Chapter One Introduction
Background of the Study
Vocabulary is one of the essential elements of language. To emphasize theimportance of vocabulary, David Wilkins (1972，p.93) claimed "that withoutvocabulary nothing can be conveyed". Norm Chomsky shared the same emphasis onlexis in his Minimalist Program for Linguistic Theory. He stated "that there is onlyone human language apart from lexicon，and language acquisition is in essence amatter of determining lexical idiosyncrasies." (Chomsky, 1995，p. 147) When it comesto applied linguistics，a great number of researchers also highlight the importance ofvocabulary，for example，Gass and Selinker aigued that the lexicon might be the mostimportant language component for learners (Gass & Selinker, 2008). Thus，ever sinceSL acquisition became an issue in SL research, practitioners and researchers havenever moved their eyes away from vocabulary acquisition and vocabulary teaching，and have looked for practical and effective methods to help with the acquisition orlearning of SL vocabulary. In the reservoir of countless experimental studies about SLvocabulary learning, there are two main categories: intentional SL vocabularylearning and incidental SL vocabulary learning. (Laufer，2001) The concept ofincidental vocabulary learning was first considered as a mode of first languagevocabulary learning by Nagy, Herman and Anderson in 1985，which was later adoptedby Krashen (1989) to describe the same kind of phenomenon in SL vocabularyacquisition. IVL refers to "the learning of vocabulary as a by-product of any activity not explicitly geared to vocabulary learning." (Hulstijn, 2001, p. 265)
Many empirical studies have been done to unveil the features of SL IVL.There were empirical studies that investigate the relationship between text type andincidental SL vocabulary acquisition (Herman, 1987), the effect of appearingfrequency on SL IVL (Webb, 2007)，the effect of form-enhancement on it (Paribkht &Wesche, 1997)，the different effect of task types on it (Griffin & Harley, 1996; Laufer,2005), and the relationship between learners' strategies and their efficiency in SL IVL(Potter，Eckardt & Feldman，1984; Tekmen & Daloglu，2006)，and so on. However,one of the most important factors, learners' mental condition，have not been givenenough examination and attention as it deserves. Until now, only quite a fewexperiments have been conducted to study the effect of learners' mental condition ontheir SL IVL, such as Pulido (2003) and Leeser (2007). The key reason that causedsuch a vacancy in study might be that it is comparatively more difficult to studymental activity during IVL in empirical researches than to design experiments withsome non-leamer based and non-mental controlled variables. Still, to the author'sknowledge, there is rarely any empirical study about the relation between ChineseEnglish learners' mental condition and their SL IVL. Thus, there is an urgent need toconduct experimental studies in this aspect, which is right the gist that the presentstudy devotes to.
.....................................
Purpose of the Study
Researching into theoretical and empirical studies on the effect of SL learners'mental condition on their IVL, the researcher found that "background knowledge"(Jacobs, Dufon & Fong.，1994; Leeser, 2007; Pulido, 2003) is one typical factoramong all kinds of factors affecting learners' mental condition. Therefore, the purposeof the present study was to investigate the effects that different levels of background knowledge had on SL IVL, In another word，the present study aimed to settle such aquestion as whether reading materials with more familiar scenarios would foster morechance for IVL and materials with less familiar scenarios would obstruct it. Byachieving such a purpose, the present study would be able to give hints for the futuredesigns of Chinese college students' reading materials, and be able to provide theoriesrelated with empirical evidences.
Significance of the Study
The present study was designed to give thought to gain theoretical andpedagogical significance.
Theoretical Significance
This study took Schmidt's (2010) Basic Assumptions of Attention and Craik &Lockhart's (1972) Levels of Depth of processing theory as the theoretical basis，thusthe statistics and results were devoted to test the two theories' validity in SL IVL.According to the Basic Assumptions of Attention, the attention recourse is limited,and learning needs the allocation of the attention recourse. Thus, readers' allocation ofattention recourse is different when they are reading texts with different levels offamiliarity. That is to say，when we are reading texts with more familiar scenarios,more attention recourses would be allocated to the new word encountered，but we areless likely to do so when reading texts with less familiar scenarios. In addition,according to the theory, allocation of attention would influence the take-ins inworking memory or short-term memory. Thus, a better SL IVL in short-term memorywould give support to Basic Assumptions of Attention; while a worse IVL inshort-term memory would deny the validity of the theory in short-term SL IVL.According to the Levels of Depth of processing theory, only can the newlyencountered knowledge processed in the deep levels of our cognitive mechanisms be stored in long-term memory and become a truly acquired knowledge. In the presentstudy，new words in the less familiar scenario texts are likely to be processed deeperthan that appearing in the more familiar scenario texts. Thus, the present study wouldtest the validity of the Levels of Depth theory in long-term SLIVL.
.....................................
Chapter Two Theoretical Foundation
This chapter elucidated the theoretical foundation which guided the presentstudy and the hypotheses for the research questions. In order to provide the necessarytheoretical background knowledge for the readers, here, in this chapter the authordemonstrated the terminologies, definitions，theories and hypotheses that involved.The elucidations began with issues about IVL; then it explained the theories andhypotheses, including Chunking theory, Comprehension Hypothesis, MentalSchemata and Representation hypotheses, Basic Assumptions of Attention, NoticingHypothesis, and Levels of Depth of processing Hypothesis.
Incidental Vocabulary Learning
TheDefmition oflVL
In Longman Dictionary of Language Teaching and Applied Linguistics, thedefinition for incidental learning is "learning something without the intention to learnit or learning one thing while intending to leam another, for example，unintentionallypicking up vocabulary, patterns, or spelling through interaction, communicativeactivities, or reading for content or pleasure. In controlled experiments, incidentallearning is usually used in a more restricted sense, operationalized as a condition inwhich subjects are not told in advance that they will be tested after treatment,sometimes contrasted with an intentional condition in which subjects are told whatthey will be tested on." (Richards & Schmidt, 2002, p. 252) Nagy, Herman &Anderson，s result in their experiment in 1985 showed that a substantial proportion of Children's vocabulary growth happened probably through incidental learning.However, when we look at the researches on IVA (Laufer & Hulstijn，2001; Joe，1995，1998; Newton, 1995; Pulido, 2003，2004; Wesche & Paribkht，1996), differentinterpretations of IVL，s definition were formulated. For example, Wesche & Paribkht(1996，p. 36) defined incidental learning of vocabulary as what happened whenlearners were "focusing on understanding meaning rather than on the explicit goal oflearning new words.” Thus, it was necessary for the present study to base strictly onthe definition given by Longman Dictionary of Language Teaching and AppliedLinguistics to constitute the definition for IVL. The definition here was learningvocabulary without the intention to leam it, through interaction, communicationactivities or reading for content or pleasure. Moreover, in the present experiment, thecontrolled condition for IVA was reading for content.
Intentional and Incidental Vocabulary Learning
As mentioned above, incidental learning only happens when one isunintended to leam. Thus it is quite necessary to explicit the difference betweenintentional vocabulary learning and IVL.
Longman Dictionary of Language Teaching and Applied Linguistics givesthe definition for the intentional learning as: intentional learning is the "learning byfollowing a deliberate programme of study to enhance vocabulary or grammar"(Richards & Schmidt, 2002，p. 263). Hatch & Brown (1995)，Laufer (2001)，andMondria & Wit (1991) all gave the same kind of definition to intentional vocabulaiylearning that intentional vocabulary learning is the intended and deliberate learningof vocabulary，for example，doing vocabulary exercises (Hulstijn，2001). All thisdescriptions of the features of intentional vocabulary learning emphasized that toaccomplish intentional vocabulary-learning goal，the learners should deliberately engage in a task with leaning the vocabulary as at least one of its learning goals. Thelearning environment is that the learners are informed of receiving some kind of teston their learning about the target words after the learning activity. (Hulstijn，2005)
....................................
Cognitive Linguistic Theories of L2 Vocabulaiy Learning
Since the present study was theoretically rooted in cognitive linguistictheories, in this section，it is reasonable for the author to elaborate the related theoriesin this field here. Theories like Chunking，Comprehension Mental Processing, mentalschemata, representation, Basic Assumptions of Attention, Noticing hypothesis andthe Levels of Processing Hypothesis are introduced here with details.
Chunking as LAP
Chunking is one of the most important terms in cormectionist view aboutlanguage learning and acquisition. Miller (1956，p. 82) first coined the term Chunkingin his review on Short-term memory，and defined it as "the development of permanent'sets of associative connections in long-term storage, which is the process thatunderlies the attainment of automaticity and fluency in language." Connectionistholds that it is the most important principle of human cognition and “a chunk is a unitof memory organization, formed by bringing together a set of already formed chunksin memory and wielding them together into a larger unit" (Ellis，2001，p. 86).According to Nick C. Ellis (2001，p. 87)，"lexical structure involves identifying thecategorical units of speech perception, their particular sequences in particular words，and their general sequential probabilities in the language." That is to say, in thelearning of lexical items，chunking functions as wielding together the letters into amorpheme or the morphemes into a word. Take the word "national" for example. Thelowest chunks in the word，which might be firstly stored separately, are "N" "A" "T"“1” “0” "N" “A，，"L". The learners later might store two higher level's chunks“NATION” and "AL" before they see the word "national". At last the learnerencounters with ‘national” and forms the larger chunk "NATIONAL". Moreover，thechunk "NATIONAL" can chunk with other chunks, for example with "INTER" to form "INTERNATIONAL"; with “FLAG” to form ‘NATIONAL FLAG", This showsthe whole process of how chunking works as the Language Acquisition Process(LAP).
Base on the Chunking as LAP theory，the present study believes that IVLcould happen in the present study's experimental condition since a one-time encounterwith the pseudo-words conceives the possibility of starting or completing such achunking process.
Mental Process Involved in Reading Comprehension
Researchers (van den Broek，1994; Singer, 1994; Graesser，Singer, &Trabasso, 1994) claimed that reading comprehension was a process and result ofdynamic interaction between the mind and the text and that readers tended to use theirprior knowledge actively during reading. During this process，the reader wouldassociate their prior knowledge about the scenario with the linguistic propositionalcontext of the text to "construct a coherent mental representation" (Cote, Goldman，&Saul, 1998) of the text. This representation is a mental situation, which goes beyondthe linguistic meaning of the text，much image richer than what is displayed in the text,and very helpful when reading. Meanwhile，based on the two kind of information, thereader constructs such a mental representation as their inferences about what has beenread. Three kinds of these inferences with different functions have been identified byempirical studies, which happens whenever there is a reading activity: explanations，associations, and predictions. Explanations occur when we try to elaborate themeaning of the target sentences by retrieving to what was read or experiencedpreviously (Keenan et al.，1984; Myers et al.，1987; Trabasso & Magliano, 1996);Association are the inferences that we make by associating with what we previouslyread in the text and experienced in life (Trabasso & Magliano，1996; Myers & Duffy,1990); predictions is the inferences about what is going to be read next in the text，which is also made based on the previous context in the text and experience in life(Duffy, 1986; Trabasso & Magliano，1996; Fletcher & Bloom，1988).
...................................
Chapter Four Design of the Study...................................28
Research Questions and Hypotheses...................................28
Experimental Design...................................30
Type of the Present Study...................................30
Variables Involved in the Study...................................31
Test of Hypotheses...................................32
Participants...................................34
Reading Materials...................................35
Target Words...................................36
Instruments...................................37
Measures of Background Knowledge...................................37
Measures of Text Comprehension...................................38
Measures of Word-form Recognition (WFR)...................................39
Measures of Translation Recognition Knowledge (TRK)............................40
Measures of Translation Productive Knowledge (TPK)...........................41
The Control of Intervening Variables...................................42
Procedure of the StudyPilot Study...................................44
Pre-test and Reading Proficiency...................................45
Posttest and Delayed Test..............................46
Statistical Methods..................................46
After-test Interview...............................48
Summary........................................48
Chapter Five Results......................................50
Results of Research Question 1.........................50
Effect of Background Knowledge on Text Comprehension..........................50
Effect of Background Knowledge on S&L-term Incidental Word-formLearning................51
Effect of Background Knowledge on S&L-term Incidental ReceptiveSemantic Knowledge Learning...........53
Effect of Background Knowledge on S&L-term Incidental ProductiveSemantic Knowledge Learning.............54
Results of Research Question 2..........................55
Relationship Between Background Knowledge and S&L-termIncidental Word-form Learning55
Relationship Between Background Knowledge and S&L-termIncidental Receptive Semantic Knowledge Learning.........56
Relationship Between Background Knowledge and S&L-termIncidental Productive Semantic Knowledge Learning........57
Results of Questionnaire After Experiment.................................58
Chapter SixAnalysis and Discussions
In this chapter, the inferential explanations and discussions were elaboratedin light of the four hypotheses. Moreover, comparison of the results of the presentstudy with that of Pulido (2004)，along with which explanation about the similaritiesand differences between the results were illustrated. Meanwhile, One importantdiscovery of the present study were pointed out at the end of this chapter.
Analysis and Discussions of Research Question 1Effect of Background Knowledge on Text Comprehension
The result of text comprehension in the present study was completely inaccordance with our expectation that readers* background knowledge (scenariofamiliarity) played a robust role in scenario text comprehension (Barry & Lazarte,1998; Carrell, 1987; Hudson, 1982; Pulido, 2003，2004). A question as which twopassages are more difficult to understand was asked in the questionnaire (see chapter4). The answer was coincident with the statistical result of the present experiment thatabout 93% of the participants found the two less familiar ones were more difficult tocomprehend. As we mentioned in chapter 4，the sentence structural complexities ofthe four texts were about the same level, thus it was background knowledge thataffected the comprehension. Therefore, we reclaimed that the relationship between background knowledge and text comprehension was so highly intimate that they werevery likely to cast impact on the dependent variables in the same way.
However, one might also notice that the participants' recall scores were not as good as we expected, since they only recalled about 20% of the total semanticpropositions. Did this mean that the design or the result of the experiment wasunreliable or invalid? The answer should be "Yes，they were reliable and valid." Thereasons are: first, we examined the recalls of the participants and found that theytended to omit the con-arguments in the texts. The author thought this might bebecause of the ideographic feature of Chinese (the participants' LI) whose usersmaintain the habit of neglect the logic words or conjunctions in written Chinese.Second, during the questionnaire, when the participants were asked whether theyfound the passages too difficult for them to understand, the answer is negative; whenasked what the biggest problem in recalling the passage was, 99% about of themthought that it was because they did not comprehend the passages thoroughly.Therefore, based on the evidences as above，it is safe to say that the statistics from thisstep of the present experiment was reliable and valid.
Effect of Background Knowledge on S&L-term Incidental Word-form Learning
To explain the result of word-form recognition test (see Table 5.3~5.4), welooked into the answer sheets of the participants. There we found that the participantshad 汪 great tendency to identify those words that did not appear in the scripts but wererelated to the scenario as appeared, which caused the unafFirmativeness of theparticipants' choices and in turn increased the alarm rate. This caused the results ofword-form recognition calculated based on the equation in Chapter 4 to be a negativeones. From another angle, the negative score means indicated that a one-timeunexpected encounter with a new word might be not enough for one to thoroughlyacquire the form of that word. According to Schmidt (2010), the forms of words arejust collocation and co-occurrence of letters. Thus，here we might just prove that aone-time exposure to a collocation or co-occurrence was not enough for it to be remembered and stored in memory. Moreover, the brain might just do not pay anyattention to the new sequence (the form of the new word). Thus, it was hard to beprocessed with attentional effort for the sake of being stored, though the sequence andcollocation of the letters were in accordance with the segmental rules of spelling of English words.
.......................................
Chapter SevenConclusion and Recommendation
Besides conclude the answers for the research questions，the author willdevote the talk here to a comprehensive summarization of the key findings of thepresent experimental and to the conservative but still plucky suggestions for thetheoretical, experimental and pedagogical implications from the results. At the end ofthis chapter, comments about the limitations are made, and recommendations forfuture researches are illustrated.
Conclusion
Results and discussion of empirical studies on the present issue weredisplayed in the last chapter and in this chapter. Answers to the research questionswere not completely coincident with the hypotheses. Result analysis about the two%research questions were made based on the patterns of the statistics.
Question 1: During narrative text reading，does background knowledge play arobust role in IVL in aspects of short-term and long-term form-recognition, receptiveknowledge and productive knowledge?
Conclusions based on the results from experiment are as follows:
(1)Under a condition of a one-time encounter during reading narrative texts,in terms of short-term learning effect, background knowledge did not play a robustrole in incidental word-form learning, but did play a robust role in new words'receptive semantic knowledge and productive semantic knowledge learning.
(2)Under a condition of a one-time encounter during reading narrative texts, in terms of long-term learning effect, it followed the same pattern, but with thesignificance degree different from that in the short-term result.
Question 2: Does background knowledge's contribution to the three aspects ofincidental vocabulary learning significantly increase along with the increase ofbackground knowledge in short-term and long-term performance?
(1)In short term, background knowledge's contribution to incidental formlearning was rather weak and insignificant, and it did not give any sign of falling orincreasing along with the increase of background knowledge. Its contribution toincidental receptive semantic knowledge learning was strong and significant, and itincreased when background knowledge increased. Its contribution to incidentalproductive knowledge learning was strong, and increased when backgroundknowledge increased.
......................................
参考文献（略）
Background of the Study
Vocabulary is one of the essential elements of language. To emphasize theimportance of vocabulary, David Wilkins (1972，p.93) claimed "that withoutvocabulary nothing can be conveyed". Norm Chomsky shared the same emphasis onlexis in his Minimalist Program for Linguistic Theory. He stated "that there is onlyone human language apart from lexicon，and language acquisition is in essence amatter of determining lexical idiosyncrasies." (Chomsky, 1995，p. 147) When it comesto applied linguistics，a great number of researchers also highlight the importance ofvocabulary，for example，Gass and Selinker aigued that the lexicon might be the mostimportant language component for learners (Gass & Selinker, 2008). Thus，ever sinceSL acquisition became an issue in SL research, practitioners and researchers havenever moved their eyes away from vocabulary acquisition and vocabulary teaching，and have looked for practical and effective methods to help with the acquisition orlearning of SL vocabulary. In the reservoir of countless experimental studies about SLvocabulary learning, there are two main categories: intentional SL vocabularylearning and incidental SL vocabulary learning. (Laufer，2001) The concept ofincidental vocabulary learning was first considered as a mode of first languagevocabulary learning by Nagy, Herman and Anderson in 1985，which was later adoptedby Krashen (1989) to describe the same kind of phenomenon in SL vocabularyacquisition. IVL refers to "the learning of vocabulary as a by-product of any activity not explicitly geared to vocabulary learning." (Hulstijn, 2001, p. 265)
Many empirical studies have been done to unveil the features of SL IVL.There were empirical studies that investigate the relationship between text type andincidental SL vocabulary acquisition (Herman, 1987), the effect of appearingfrequency on SL IVL (Webb, 2007)，the effect of form-enhancement on it (Paribkht &Wesche, 1997)，the different effect of task types on it (Griffin & Harley, 1996; Laufer,2005), and the relationship between learners' strategies and their efficiency in SL IVL(Potter，Eckardt & Feldman，1984; Tekmen & Daloglu，2006)，and so on. However,one of the most important factors, learners' mental condition，have not been givenenough examination and attention as it deserves. Until now, only quite a fewexperiments have been conducted to study the effect of learners' mental condition ontheir SL IVL, such as Pulido (2003) and Leeser (2007). The key reason that causedsuch a vacancy in study might be that it is comparatively more difficult to studymental activity during IVL in empirical researches than to design experiments withsome non-leamer based and non-mental controlled variables. Still, to the author'sknowledge, there is rarely any empirical study about the relation between ChineseEnglish learners' mental condition and their SL IVL. Thus, there is an urgent need toconduct experimental studies in this aspect, which is right the gist that the presentstudy devotes to.
.....................................
Purpose of the Study
Researching into theoretical and empirical studies on the effect of SL learners'mental condition on their IVL, the researcher found that "background knowledge"(Jacobs, Dufon & Fong.，1994; Leeser, 2007; Pulido, 2003) is one typical factoramong all kinds of factors affecting learners' mental condition. Therefore, the purposeof the present study was to investigate the effects that different levels of background knowledge had on SL IVL, In another word，the present study aimed to settle such aquestion as whether reading materials with more familiar scenarios would foster morechance for IVL and materials with less familiar scenarios would obstruct it. Byachieving such a purpose, the present study would be able to give hints for the futuredesigns of Chinese college students' reading materials, and be able to provide theoriesrelated with empirical evidences.
Significance of the Study
The present study was designed to give thought to gain theoretical andpedagogical significance.
Theoretical Significance
This study took Schmidt's (2010) Basic Assumptions of Attention and Craik &Lockhart's (1972) Levels of Depth of processing theory as the theoretical basis，thusthe statistics and results were devoted to test the two theories' validity in SL IVL.According to the Basic Assumptions of Attention, the attention recourse is limited,and learning needs the allocation of the attention recourse. Thus, readers' allocation ofattention recourse is different when they are reading texts with different levels offamiliarity. That is to say，when we are reading texts with more familiar scenarios,more attention recourses would be allocated to the new word encountered，but we areless likely to do so when reading texts with less familiar scenarios. In addition,according to the theory, allocation of attention would influence the take-ins inworking memory or short-term memory. Thus, a better SL IVL in short-term memorywould give support to Basic Assumptions of Attention; while a worse IVL inshort-term memory would deny the validity of the theory in short-term SL IVL.According to the Levels of Depth of processing theory, only can the newlyencountered knowledge processed in the deep levels of our cognitive mechanisms be stored in long-term memory and become a truly acquired knowledge. In the presentstudy，new words in the less familiar scenario texts are likely to be processed deeperthan that appearing in the more familiar scenario texts. Thus, the present study wouldtest the validity of the Levels of Depth theory in long-term SLIVL.
.....................................
Chapter Two Theoretical Foundation
This chapter elucidated the theoretical foundation which guided the presentstudy and the hypotheses for the research questions. In order to provide the necessarytheoretical background knowledge for the readers, here, in this chapter the authordemonstrated the terminologies, definitions，theories and hypotheses that involved.The elucidations began with issues about IVL; then it explained the theories andhypotheses, including Chunking theory, Comprehension Hypothesis, MentalSchemata and Representation hypotheses, Basic Assumptions of Attention, NoticingHypothesis, and Levels of Depth of processing Hypothesis.
Incidental Vocabulary Learning
TheDefmition oflVL
In Longman Dictionary of Language Teaching and Applied Linguistics, thedefinition for incidental learning is "learning something without the intention to learnit or learning one thing while intending to leam another, for example，unintentionallypicking up vocabulary, patterns, or spelling through interaction, communicativeactivities, or reading for content or pleasure. In controlled experiments, incidentallearning is usually used in a more restricted sense, operationalized as a condition inwhich subjects are not told in advance that they will be tested after treatment,sometimes contrasted with an intentional condition in which subjects are told whatthey will be tested on." (Richards & Schmidt, 2002, p. 252) Nagy, Herman &Anderson，s result in their experiment in 1985 showed that a substantial proportion of Children's vocabulary growth happened probably through incidental learning.However, when we look at the researches on IVA (Laufer & Hulstijn，2001; Joe，1995，1998; Newton, 1995; Pulido, 2003，2004; Wesche & Paribkht，1996), differentinterpretations of IVL，s definition were formulated. For example, Wesche & Paribkht(1996，p. 36) defined incidental learning of vocabulary as what happened whenlearners were "focusing on understanding meaning rather than on the explicit goal oflearning new words.” Thus, it was necessary for the present study to base strictly onthe definition given by Longman Dictionary of Language Teaching and AppliedLinguistics to constitute the definition for IVL. The definition here was learningvocabulary without the intention to leam it, through interaction, communicationactivities or reading for content or pleasure. Moreover, in the present experiment, thecontrolled condition for IVA was reading for content.
Intentional and Incidental Vocabulary Learning
As mentioned above, incidental learning only happens when one isunintended to leam. Thus it is quite necessary to explicit the difference betweenintentional vocabulary learning and IVL.
Longman Dictionary of Language Teaching and Applied Linguistics givesthe definition for the intentional learning as: intentional learning is the "learning byfollowing a deliberate programme of study to enhance vocabulary or grammar"(Richards & Schmidt, 2002，p. 263). Hatch & Brown (1995)，Laufer (2001)，andMondria & Wit (1991) all gave the same kind of definition to intentional vocabulaiylearning that intentional vocabulary learning is the intended and deliberate learningof vocabulary，for example，doing vocabulary exercises (Hulstijn，2001). All thisdescriptions of the features of intentional vocabulary learning emphasized that toaccomplish intentional vocabulary-learning goal，the learners should deliberately engage in a task with leaning the vocabulary as at least one of its learning goals. Thelearning environment is that the learners are informed of receiving some kind of teston their learning about the target words after the learning activity. (Hulstijn，2005)
....................................
Cognitive Linguistic Theories of L2 Vocabulaiy Learning
Since the present study was theoretically rooted in cognitive linguistictheories, in this section，it is reasonable for the author to elaborate the related theoriesin this field here. Theories like Chunking，Comprehension Mental Processing, mentalschemata, representation, Basic Assumptions of Attention, Noticing hypothesis andthe Levels of Processing Hypothesis are introduced here with details.
Chunking as LAP
Chunking is one of the most important terms in cormectionist view aboutlanguage learning and acquisition. Miller (1956，p. 82) first coined the term Chunkingin his review on Short-term memory，and defined it as "the development of permanent'sets of associative connections in long-term storage, which is the process thatunderlies the attainment of automaticity and fluency in language." Connectionistholds that it is the most important principle of human cognition and “a chunk is a unitof memory organization, formed by bringing together a set of already formed chunksin memory and wielding them together into a larger unit" (Ellis，2001，p. 86).According to Nick C. Ellis (2001，p. 87)，"lexical structure involves identifying thecategorical units of speech perception, their particular sequences in particular words，and their general sequential probabilities in the language." That is to say, in thelearning of lexical items，chunking functions as wielding together the letters into amorpheme or the morphemes into a word. Take the word "national" for example. Thelowest chunks in the word，which might be firstly stored separately, are "N" "A" "T"“1” “0” "N" “A，，"L". The learners later might store two higher level's chunks“NATION” and "AL" before they see the word "national". At last the learnerencounters with ‘national” and forms the larger chunk "NATIONAL". Moreover，thechunk "NATIONAL" can chunk with other chunks, for example with "INTER" to form "INTERNATIONAL"; with “FLAG” to form ‘NATIONAL FLAG", This showsthe whole process of how chunking works as the Language Acquisition Process(LAP).
Base on the Chunking as LAP theory，the present study believes that IVLcould happen in the present study's experimental condition since a one-time encounterwith the pseudo-words conceives the possibility of starting or completing such achunking process.
Mental Process Involved in Reading Comprehension
Researchers (van den Broek，1994; Singer, 1994; Graesser，Singer, &Trabasso, 1994) claimed that reading comprehension was a process and result ofdynamic interaction between the mind and the text and that readers tended to use theirprior knowledge actively during reading. During this process，the reader wouldassociate their prior knowledge about the scenario with the linguistic propositionalcontext of the text to "construct a coherent mental representation" (Cote, Goldman，&Saul, 1998) of the text. This representation is a mental situation, which goes beyondthe linguistic meaning of the text，much image richer than what is displayed in the text,and very helpful when reading. Meanwhile，based on the two kind of information, thereader constructs such a mental representation as their inferences about what has beenread. Three kinds of these inferences with different functions have been identified byempirical studies, which happens whenever there is a reading activity: explanations，associations, and predictions. Explanations occur when we try to elaborate themeaning of the target sentences by retrieving to what was read or experiencedpreviously (Keenan et al.，1984; Myers et al.，1987; Trabasso & Magliano, 1996);Association are the inferences that we make by associating with what we previouslyread in the text and experienced in life (Trabasso & Magliano，1996; Myers & Duffy,1990); predictions is the inferences about what is going to be read next in the text，which is also made based on the previous context in the text and experience in life(Duffy, 1986; Trabasso & Magliano，1996; Fletcher & Bloom，1988).
...................................
Chapter Four Design of the Study...................................28
Research Questions and Hypotheses...................................28
Experimental Design...................................30
Type of the Present Study...................................30
Variables Involved in the Study...................................31
Test of Hypotheses...................................32
Participants...................................34
Reading Materials...................................35
Target Words...................................36
Instruments...................................37
Measures of Background Knowledge...................................37
Measures of Text Comprehension...................................38
Measures of Word-form Recognition (WFR)...................................39
Measures of Translation Recognition Knowledge (TRK)............................40
Measures of Translation Productive Knowledge (TPK)...........................41
The Control of Intervening Variables...................................42
Procedure of the StudyPilot Study...................................44
Pre-test and Reading Proficiency...................................45
Posttest and Delayed Test..............................46
Statistical Methods..................................46
After-test Interview...............................48
Summary........................................48
Chapter Five Results......................................50
Results of Research Question 1.........................50
Effect of Background Knowledge on Text Comprehension..........................50
Effect of Background Knowledge on S&L-term Incidental Word-formLearning................51
Effect of Background Knowledge on S&L-term Incidental ReceptiveSemantic Knowledge Learning...........53
Effect of Background Knowledge on S&L-term Incidental ProductiveSemantic Knowledge Learning.............54
Results of Research Question 2..........................55
Relationship Between Background Knowledge and S&L-termIncidental Word-form Learning55
Relationship Between Background Knowledge and S&L-termIncidental Receptive Semantic Knowledge Learning.........56
Relationship Between Background Knowledge and S&L-termIncidental Productive Semantic Knowledge Learning........57
Results of Questionnaire After Experiment.................................58
Chapter SixAnalysis and Discussions
In this chapter, the inferential explanations and discussions were elaboratedin light of the four hypotheses. Moreover, comparison of the results of the presentstudy with that of Pulido (2004)，along with which explanation about the similaritiesand differences between the results were illustrated. Meanwhile, One importantdiscovery of the present study were pointed out at the end of this chapter.
Analysis and Discussions of Research Question 1Effect of Background Knowledge on Text Comprehension
The result of text comprehension in the present study was completely inaccordance with our expectation that readers* background knowledge (scenariofamiliarity) played a robust role in scenario text comprehension (Barry & Lazarte,1998; Carrell, 1987; Hudson, 1982; Pulido, 2003，2004). A question as which twopassages are more difficult to understand was asked in the questionnaire (see chapter4). The answer was coincident with the statistical result of the present experiment thatabout 93% of the participants found the two less familiar ones were more difficult tocomprehend. As we mentioned in chapter 4，the sentence structural complexities ofthe four texts were about the same level, thus it was background knowledge thataffected the comprehension. Therefore, we reclaimed that the relationship between background knowledge and text comprehension was so highly intimate that they werevery likely to cast impact on the dependent variables in the same way.
However, one might also notice that the participants' recall scores were not as good as we expected, since they only recalled about 20% of the total semanticpropositions. Did this mean that the design or the result of the experiment wasunreliable or invalid? The answer should be "Yes，they were reliable and valid." Thereasons are: first, we examined the recalls of the participants and found that theytended to omit the con-arguments in the texts. The author thought this might bebecause of the ideographic feature of Chinese (the participants' LI) whose usersmaintain the habit of neglect the logic words or conjunctions in written Chinese.Second, during the questionnaire, when the participants were asked whether theyfound the passages too difficult for them to understand, the answer is negative; whenasked what the biggest problem in recalling the passage was, 99% about of themthought that it was because they did not comprehend the passages thoroughly.Therefore, based on the evidences as above，it is safe to say that the statistics from thisstep of the present experiment was reliable and valid.
Effect of Background Knowledge on S&L-term Incidental Word-form Learning
To explain the result of word-form recognition test (see Table 5.3~5.4), welooked into the answer sheets of the participants. There we found that the participantshad 汪 great tendency to identify those words that did not appear in the scripts but wererelated to the scenario as appeared, which caused the unafFirmativeness of theparticipants' choices and in turn increased the alarm rate. This caused the results ofword-form recognition calculated based on the equation in Chapter 4 to be a negativeones. From another angle, the negative score means indicated that a one-timeunexpected encounter with a new word might be not enough for one to thoroughlyacquire the form of that word. According to Schmidt (2010), the forms of words arejust collocation and co-occurrence of letters. Thus，here we might just prove that aone-time exposure to a collocation or co-occurrence was not enough for it to be remembered and stored in memory. Moreover, the brain might just do not pay anyattention to the new sequence (the form of the new word). Thus, it was hard to beprocessed with attentional effort for the sake of being stored, though the sequence andcollocation of the letters were in accordance with the segmental rules of spelling of English words.
.......................................
Chapter SevenConclusion and Recommendation
Besides conclude the answers for the research questions，the author willdevote the talk here to a comprehensive summarization of the key findings of thepresent experimental and to the conservative but still plucky suggestions for thetheoretical, experimental and pedagogical implications from the results. At the end ofthis chapter, comments about the limitations are made, and recommendations forfuture researches are illustrated.
Conclusion
Results and discussion of empirical studies on the present issue weredisplayed in the last chapter and in this chapter. Answers to the research questionswere not completely coincident with the hypotheses. Result analysis about the two%research questions were made based on the patterns of the statistics.
Question 1: During narrative text reading，does background knowledge play arobust role in IVL in aspects of short-term and long-term form-recognition, receptiveknowledge and productive knowledge?
Conclusions based on the results from experiment are as follows:
(1)Under a condition of a one-time encounter during reading narrative texts,in terms of short-term learning effect, background knowledge did not play a robustrole in incidental word-form learning, but did play a robust role in new words'receptive semantic knowledge and productive semantic knowledge learning.
(2)Under a condition of a one-time encounter during reading narrative texts, in terms of long-term learning effect, it followed the same pattern, but with thesignificance degree different from that in the short-term result.
Question 2: Does background knowledge's contribution to the three aspects ofincidental vocabulary learning significantly increase along with the increase ofbackground knowledge in short-term and long-term performance?
(1)In short term, background knowledge's contribution to incidental formlearning was rather weak and insignificant, and it did not give any sign of falling orincreasing along with the increase of background knowledge. Its contribution toincidental receptive semantic knowledge learning was strong and significant, and itincreased when background knowledge increased. Its contribution to incidentalproductive knowledge learning was strong, and increased when backgroundknowledge increased.
......................................
参考文献（略）
收藏