| お買い物カゴを見る | お買い物ご利用方法 | 教科書をご購入する場合 | ウェブショップ会員 |
| 特長 | 目次 | 仕様 | 関連教材 |
Abstractより
When we begin reading text, we do not have any idea of the text structure or content that we will encounter. As we read sentence by sentence, or paragraph by paragraph, we apply the most appropriate schema to what we have encountered to that point. The schema concept has been referred to by various names: "scenario" (Stanford & Garrod, 1981), "frame" (Minsky,1975), and "script" (Schank & Abelson, 1977). All describe a similar knowledge structure vital to mental activities such as reading comprehension. While some readers apply an appropriate schema and accurately predict what will come next, others do not, having applied an inaccurate schema from what the writer intended. EFL (English as a foreign language) readers may have difficulty applying appropriate schemata because of insufficient linguistic ability.
Although many studies address schema theory in reading comprehension, few empirical studies look at schema change (e.g., Cook, 1994; Schank, 1982). As its departure point in explicating the schema modification process, this study uses the framework of Cook (1994, p.182). Schema disruption happens when textual information does not fit with readers' schemata. As a result, readers have to change their schemata ("schema refreshing"), destroy their old schemata ("schema destroying"), construct new schemata ("schema constructing"), and establish new connections among existing schemata ("schema connecting"). Schema disruption is "a general effect on existing schemata" and "disruption is a pre-requisite of refreshment" (p.191). As a result, readers have to engage in at least one of the following processes: schema refreshing, schema destroying, schema constructing, and schema connecting (establish new connections among existing schemata). Moreover, schema disruption is "a general effect on existing schemata" and "disruption is a pre-requisite of refreshment" (p.191). Readers may also alter existing schemata by "schema reinforcing," "schema preserving" (i.e., confirmation), and "schema adding," which adds new information. Schema refreshing and schema adding are the same concepts as "tuning" and "accretion" in Rumelhart and Norman (1978). The present study will consider the schema modification process relative to all the above elements.
This thesis is concerned with cases of schema failure and recovery. Failure of schema activation in L2 reading may be caused because readers either lack or fail to activate the appropriate schema (Carrell, 1988). The schema modification process is one of the important components of the reading process. The purpose of the present thesis is to investigate whether there is any difference in the schema modification process from two experimental experiments.
Study 1 aims to elucidate Japanese EFL readers' schema modification processes. These processes are examined sentence-by-sentence, and the data are analyzed in the following two stages. First, this paper focuses on the proportion of correct answers (reaching proper schemata) in each passage, which the second focus is when each reader achieved the correct schemata. In order to see the relationship between these scores and factors of L2 reading proficiency, working memory span, and inference ability, a Pearson correlation analysis, a partial correlation analysis and a multiple regression analysis were conducted. Thereafter qualitative analyses are reported.
The results of the regression analyses showed that L2 reading proficiency was a significant predictor, especially for when the readers reached the correct schemata. Furthermore, although poor readers often failed to recover from overly specific schemata, good readers rarely activated specific schemata but tended to keep open their schemata for future modification.
Study 2 investigates Japanese EFL learners' schema activation and modification processes. One hundred and thirty-two university students read six passages composed of five to seven sentences, and responded sentence by sentence with the verbal protocol method. The three main focal points are: (a) the rate at which readers reach the correct schemata, (b) the number of schemata readers simultaneously hold in each sentence, and (c) the schema modification patterns. These perspectives were compared across language groups, L2 reading proficiency and working memory span. In addition to the verbal protocol, the processing times of garden path sentences and schema modification were measured.
The results of the regression analyses showed that readers with Japanese passages held multiple schemata at first, and then narrowed them down to more appropriate ones as they read. Readers with English passages, however, tended to hold only one schema in each sentence. This makes schema changes difficult, and leads to delay or failure in reaching the correct answers. Compared with the difference of language, the effect of readers' L2 reading proficiency and working memory span were not so large. In reviewing qualitative analyses, we can see that there was a tendency for readers to create schema based on the present and only one preceding sentence (i.e., not earlier sentences). This tendency was much more obvious with the English version. In this way, EFL readers often failed to stock previous information and therefore failed to grasp the meaning of the whole story as a result.
As for the processing times, readers with larger working memory span showed longer processing times than those with smaller span. Although there has been a controversy between those who claim that readers with larger working memory span show shorter processing times and those who assert that there is no difference between different working memory span groups for the processing times, there is little research indicating the larger-memory group showed longer processing times than the smaller-memory group. Furthermore, the results of this study showed a different tendency as to where readers spend longer processing times. L1 studies have shown a significant difference between larger and smaller working memory groups at the garden path. However, in this study, the longer processing time was occurred not at the garden path phenomenon but at the end of the sentence, and therefore L2 sentence processing may be relatively tentative.
The results of two experiments are discussed based the theory of working memory and schema modification patterns. What we can conclude from this study is that the way of activation of schemata affects the ease of the schema modification process.
