Talmy lexicalization pdf
The data were analyzed both quantitatively and qualitatively, and compared with the English corporal data also obtained for this study. The result will then be compared to the suggested typology of S- and V-languages. In the section that immediately follows, the suggested typology of S- and V-languages will first be presented. The design of the study, as well as the results and discussion will follow. Finally, implications for SLA and limitations of the present study will be discussed.
An event where an owl popped out of the hole translational , for example, can be often captured as a change of state the state changed from invisible to visible as in an owl appeared from a hole in the trunk in Romance languages Slobin, Talmy defines six basic components of motion events: 1 the presence or absence of the translational motion Motion , 2 the moving entity Figure , 3 the object with respect to which the Figure moves Ground , 4 the course followed by the Figure with respect to the Ground Path , 5 the manner in which the motion takes place Manner and 6 the cause of its occurrence Cause b, p.
Based on the patterns of how the semantic components are mapped onto the surface forms, Talmy b proposes two main cross-linguistic typologies: Satellite-framed languages S-languages , and Verb-framed languages V-languages. Scene settings and Ground are often left to inference. In contrast, V-languages, which include Romance, East-Asian languages, tend to conflate Motion and Path into verbs, and Manner is often left to inference or expressed in adverbial phrases when mentioned Cadierno, a.
V- languages also prefer to capture motion events as change of state, such as being in to out or invisible to visible. Cadierno a, p. English S-language The man ran out of the house. The man went out of the house running. In an English narration, Path can be elaborated on with additional PPs, such as into the backyard, while Spanish needs verbs to elaborate Path Cadierno, Consequently, S-languages tend to have detailed sub-trajectories of a motion event by having many Path PPs associated to one verb.
This way, many sub-trajectories are efficiently packed into a motion event, whereas V-languages need multiple verbs to describe different sub-trajectories of a motion. See the example below: 2 a. English S-language The man ran out of the house into the backyard.
The man went out of the house and entered into the backyard running. In capturing the same motion event, both 2a and 2b divide the motion event i. Spanish 2b , however, has to utilize three verbs go, enter, run to describe the same event while English needs only one. Slobin asserts that having multiple verbs to describe sub-trajectories of a movement is costly in terms of processing, and thus V-languages do not prefer segmentation—dividing a motion into multiple sub-trajectories.
Therefore, the trajectory of motion is usually left to inference in V-languages. S-languages, by contrast, prefer to have more segments sub-trajectories than V-languages, because the grammatical structure allows Path to be easily elaborated with satellites and PPs Slobin, In the English example above, the multiple sub-trajectories out of the house, into the backyard are united with one motion verb run , while in the Spanish example, each sub-trajectory is coordinated by a null subject e.
So far, only the signifying packaging in S-languages—that is, having multiple sub-trajectories tied to one verb—has been given attention in SLA research. In contrast, packaging in V-languages, is loose because predicates motion verbs in this case cannot have multiple Path components. Here, Slobin and Berman define a unit that is unified with a predicate as a clause.
It thus can be predicted that S- languages will tend to have longer, fewer clauses while V-languages may have shorter, but more clauses in a text.
However, as Talmy himself asserts, these lexicalization patterns are general tendencies rather than absolute cross-linguistic differences. The man ran into the house. The man entered the house running. Both sentences in 3 are grammatically possible constructions in English.
However, 3a , in which a verb conflates Motion and Manner, is preferred over 3b , which conflates Path and Motion into the main verb with the Manner encoded in a gerund. Slobin suggests 3a is preferred as it can be processed more easily than 3b , because 3a economically encodes multiple semantic components Motion, Manner[run], and Path[inward] , with fewer words and simpler structure. Compare the Spanish sentences below Cadierno, a, p. The man ran in the house.
Another problem with current lexicalization pattern studies is that their focus is primarily on the S-V dichotomous contrast, ignoring the fact that some languages do not fall into either category. Slobin defines Chinese as an equipollently-framed language E-language , where Path, Manner and Motion are encoded in equipollent elements.
Nage nanren paoxiang fangzi qule The man ran toward the house. Nage nanren paojin fangzi qule The man ran into the house. None of them are grammatically marked as subordinating phrases like satellites.
The elaboration of Paths modifying one manner verb pao run in these sentences explains why Chinese was originally classified as an S-language Talmy, , even though these Path elements are not satellites.
Overall, studies on Chinese lexicalization patterns e. As such, languages should be placed alongside a continuum of Manner saliency—that is, how accessible and codable Manner of motion is may depend on the particular language, rather than being classified into dichotomous or a ternary typology Slobin, The suggested characteristics of the language types discussed are summarized in Table 1.
Overall, S-languages elaborate more on Path and Manner, having multiple Path segments packaged into one clause and often encoding Manner into verbs. V-languages, on the other hand, leave Manner and details of Path to inference.
Elaboration of Path and Manner is costly for V- languages, because they would need verbs in each description of sub-trajectories. V-languages tend to capture motion events as change of state, thus providing elaborated Ground information instead. In regards to the degree of Manner and Path saliency, E-languages seem to fall in- between S- and V-languages. The current literature then would give us the impression that Japanese V-languages will have to spare more clauses and sentences to describe the same amount of information of motion events, or otherwise, would deliver less information of Motion than English S-language.
This study, hopefully, will reveal whether or not Japanese in fact pays less attention to Motion, and how influence from the Chinese language interacts with the way Japanese narrates motion events. Unlike previous studies on Japanese, the present study will exclusively focus on the lexicalization patterns of motion events. More specifically, the study is interested in: 1 whether or not English and Japanese focus on Motion at all when reporting events; 2 if so, how each language does it at the discourse level, and what lexicalization patterns are used to encode Motion.
Excerpts from newspaper articles reporting on the aircraft movement were extracted and analyzed at the sentential and discourse levels. Results from the Japanese articles were compared to the English ones, as well as to the suggested lexicalization patterns of V-languages see Table 1. The four sets were 1 English articles reporting the American accident as domestic news, 2 Japanese articles reporting the American accident as international news, 3 Japanese articles reporting the Japanese accident as domestic news, and 4 English articles reporting the Japanese accident as international news.
All four sets were used to identify the cross-linguistic differences in lexicalization patterns rather than difference in wording due to translation or due to different degree of interest in domestic versus. The two airplane accidents were both national news events.
One occurred in the US, in which a passenger flight from New York to Charlotte, North Carolina, successfully ditched in the Hudson River after striking a flock of geese on January 15, henceforth Hudson. The two accidents were chosen because the movements of the aircrafts are somewhat important, though not obligatory, in reporting. This allows an analysis on how much of the attention is paid to Motion in English and Japanese narratives, as well as the lexicalization patterns themselves.
All articles were located online and copied and pasted to word document files for the analysis. Overall numbers of English articles are smaller 20 each for Hudson and JAL than Japanese articles 55[Hudson] and 33[JAL] but the size of the four data sets are similar pages, lines. Ratios, rather than the raw numbers of clauses, that describe motion events were utilized to make the results comparative between English and Japanese.
Coding and Analysis In order to identify the patterns in narrating motion events at the sentential and discourse levels, the following was taken as the foci of the analyses: 1. The number and types of sub-topics reported — in order to identify what aspects of the incident were reported; 2.
The number of clauses divided by the number of sentences roughly would reveal how loosely the motion events are packaged. The sentences would contain more clauses if the events are packaged loosely. However, just because the sentences have fewer clauses does not necessarily mean that the events are tightly packaged — for the fact that one clause may only have one predicate, and that it does not mean the clause packages multiple path elements in it.
The more semantic components are encoded into segments, the tighter the packaging is. In the corpora, the author identified 34 sub-topics for Hudson, and 32 sub-topics for JAL e.
Then, the entire corpora were color-coded into sub-topics. Next, the texts were divided into clauses by the author. The inter-coder reliability was Also, the intra-coder reliability was Then, motion verbs and other motion segments that were tied to the verbs e. Then the types and tokens of the verbs were counted. Finally, the semantic components encoded into the motion verbs and segments that describe motion motion segments were identified and tabulated by the author.
The differences between completely vs. When this type of specific identification was not possible e. The percentages of encoded semantic components were calculated for the analysis. In terms of the focus of the articles, the English ones included more varieties of sub- topics per article This means that English tended to focus on different aspects of the incident in one long text whereas each of Japanese news articles tended to focus on fewer aspects of the accidents in a short text. Moreover, the English articles used a slightly lower percentage of clauses than the Japanese ones in describing motion events e.
Token frequency of semantic components The token frequency and percentage of each semantic component encoded into verbs and other motion segments are presented in Table 2. The rows represent the grammatical elements and the types of encoded semantic components.
Because clauses may consist of multiple segments, and verbs and segments conflate multiple meanings, the total number of semantic components does not add up to the total number of clauses. This is understandable, given the fact that both English and Japanese complete sentences require verbs, but not other motion segments, such as adverbial phrases or PPs.
Also, Japanese verbs instead encoded Figure Contrary to the suggested typology, Japanese verbs encoded Manner Overall, the results confirm the suggested typology of S- and V- languages to a certain extent: English S-language elaborates on Path with motion segments and Manner and Path with verbs, while Japanese verbs encode Path.
However, the interesting finding here is how similar Japanese and English are in elaborating on Manner and Path via verbs. As a result, when Manner has decreased salience, English may make use of V-framed grammar.
Finally, what is also noteworthy is that Japanese does elaborate on Manner with motion segments, and Figure and Ground with verbs. Verb types Overall, the English data contained 79 verbs, whereas the Japanese contained Among the verbs that appeared in the data, eight verb-types were identified, based on the encoded semantic components.
The numbers and examples of each verb type are presented in Table 3. The columns represent the languages, and the rows represent the verb types. Verbs listed under one category in Table 3 may have been coded as multiple verb types for the analysis due to the semantic conflation e.
Examples No. Moreover, Japanese data contained more varieties of Figure and Ground verbs. In addition to the eight semantic verb types, two formal verb types were observed among Japanese verbs: Chinese-loan-verbs CLs and native-Japanese-verbs NJs See Table 3 for their distribution.
CLs consist of a Chinese phrase that was originally loaned as a noun in Japanese, and turned into a verb via the Japanese verb suru do. Since the Chinese loanwords are already compounds, most CLs conflate multiple meanings.
These CLs, moreover, tend to convey Figure as well. Ririku-suru, for example, only refers to departures of flying transportation, and cannot express take-offs of birds, for example. NJs, in contrast, tend to encode one semantic component.
Out of the 36 7 manner and 33 path verbs NJs in the data, only four conflated different semantic components, and five conflated two Paths. And these semantic conflations were done via compound verbs, such as tobikoeru, each consisting of two verbs tobu jump and koeru go over , thus conflating Manner and Path.
Consequently, since path NJs can encode only one sub-trajectory e. For example, mukau head , sagaru descend , and magaru turn almost always co-occurred with one Goal NP.
The NPs that serve as scene setting seem to semantically weigh more than NJs in narrating on motion events in Japanese. The ground NP in 6 , consisting of a clause and a subordinating relative clause, illustrates the relative locations between the Hudson River, the crash-site, George Washington Bridge and the height of the flight, as if unfolding a map for readers. Fuji etc … The NPs in 7 describe Motions statically, as if showing a snapshot of the actions returning-route, or sways and turns , and the NPs, after all, set the scene of the motion events.
The verbs that come after the NPs describe what people did given the settings. As discussed above, CLs consist of a noun, which mostly captures course of actions statically by providing Path, and a verb suru do that describes the execution of the action. Therefore, the noun e. Because CLs elaborate Path, they co-occurred with adjectives, adverbs or extra nouns that provide Manner information as shown in 8 : 8 a.
The plane flies unsteadily. The plane flew in an S-shape. The plane was flying as planned. Most Japanese sentences or clauses tended to elaborate Path or Ground, with NPs preceding verbs as shown in 6 and 7.
Thus nouns, not verbs, packaged segments in Japanese. By contrast, English manner verbs were more semantically loaded, unifying different segments of motion events. The trajectories and spatial information encoded in motion verbs eventually infer Ground.
Thus, English verbs are required to encode richer semantic elements than the following segments so that the Ground can be inferred. It is interesting to note that Manner CLs 10e show a similar lexicalization pattern. As the six patterns in 10 show, Path is almost always encoded in Japanese motion phrases, and Manner is optionally, though frequently, encoded via various grammatical elements onomatopoeia, adverbs, adjective, nouns and gerunds.
In terms of boundary-crossing motion events, Japanese Manner verbs never expressed boundary-crossing events and the Motions expressed via verbs took place within the bounded area which is delimited by the sentence initial NPs.
Regarding the first question, Japanese encoded as many semantic components on Motion as English did, suggesting that English and Japanese focus on Motion roughly equally when reporting events. This finding contrasts to what is suggested by the dichotomous typology, which predicts that Japanese, a V-language, would leave more information to inference than English would.
In terms of the second question, this study confirmed that, at the sentential level, Japanese lexicalization patterns follow partially the suggested characteristics of V-languages, encoding Path into verbs with elaboration of Ground. Further qualitative analysis of Japanese lexicalization patterns revealed intricate relationships between the syntactic characteristics of Japanese and its lexicalization patterns at the sentential and discourse level, which in turn suggests a need for a more fine-grained account of the morphosyntax of motion events beyond the typological approach.
At the first layer of analysis, as Talmy predicts, Japanese V-language needed more predicates clauses than English S-language to describe the same aircraft movements, meaning English indeed tightly packages Path segments into one clause, and Japanese does not.
The fact that a relatively high percentage of Japanese motion verbs encoded Manner may indicate that some clauses are solely dedicated to express Manner, modifying path verbs. The verb-type analysis also confirmed that S-languages have more manner verbs than V-languages, and V- languages prefer to encode Path in verbs.
However, the present data suggest that Japanese V-language indeed elaborates on Path via abundant Path verbs that conflate other semantic components like Figure and Ground, while the typology predicts that V-languages leave Path to inference.
The conflation of Figure in verbs may be compensating the fact that Japanese prefers to drop sentence subjects whenever they are recoverable from the context. The fact that English did not conflate Figure in verbs may be because English never drops sentence subjects. This relationship between the meaning conflation at the lexical level and the sentence structure illustrates how information organization at the local lexical and global sentential or discourse level interact in intricate ways.
In addition, the qualitative analysis of Japanese verb-types also provided evidence of the interaction between local and global information organization patterns: Japanese captures motion events statically because noun phrases, which are static in nature, appear at the sentence-initial position, while English captures motion events dynamically because verbs appear sentence initially.
Consequently, Japanese elaborates on scene setting, which in turn enhances the clarity of Path encoded in Japanese verbs. English, on the other hand, provides detailed information of Motion via verbs, often conflating Manner. The semantic-pragmatic consequences of this ordering are clear: by the time the verb is produced … the complements have already provided all the necessary information about the motion event.
It may also explain why Chinese-loaned-verbs have similar lexicalization patterns to English, because Chinese often takes verbs at the second position in the word order. Additionally, Japanese elaborated on Manner as much as English, but via different means. First, we assume we can isolate elements scparately within the domain of meaning and within the domain of surface expression.
Second, we examine… Expand. Save to Library Save. Create Alert Alert. Share This Paper. Background Citations. Methods Citations. Results Citations. Figures, Tables, and Topics from this paper.
Biological anthropology Programming Languages medullipin I Procedural programming. Citation Type. Has PDF. Publication Type. More Filters. One of the most pervasive problems in machine translation is that of lexical ambiguity.
In this paper we suggest to apply a Frame Semantic approach to the construction of the Danish monolingual … Expand.
View 1 excerpt, cites background. This list contains about a hundred verbs like passer … Expand.
0コメント