proposes a hypothesis of “thinking for speaking” which claims that there is a kind of thinking that is intimately tied to language, namely, the thinking carried out online in the process of speaking, writing, signing or listening （and possibly also in online translation and mental imagery）.“Thinking for speaking” involves identifying those characteristics of entities and events that fit some conceptualization of the event and are readily encodable in the language.In this sense, the habitual pattern of using a given language represents a particular way of thinking about the world, and a typological study of languages provides a window through which we can further examine whether motion events are mentally represented for verbal expressions in a typologically specific way.