Chapter 1 Introduction
1.3 Organization
The current study is divided into six chapters. Chapter 1 provides background information on motion encoding of language in children and explains the purpose of this current research. Chapter 2 is a literature review of relevant studies on the discussion of the typology of motion events, the acquisition of motion events, and the difference between children’s expressions in narration and conversation. Chapter 3 elaborates the research methods used in this study. Chapter 4 presents the findings of the study, including the analysis on motion verbs and motion constructions. Chapter 5 is the discussion on the findings. Finally, a conclusion is presented in Chapter 6.
‧
國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Chapter 2 Literature review
2.1 Motion events encoding
2.1.1 Motion event typology
The expressions of motions in various languages have drawn the attention of many researchers for the past few decades (Talmy, 1975, 1985, 1991, 2000; Berman
& Slobin, 1996, 2004). Talmy (1975) was the first to address the issue of the relation between the semantic structures and the lexical patterns in motion events. A motion event is defined as “a situation containing movement or the maintenance of a stationary location alike” (Talmy 1985, p.60). He categorized the semantic elements into several categories, including “Motion”, “Figure”, “Ground”, “Path”, “Manner”, and “Cause”. The lexical elements, or surface elements, include “verb”, “adposition”, and “subordinate clause”. “Motion” refers to its presence per se in motion events. A motion clause consists of at least one linguistic unit describing the Figure changing its location in the time period under consideration, including the Figure moving between two macro-locations, changing posture, disappearing or appearing with respect to the ground, and engaging or disengaging to or from the ground (Talmy, 2000; Guo & Chen, 2009). “Figure” refers to the entity that moves or is located.
Figure can move autonomously or be caused to moved. “Ground” is the referent object or referent point with respect to the Figure’s movement. “Path” is the course that Figure follows when moving or that Figure occupies with respect to the Ground.
‧
國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Path is the core element in a motion event, since there must be a course when the Figure moves in every motion event. The example (1) presents an example with these four basic elements encoded in an English sentence:
(1) The man came into the house.
[Figure] [Motion] [Path] [Ground]
The last two semantic elements “Manner” and “Cause” are considered to be the external semantic elements, since they may not be mentioned in a motion event.
“Manner” is the way in which the figure moves. It’s the co-events of motion. Slobin (2004) mentioned that Manner indicates “motor pattern, rate, rhythm, posture, affect, and evaluative factors” that modulate Motion. Pourcel (2003) mentioned that Manner also includes force dynamics, attitude, and encoding instrument that concerns Motion. “Cause” is the external force that moves the Figure. The example is shown in (2) (Talmy, 1985):
(2) The pencil rolled off the table.
[Manner]
The pencil blew off the table.
[Cause]
Talmy compared the lexicalization structures across different languages, and found that Motion, as a universal phenomenon, could be expressed in different ways
‧
國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
in typologically different languages. The typological characteristics occurred frequently and pervasively in colloquial style. Different language families tended to conflate Motion with different semantic elements and different surface element as well. The most distinct character was that “Path”, the core semantic element, was encoded in syntactically different fashions (Talmy, 1985, 1991). In some languages, Path was encoded in the main verb, occupying the main position of a sentence. Such languages included Spanish, Turkish, and Hebrew. These languages were categorized as verb-framed languages (V-languages). In other languages, Path was expressed by a subordinate component as a satellite element associated with the verb (e.g.
prepositions or particles). These kinds of languages included English, Russian, and German. They were called satellite-framed languages (S-languages).
For example, when describing the motion event of a balloon (Figure) moving upward (Path) through the chimney (Ground) by floating (Manner), a Spanish (V-language) speaker normally makes a sentence as (3) (Talmy, 1985):
(3) El globo subió por la chimenea (flotando).
the balloon rose through the chimney (floating)
In Spanish, Path (moving upward) is encoded in the main verb subir (rise). The verb subir clearly indicates the meaning of moving upward. Manner, on the other hand, is expressed by a dispensable particle flotando (floating). As for English, an S-language, the same motion event is often encoded as:
‧
國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
(4) The balloon floated up through the chimney.
Different from Spanish, Path in English is encoded in the particle “up”, and the main verb “float” bears the meaning of Manner.
The typological tendencies of motion expressions have aroused the interests of many other researchers. Slobin (1996, 2000, 2004) believed that motion event typology influenced the habitual language use of speakers of different languages. He emphasized the importance of “the typologies of language use” (Slobin, 2004). That is, linguistic patterns occurred in dynamic and interactive situations when language was used. Therefore, it was necessary to study actual language use in a diversity of languages and with different types of data. Slobin (2004) examined children and adults’ narratives elicited from the wordless picture book Frog, where are you?
(Mayer, 1969) in different languages, and found several differences in their rhetorical style. First, the usage of Manner was different in type and frequency. S-language speakers tended to use more diverse lexicon of Manner than V-language speakers. S-language speakers also mentioned Manner more frequently than V-S-language speakers. For example, in Frog story, there was a scene where an owl emerges from the tree. English (S-language) speakers used Manner verb to describe the scene, such as: An owl popped out (Slobin 2004, p.224). Spanish speakers use Path verb describing the same scene, such as: Sale un buho “Exits an owl”. Second, different languages differed in the packaging of Path components. S-language speakers seemed to encode more Path segments per sentence by using multiple Path satellites prepositions, e.g. The frog crawled out of the jar and through the window into the
‧
國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
woods. (Slobin, 2004). On the other hand, V-language speakers encoded Path by
using one bare verb per clause, such sentences translated in English would be like:
The frog exiled the jar, passed through the window, and entered the woods (Slobin,
2004). Third, the description of Ground was different. S-language speakers provided more elaborated Ground information, and provided such information more frequently than V-language speakers. Finally, Slobin (2004) suggested that the motion expressions could be affected by other factors beyond the typological one. Those possible factors included patterns of motion expression which arise in language use, in the culture, and in the communicative aims of the speakers. Slobin believed these factors above all contributed in shaping the motion encoding among different languages but he had not mentioned further evidence supporting how communicative aims could affect motion event expressions.
2.1.2 The acquisition of motion events encoding
Along with the discussion of motion event typology, the interest in how children come to learn the expressions of motion event appeared. Researchers have investigated the development of children’s motion expressions in various languages.
The issues frequently addressed in previous studies were whether children’s motion expressions had language-specific patterns from the beginning of children’s language acquisition, and how the language-specific patterns affect children’s motion expressions as they grow older. The different language-specific patterns of motion encoding found in previous studies include the typological differences suggested by
‧
國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Talmy’s, the differences of the rhetorical style suggested by Slobin (1996), or other more specific usages of motion verbs and prepositions different from language from language. Such studies often compare two or more typologically different languages.
Choi and Bowerman (1991) studied how Korean and English children lexicalized motion events with the aim of examining the language-specific factor played an important role in children’s language expressions from early on. They discovered that children are sensitive to the language-specific patterns since the age of two. For example, English children could master the particle use of Path such as up and down in both spontaneous and cause motion, and Korean children are able to distinguish verbs used for spontaneous and cause motion. Özçaliskan and Slobin (1999) studied English, Spanish and Turkish children’s motion verbal expressions. They found that from the age of three, children talked about motions in distinct ways that are specific to their native language. English (S-language) children preferred to use more Path verbs, while Spanish and Turkish (V-language) children had the preferences of using more Manner verbs. Allen et al. (2003) examined English and Turkish children’s motion expressions, and discovered that children from the age of three encoded motion with language-specific patterns. English-speaking children did not use two separate verbs to encode Manner and Path, while Turkish children did. Hickmann and Hendriks (2010) compared English and French children’s motion expressions. It was found that the English children utterances showed higher semantic density than French children. Also, French-speaking children express both Manner and Path with verbs, while English-speaking children usually express only Path with verbs. As
‧
國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
these previous studies suggest, children learning different languages encoded motion events in a language-specific way from early on, around the age of two to three.
The studies of children’s motion expressions also concern the issue of Slobin’s (1991, 1996b) “thinking for speaking” hypothesis. Slobin proposed a dynamic approach to deal with the relation between thought and language. “Thinking for speaking” was a special form of thought. It was an on-line mental activity, which took place while one formulated utterances. When we spoke, the grammar of the language guided us the speaker to take a language-specific perspective of what we experienced. The grammatical structure of language provided “a set of options for schematizing experience for the purpose of verbal expression” (Slobin,1996b). In other words, the patterns of our utterances were the realization of the thinking process. It was the communication purpose that characterizes the nature of thinking for speaking. Because of the need to communicate, the special form of thought was accessed and mobilized for language use.
The evidence of thinking for speaking is found in several areas, such as the language use of different genres, the use of gestures when speaking, and language acquisition (Slobin, 2000; Kita and Özyürek, 2003; Stam, 2006). Slobin (1996b) pointed out the issue of learning to think for speaking. Since thinking for speaking required the speakers to pick “those characteristics of objects and events that (a) fit some conceptualization of the event, and (b) are readily encodable in the language”
(1996b, p.76), he hypothesized that children needed to learn to think in accordance with how their native language allowed them to speak. Children were guided by the semantic distinctions and preferred constructions of their native language to attend to
‧
國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
the specific features of events when they spoke. If true, the typological differences of motion encoding could affect children’s “thinking for speaking” as well.
Slobin (1991, 1996) examined thinking for speaking theory through the study of the motion expressions of children learning typologically different languages. He found systematic contrasts between English and Spanish children’s utterance. The contrasts reflected that children had different thinking for speaking patterns in respect to their mother tongue. Children attended to the particular aspect they needed to focus on in order to encode their experience of the world into language.
Furthermore, thinking for speaking functioned in various genres. Slobin (2000) tested the theory in pervasive genres. In his study, the expressions of motion events were examined in narrative, fiction, and conversation. The result showed that the motion encoding of different languages remained their own typological patterns regardless of the contexts. For example, S-language speakers always used more Manner token than Path token while V-language speakers always used more Path token than Manner token in all genres. In this present study, accordingly, it is assumed that children’s motion expressions should present the same typological character in different genres.
2.2 Motion events encoding of Mandarin Chinese
2.2.1 Motion encoding typology of Mandarin
The typology classification of Mandarin has been a dispute over the years.
Talmy (1991) classified Mandarin as an S-language. S-languages is said to encode
‧
國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Path in a satellite position in a sentence. He argued that Path verbs did not usually function as full verbs and that Path verb type was rather limited. Therefore, he believed that, in Mandatrin, the second verb (usually a Path component) of the motion serial verb was generally viewed as the satellite part of the serial verb, so Manner was the more significant component. Such argument was supported by scholars such as Chao (1968) and Li and Thompson (1989), who also regarded the second verb “Path” as a complement to the first verb “Manner”. However, later Tai (2003) proposed that the second verb (Path) was the center of predication, or the head of the serial verb construction. He argued that because the second verb referred to the foreground information, it should be considered more important than the first verb (Manner) of the serial verb. This means that Mandarin should be considered a V-language, since V-languages encode Motion mainly in Path verbs. The argument of which segment takes the primary position has long been discussed until Slobin (2004) proposed an alternate view on Mandarin’s categorization. He believed that Mandarin belonged to the third type — equipollently-framed languages. He believed that V1 (Manner) and V2 (Path) were equally important in its morphosyntactic status, because it was very natural for the Mandarin speakers to express Motion with either Path or Manner verbs, and they appeared “to be equal in force or significance” (Slobin, 2004). Chen (2005) and Guo and Chen (2009) supported this view by examining Mandarin speakers’ use of motion verb type and the encoding of Ground information. They found that the rhetorical style behaved not so much like V-languages nor S-V-languages. Therefore, they believed that Mandarin indeed belonged to another type — the equipollently-framed language.
‧
國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
In Mandarin, Path, Manner and Deixis are all allowed to be coded individually or jointly in a serial verb (such as: ⾶飛上 fei shan, “fly up”; 跑出來 pau guo lai, “run out over”). Path verb is coded simultaneous with the semantic element of Motion and Path. Furthermore, Path can appear alone (e.g. 進 jing “enter”), or be combined with other Path, Manner or Deictic verbs. When it appears in a serial verb, it may occupy the first slot of the verb (e.g. 進來 jing lai “come in”). But most of the time, it occupies the second slot of the serial verb (e.g. 跑進 pao jing “run in”). Manner verb, naturally, is encoded with the semantic element of Manner and Motion. Manner verb is one of the principle linguistic devices for expressing Manner of motion events in Mandarin (Chen, 2005). Like Path verbs, Manner verbs can be used individually (e.g.
跑 pao “run”). When it is used jointly with Path or/and Deixis, it always stays in the
first slot of the verb (e.g. 跑掉 pao diao “run away”; 跑出來 pao chu lai “run out”).
Deixis verbs indicate whether the Figure moves away or toward the speaker. There are two kinds of Deixis in Mandarin: 來 lai “come” and 去 qu “go”. Deixis can either be used alone, or combined with Path or Manner in the final slot of the serial verb (e.g. 跑來 pao lai “run near”; 跑出來 pao chu lai “run out toward”). One of the language-specific characteristics of Mandarin was that Manner, Path and Deixis verbs could occur simultaneously in a serial order to form serial verb (Chen, 2005).
Naturally, the three categories of the motion verbs can generate several combinations of different constructions. However, among all the constructions, Manner + Path
‧
國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
(+Deixis) was the most common construction found in Mandarin speakers’ motion encoding (Chen, 2005; Guo & Chen, 2009).
Apart from encoding simultaneously Manner and Path (including Deixis) into a construction of serial verb, there are some other language-specific features of Mandarin in coding motion events mentioned by scholars such as Chen (2005). Chen discussed adults’ motion encoding in the aspects of motion verb token, motion verb type, Ground description, motion segment across sentences, and static settings versus dynamic movement. According to his study, in verb type usage, Mandarin speakers used a lot more types of Manner verbs (including transitive and intransitive ones) than Path verbs (45 vs. 16). As for verb token usage, Manner frequency is a lot higher than that of only Path (65% vs. 26%). As for Ground information, 52% of their motion expressions were found with at least one type of ground information.
Furthermore, the Mandarin speaking adults used 3.5 event segments on average in the deer scene. Last, the speakers tended to describe more physical settings than dynamic movement.
In sum, there are several features in Mandarin’s motion encoding: 1) Construction: Manner+Path (+Deixis) can all be coded into a serial verb, and are the most common motion construction. 2) Verb type: Manner lexicon was larger than Path lexicon. 3) Verb token: Manner usage was more frequent than Path. 4) Ground:
Ground information occurs in high frequency, about half of the time. 5) Multiple-action clauses are used in a complex event. 6) Descriptions of physical settings are frequent. These characteristics above are not entirely applicable to either S-language or V-language, thus Mandarin appears to be a unique language type in motion events
‧
國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
encoding, and remains to be studied more. This present study focused on the analysis of the verbs, thus 1), 2) and 3) were used altogether as the criteria to examine the language-specific patterns of the motion verbs.
2.2.2 The acquisition of Mandarin speaking children’s motion encoding
Several scholars have carried out much research on Mandarin children’s development of motion expressions (Chen, 2005; Guo & Chen, 2009; Lin, 2006;
Huang, 2012). These researchers examined many aspects of the motion expressions of children, including verb type, verb frequency, verb construction, and syntactic construction. However, the results of different genres of data seem to contradict one another. Most of the studies investigated narrative data (Chen, 2005; Guo & Chen, 2009; Lin, 2006). Chen (2005) elicited children’s narration by using the picture book Frog, where are you (Mayer, 1969), and found that children used more types of
Manner verbs than Path verbs starting from the age of three, and that the frequency of Manner was higher than that of Path in all age groups. The frequency of Ground information was lower than that of the adults, but would increase with age. Guo and Chen (2009) used the same method to collect their data, and found similar results, which is that Manner prevailed Path with respect to verb type. They have also investigated children’s motion constructions, and found the frequency order resembled to that of adults, and that M+P(+D) was the most productive construction from the age of three. Lin (2006) used animated films to collect Taiwanese children’s narrative data. The results of verb type and token concurred with that of Chen (2005) and Guo and Chen (2009). Motion construction is similar to the results of Guo and
‧
國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Chen, presenting M+P+D to be the most productive construction. These three studies above suggested that children tended to use Manner more frequently and more diversely than Path, and the frequency and diversity of Manner increased with age.
Also, the construction of M+P+D occurred with high frequency in children’s speech.
This construction is a typical characteristic of the equipollentlly-framed language,
This construction is a typical characteristic of the equipollentlly-framed language,