The Texture of Inefficiently Self-Regulating ESL Systems

The Texture of Inefficiently Self-Regulating ESL Systems (8, 858 words, with 141-word abstract and 68-word biographical note) by Terence Patrick Murphy Dept of English, Yonsei University, Seoul, Korea 120-749 Biographical Note Terence Patrick Murphy teaches in the Department of English and the Underwood International College at Yonsei University in Seoul, South Korea. In addition to a pioneering essay on the concept of emergent texture that appeared in Language Learning and Technology, he has published numerous essays on literary stylistics in such journals as the Journal of Narrative Theory; the Journal of Literary Semantics; Language and Literature; Style and Narrative. Dept of English Tel. 82-2-2123-2300 Dept of English Fax: 82-02-392-0275 [email protected] Home Telephone: 82-2-391-4307 February 2007 Abstract This chapter addresses the question of how to measure the student’s English second language (ESL) textual sophistication. It suggests that the second language text is an inefficiently self-regulating system, at the levels of grammar, lexis and logico-rhetorical structure. Learner texts use a narrow or even fixed set of key lexical phrases; they deploy cohesive ties that bind the text incorrectly, they omit cohesive ties altogether, or redundantly retain items that are easily recovered from the situational context. Following a review of some typical second language cohesion problems, the chapter offers an analysis of the emergent texture of four versions of the same paper, each written by a different ESL student. The results suggest that a learner text-maker is unable to perceive the ineffective choices in texts written at levels of sophistication higher than those he or she is capable of creating. 2 Introduction In the last ten years, the investigation of the written texts of English second language (ESL) learners has turned increasingly to the use of computer-aided corpus analysis (Freedman et al, 1979; Kroll, 1990; Beaugrande, 1997; Granger, 2002; Yoon and Hirvela, 2004). Since it is strongly committed to the acceptance of the evidence found in large sources of natural format data, corpus research offers a means for establishing Robert de Beaugrande’s triumvirate of normal science attributes at the heart of second language acquisition research: convergence, consensus, and coverage (Sinclair, 1991; Beaugrande, 1997). One major initial concern within this new field of study has been the issue of how to measure the student’s growing ESL sophistication. To date, a majority of the applied linguists who have investigated this issue believe that an adequate explanation must focus predominantly on some notion of lexical richness (Laufer and Nation, 1998; Shaw and Liu, 1998; Wolfe-Quintero, Inagaki and Kim, 1998; Meara, 2005; Laufer, 2005). In this chapter, however, I will argue that an exclusive focus on lexis offers an oversimplified account of the ability of the second language writer to produce written English texts. As an alternative, I will suggest that applied linguists need to offer a three-fold account of second language textual development, one that incorporates grammar, lexis and logico-rhetorical structure. The Text as a Self-Regulating System A well-written text may be defined as a self-regulating system. The wellwritten text exhibits an informational dialectic of ease and efficiency in a register appropriate for the situational context (Iser, 1978; Beaugrande, 1980). As a selfregulating system, the well-written text provides a set of directives to enable the reader 3 to probe a variety of intra-textual (and extra-textual) relations to a certain depth at distinct grammatical, lexical and logico-rhetorical levels, without giving rise to insoluble discrepancies, paradoxes, ambiguities, or contradictions (Beaugrande, 1980; Murphy, 2005a). If the texts of ESL students are considered as a special case of the more general category of the poorly written text, what this means is that the typical learner text is inefficiently self-regulating. One of the major reasons why this is so is because the second language text possesses emergent texture (Murphy, 2001). In other words, because the grammatical, lexical and logico-rhetorical relations in the second language text remain underdeveloped, many of its textual directives lead to ambiguities, discrepancies, paradoxes, contradictions and redundancies. Particularly for those readers who have had limited experience working with learner texts, the result is frequently frustration and confusion. The Murphy-Lee Second Language Corpus The Murphy-Lee Second Language Corpus is a personal collection of Korean English-language learner texts that I have been assembling with the assistance of my wife, Lee Joon-kyoung, in the Department of English at Yonsei University in Seoul, Korea since 1999. As of Fall 2006, the total word count is approximately 100,000 words, with the corpus increasing at the rate of about 15,000 words per year. Although the corpus contains the work by students in a number of different humanities departments at the university, the bulk of the material is by students in the English Department. For the past five years, I have focused on the collection of five-paragraph essays. All of the five-paragraph essays in the corpus include at least one revision; many of them include two. These essays are of two principle types: discussions of movie themes and literary 4 interpretation of short stories, mostly by James Joyce. However, there are also a number of other genres represented in the corpus, including about 120 single paragraphs as well as an increasing number of graduate-level summaries and critiques of published linguistics papers. In addition, a small set of assignments have involved the revision of a low-level learner text. These last assignments were undertaken as part of a one-week take-home end-of-term examination. Lexis or Grammar? In the Hallidayean account of English grammar, the formation of the five key word groups—the nominal group, the verbal group, the adverbial group, the adjectival group, and the prepositional group—involves the text-maker in simultaneous grammatical and lexical word choices. The grammatical items consist of that set of words involved in a small fixed number of obligatory choices. As Halliday, McIntosh and Strevens suggest: “There are some places in every language where we have to make such choices; we cannot avoid them or remain neutral, and there is a limited number of possibilities to choose from” (1964, p. 21). For example, in English, there is the obligatory choice between this and that; or the obligatory choice among who, whose, what, and which; the obligatory choice between the use of the singular and the plural; the obligatory choice between the use of the present or another tense; and the choice between the positive and negative use of language items. This first set of words consists of those language items where “we face a choice among a very small number of possibilities” (Halliday, McIntosh and Strevens, 1964, p. 21). In contrast, lexis refers to language events where meaningful choice involves “a very large number of possibilities” and where it is very difficult to separate out “what is possible from what is 5 impossible” (Halliday, McIntosh and Strevens, 1964, p. 21). The second language text-maker’s task thus unfolds on at least three different levels. Moreover, each of these three levels appears to develop not simultaneously but rather unevenly. In other words, grammar, lexis, and logico-rhetorical structure present the second language text-maker with different kinds of difficulty. Indeed, for Korean learners for whom the mastery of the article system is a life-long undertaking, it may be that lexical collocation is not necessarily the most difficult among these tasks (But cf. Wray 2002). If this is so, a measure of lexical sophistication will not be the most accurate means for determining a text-maker’s second language development, as supporters of the concept of the Lexical Frequency Profile would maintain (Laufer & Nation, 1998; Laufer, 2005). A more sensitive measure, it is suggested, is the development of that aspect of the learner’s ESL grammar that is involved in the formation of texture. The Cohesive Ties and Emergent Texture Textual cohesion is established by means of the five major forms of cohesive tie. The first four sets are the ties of reference, substitution, ellipsis and conjunction (Halliday and Hasan, 1976, p. 4). A central measure of the text-maker’s language abilities is therefore the ease with which he or she can deploy the full range of ties. In contrast, a limited ability on the part of the text-maker to form some or all of these ties restricts the learner’s text-making abilities. According to Halliday and Hasan, “a text has texture, and this is what distinguishes it from something that is not a text. It derives this texture from that fact that it functions as a unity with respect to its environment” (1976, p. 2). If competent first language texts exhibit texture, learner texts possess a 6 differentiated range of emergent textures. At one extreme, examples exist of virtually unreadable second language texts. These texts are presumably created with almost no consideration at all for grammar, lexis and logico-rhetorical structure. The textual relationships among their clauses, sentences, text segments and the text are therefore sometimes unrecoverable: In our society, sex merchandising come in touch with the prosperity of merrymaking place that try conclusions by barmaids’s nude service. the ‘Boss landing operations’ uncovered the merrymaking culture frankly. the sticky scenery of merrymaking place shield the subject of this comic movie that a public prosecutor confront with gang.1 In this example, it is virtually impossible to reconstruct the intended meaning of the text-maker. It is clear that the subject of the text is a comic movie, which is apparently entitled “Boss Landing Operations”. The movie seems to be about a public prosecutor who is in conflict with a criminal gang. For reasons that are obscure, the movie is representative of a society in which sex has become commercialized. Beyond this, it is difficult to say much more. The text-maker’s inability to recognize that he has presented a virtually incomprehensible text is strong evidence of a very low second language ability. Near the middle of this continuum, readable texts with highly unusual wording occur. These texts also occasionally incorrectly signal the relations among the word groups, clauses and sentences in a way that is initially confusing: People sometimes want to have inner resources and so do I. All That Jazz—the jazz 1 This first example is so obscure that I have begun to suspect that it may have been produced by the automatic translation of an original Korean text into English, using the freely available Internet program, Babelfish. If this is the case, the inability of the student to recognize the incoherence of the translated text is evidence of a very low ESL ability. 7 bar which is located in Itaewon is my favorite bar for that reason. There is a wooden stair which leads to the bar, because it is upstairs. Old jazz musicians pictures are hung on the both side of the stairway. End of the stairway, there is a door which moved slightly by the jazz harmony. In the first sentence, “People sometimes want to have _________ __________.”, the Googled choice of the next two words includes such items as a baby or a relationship or a scrap or more fun. For Halliday, what this means is that “lexis seems to require the recognition merely of linear coexistence together with some measure of significant proximity, either a scale or at least a cut-off point” (Halliday, 1966, p. 152). Theoretically, of course, the level of individual word detail required to specify such a measure could be written down in a sufficiently large dictionary. As Halliday suggests: No grammar has, it is believed, achieved the degree of delicacy required for the reduction of all such items to one-member classes, although provided the model can effectively handle cross-classifications, it is by no means absurd to see this as the eventual aim: that is, a unique description for each item by its assignment to a ‘microclass’, which represents its value as the product of the intersection of a large number of classificatory dimensions. (1966, p. 149) It is Halliday’s idea of significant proximity or collocation that the first language writer draws on when he or she intuitively recognizes, as a second language writer may not, that the class of items that can successfully conclude a sentence like: People sometimes want to have _________ __________. does not include the unusual second language lexical collocation inner resources. Finally, there are those learner texts that employ mostly correct phrasing and clause structure. In this sense, these learner texts are barely distinguishable from the 8 sophisticated self-regulating systems produced by first language writers: Jae-na is a nice girl, but she is so shy that she cannot speak to an audience well. Because of her timidity, a pink glow often mounts her cheeks. Her personality influences on her style of clothing. She likes clothes which expose her body little. Because she wants not to attract public attention, she wears dark-colored garments. The major difference between first language texts and this second language text is the latter’s limited ability to deploy a sophisticated lexical repertoire. As a result, it has a somewhat sophomoric quality to it. Methodology: Markedness and Emergent Texture A large number of problems relating to the inefficient cohesion achieved by ESL texts appear to be related to the text-maker’s selection of the unmarked, rather than the marked, textual option (Murphy 2001). In other words, the text-maker apparently does not understand the textual motivation that lies behind the choice of the marked term. As a result, many learner texts consist of what reads like a series of discrete sentences rather than being unified textual products tied together with the appropriate cohesive devices. In learner texts, the inability to plan prospectively (or, more rarely, to plan retrospectively) means the text may falter at the level of the next prepositional phrase or the next sentence or the next discrete text segment. For example, in the following example, the learner text-maker begins with a context-specifying prepositional phrase but then follows this repeatedly with the unmarked form of the definite article: For example, in "Notting Hill", a main woman character is the famous actress in Hollywood but a main man character is 35-year-old divorcee who runs a travel 9 bookstore in Notting Hill, London. In "Pretty Woman", a man is a rich businessman who visits Hollywood on business but a woman is a street girl. As Halliday and Hasan write, “the definite article … merely indicates that the item in question IS specific and identifiable; that somewhere the information necessary for identifying it is recoverable” (1976, p. 71). In the sample learner text, although the learner text-maker has signaled that the information is specific (in “Notting Hill” … in “Pretty Woman”), she does not coordinate this logico-rhetorical structure with the choice of the definite article. The next example demonstrates the tendency of the learner text-maker to favour the unmarked or non-selective it (instead of the marked cohesive this) in the signaling of reference continuity: It's free to enter and exit, but if students check out library materials without appropriate steps and permission, it will be detected by book detection system and students will be punished according to the regulations. Learner text-makers also tend to rely on the neutral non-selective definite article the (rather than the cohesive demonstrative this) in establishing reference back to a previously mentioned lexical item: In front of the door or the counter, sofas are put from left to right facing a large screen which displays all sorts of rock music clips. On the screen you can see several genre clips that are from USA, Japan, Europe and even the Third World. The learner text-maker may also demonstrate confusion in choosing the appropriate cohesive device: In addition when they shall be old, they want to live with son's family and take it for granted that son should support their old parents. They rarely expect daughters to do so for their parents. 10 This final example may suggest that the various forms of cohesive substitution— nominal, verbal, and clausal—represent a hierarchy of grammatical difficulty. In this regard, Jo Soo-jin, one of my former graduate students, has observed that the instances of nominal substitution that occurred in a small 70-item sub-corpus of five-paragraph essays from the Murphy-Lee Corpus showed the single pattern of a/the + adjective + one(s) (Jo, 2003). Without empirically verifying this, Jo posits the following learning order for the cohesive ties of substitution: nominal one, verbal do, nominal same, plural ones, clausal so, combined do so, and clausal not. Lexical directives constitute the fifth set of cohesive ties. According to Halliday and Hasan, “reiteration is a form of lexical cohesion which involves the repetition of a lexical item, at one end of the scale; the use of a general word to refer back to a lexical item, at the other end of the scale; and a number of things in between—the use of a synonym, near-synonym, or superordinate (Halliday and Hasan, p. 278). Reiteration in low-level learner texts takes place at the end of the scale marked out by repetition. This form of cohesion involves simple lexical repetition and the neutral non-selective use of the definite article as an anaphoric device. In low-level learner texts, the tendency is for the text-maker to introduce a specific lexical item using the indefinite articles a or an and then to switch to the use of the non-selective definite article, the (Murphy 2001). It is also quite common in such texts for this kind of lexical repetition to take the simplest form of unadorned Modifier and identical Head. In other words, each of the other possible elements—Numerative, Epithet, Classifier, Thing and Qualifier—remain unrealized (Murphy 2001). The upshot of this is that second language texts are inefficient at the level of lexis. By relying on the use of exact lexical repetition to establish continuity of reference, these texts fail to take advantage either of the effort- 11 saving substitution of appropriate pro-forms or of the appropriate curtailment of language items that are immediately recoverable from the situational context in establishing continuity of reference: Feminism is a movement of extension of women's rights that were reduced in the past society of patriarchal system that man were believed to be superior to women. In other words, feminism movement is a movement for equality between men and women. So, it is needed to try to know how women have been treated discriminately and how women try to be treated on the same footing as men. At first, we need to know how women have been treated at home (my emphasis). In this second language text, the burden of establishing overall continuity of reference is presumably so great that little consideration is given to the high degree of redundant items that are present in this text. The most appropriate frame of reference for this aspect of second language texture is the range of lexical choice in comparable first language texts. Consider, for example, the New York Times article entitled “Another Attack in Central Iraq Kills Another U.S. Soldier”. In this first language text, a central chain of reference is grown out of the initial title reference to “Another U.S. soldier”: American soldiers … one American … military officials … the third American soldier … Five other American soldiers … American commanders … American forces … United States commanders … 4,000 soldiers … a 1,200-member armored cavalry squadron … Military commanders … the increased American presence … American military might … American soldiers … United States forces … the new American approach … A dozen American soldiers … they … Americans … American forces … soldiers … officials from the Third Armored Cavalry Regiment … Reinforcements … military officials … more American deaths … American 12 troops … soldiers … American troops … their troops … them … American misconduct … the American search Among the lexical choices in this newspaper text, there are superordinate items (American forces, American troops), synonyms (American) and near-synonyms (American presence, American military might), as well as word items indicating a range of hierarchical (military commanders, military officials) and horizontal relations (officials from the Third Armored Cavalry Regiment, Reinforcements). These first language choices provide the most appropriate frame of reference for analyzing second language choices in similar situational contexts. In contrast, here are the textually established lexical ties in a learner text dealing with the subject of the US military presence in Korea: the presence of U. S. military in Korea…the legitimacy of the U. S. military in Korea…the U.S. military in Korea…it…it…the presence of U. S. military in Korea…it the very U. S. military…the presence of U. S. military…the legitimacy of U. S. military in Korea…the U. S. military…the one in Korea…the U. S. military in Korea…the U. S. military in Korea…the U. S. military in the region…the U. S. military in Korea…the U. S. military in Korea…it in the Peninsula…the U. S. military in Korea…the U. S. military What these textual choices reveal is that at a certain level of sophistication learner textmakers demonstrate little or no ability to vary lexical reference by means of synonyms and near-synonyms. In this text, the full collocation the U.S. military in Korea occurs ten times, the shorter collocation the U.S. military occurs four times, and the variant the U.S. military in the region occurs once. Besides this, the learner text-maker is only able to vary this monotonous set unsteadily with the use of the pro-form it and the attempted 13 use of the substitute one (Hwang, 2003). Learner texts, in other words, are inefficiently self-regulating systems. They are highly repetitive in their use of a narrow or even fixed set of key lexical phrases; they deploy cohesive ties that bind portions of the text incorrectly, they omit to deploy cohesive ties altogether, or else they redundantly retain items that are immediately recoverable from an established situational context. Logico-Rhetorical Structure Within functional grammar, grammar and lexis are both well articulated concepts. These two concepts, however, do not exhaust the description of the self-regulating nature of textual systems. As Halliday suggests, “a text has structure, but it is semantic structure, not grammatical. Just as a syllable has a phonological structure, and a clause has a grammatical structure, a text has a semantic structure…. For a text to be coherent it must be cohesive; but it must be more besides” (Halliday, 1994b, p. 339). Textual coherence—that “more besides”—is established principally by the text’s deployment of a specific logico-rhetorical structure. The text’s logico-rhetorical structure is the work of the punctuation and a specific set of word items distinct from both the grammar and the lexis. Within Hallidayean grammar, the clause consists of two major elements: a Given Element and a New Element. The Given Element is what the text-maker supposes the reader already to know, while the New Element is what the text-maker suggests the reader ought to find interesting or unusual or newsworthy. According to Halliday, “what constitutes the ‘main point’ of the discourse is any motif that figures regularly as New” (1994a, p. 146). The New element, which is realized “prosodically, by greatest pitch movement in the tone group” is what the writer suggests the reader is “to attend to” (1994a, p. 140). 14 Well-written texts progressively accumulate information through a skillful deployment of the dialectic of Moves and Comments. Move clauses are used to drive the text forward, by picking up the New Information that has previously been presented and extending it in some way. In contrast, Comment clauses purposefully slow down the rate of information presentation by elaborating a portion of the previous sentence’s Given Information, typically reaching back for that information with cohesive devices. Well-written texts skillfully utilize this dialectic to drive toward clear textual goals. What is more, well-organized texts regularly utilize normal or typical or unmarked order to present their information, while reserving the choice of unusual or untypical or marked order for special informational purposes. Poorly organized texts, on the other hand, present information in the form of discrete text segments but fail to relate the semantic meaning of these individual text segments to the overall textual goal. Poorly organized texts show only haphazard recognition of the fundamental dialectic of Moves and Comments. They work with formalistic notions of textual organization and overall textual meaning. The occasional incoherent stretches of information presentation in inefficiently organized texts result from the text-maker’s choice of a logico-rhetorical structure that does not facilitate accurate processing. Essentially, what badly organized texts do is to use an inappropriate or poorly elaborated structure to signal the text’s main point. One way in which badly organized texts do this is by offering inappropriate Thematic signals for the information they haphazardly present as the main point of the text (cf. Fries, 1994; Halliday, 1994a). In other words, what gets construed as the “main point” is often not really the main point at all. In poorly organized texts, the relations between the main and the subordinate information are confused. The text-maker often sidelines what 15 ought to be central and erroneously highlights what ought to be peripheral. The founder of functional grammar has suggested that a written text represents a system of fixed sentence choices. In other words, in his theoretical writing, Michael Halliday haphazardly defends the view that choice at the level of the individual sentence in a given text is already optimized. The reason that the founder of functional grammar suggests this is because he had not found a way to incorporate the notion of valid choice above the level of the sentence into his theory. However, it is more accurate and productive to suggest the following proposition: any written text exists as merely one among a number of possible textualizations. In other words, the written text needs to be construed as a potential system for optimal information clausing. The concept of alternative textualization is important for all aspects of textual linguistics, but it has a particular significance within ESL textual studies. This is because the ability to retextualize or rewrite a poorly self-regulating learner text is one of the most significant tests of a learner’s second language ability. What makes this task possible is the recognition on the part of the text-maker of the possibility for deploying alternative logico-rhetorical structuring of the same information. What follows are four versions of the same text. The first text is the original, composed by a low-level second language text-maker. The second, third and fourth are all rewritten versions composed by other students as part of an end-of-term one-week take-home examination for an intermediate English course at Yonsei University. The students in the class were drawn from a variety of disciplines within the humanities, including law, dentistry and business administration. The second text has been composed by a second language text-maker of roughly comparable ability to that of the original text-maker; the third, by a mid-level second language text-maker; the fourth, by 16 a text-maker whose language abilities approach that of a first language writer. Comparison and Contrast of Two Popular Movies The movie "Notting Hill" which was released in this summer can be compared to the movie "Pretty Woman" which was made in 1990. In first, a leading actress is the same person, Julia Roberts. She was very proper and well-matched to the movies. Especially in "Notting Hill", playing a role not too far removed from her own existence, Roberts gave an understated, nuanced performance. Besides, the basic structure of two movies' story are very similar. Main characters who are different in social class and living circumstance meet accidentally and fall in love with each other. However, they have misunderstandings, conflicts, and obstacles in their love because they have lived in very different ways. So they break up. In spite of that, they meet again and certify their own love. Someone say that the movies are modern versions of 'Cinderella'. Both of two movies show us love story which is hard to exist in real world. Though entire structure is very similar, the rest— occupation of character, background, and happenings—are very different. For example, in "Notting Hill", a main woman character is the famous actress in Hollywood but a main man character is 35-year-old divorcee who runs a travel bookstore in Notting Hill, London. In "Pretty Woman", a man is a rich businessman who visits Hollywood on business but a woman is a street girl. "Notting Hill" and "Pretty Woman" are similar and different In this low-level learner text, many of the language choices have not been optimized. For example, at the level of grammar, the text-maker makes such choices as “a leading actress”, “to the movies”, “their own love”, “Both of two movies”, “in [the] real world”, “a main woman character”, “a main man character”, “[a] 35-year-old divorcee”, “a 17 man”, “a woman”. At the level of lexis and collocation, the text-maker utilizes such non-standard choices as “very proper and well-matched”, “certify their own love” and “hard to exist in the real world”. At the level of logico-rhetorical structure, the textmaker demonstrates a lack of ability to achieve overall textual unity. This is chiefly signaled by the utilization of the Thematically marked sequence, “The movie "Notting Hill" which was released in this summer can be compared to the movie "Pretty Woman" which was made in 1990. In first, … Besides, … Someone say that … Though entire structure is very similar,…”. In addition, the text-maker frequently end-focuses subordinate information: “which was made in 1990”, “because they have lived in very different ways” and “So they break up”. Indeed, the most glaring aspect of this more general problem is the three-sentence text segment that deals with the acting ability of Julia Roberts. Finally, we might note that the self-regulating nature of the textual system itself is occasionally impaired. For example, whereas the reader must move forward in the text to discover that “the rest” may be resolved as “occupation of character … the famous actress … 35-year-old divorcee who runs a travel bookstore … a rich businessman … a street girl” and as “background … in Hollywood … in Notting Hill, London”, the textual resolution of “happenings” requires a different strategy. The meaning of this lexical item must be resolved by either a backwards search to the language item “basic structure” or by appeal to the reader’s outside world knowledge (Beaugrande and Dressler, 1981). Whichever of these two search routes is utilized, in logico-rhetorical terms, what this means is that the text has requested through the use of a co-ordinate arrangement of lexical items that a particular feature shared by the two movies (“happenings”) be processed in a like manner to two in which they differ (“occupation”, “background”). 18 Here is a second version of the text, rewritten by another text-maker of broadly similar ability: The movie “Notting Hill” which was released in this summer can be compared to the movie “Pretty woman” which was made in 1990. Though entire structure is very similar, the character’s social class and living circumstances are very different. For example, in “Notting Hill”, a main woman character is the famous actress in Hollywood but a main man character is 35-year-old divorcee who runs a small travel bookstore in Notting Hill, London. In “Pretty woman”, a man is a rich businessman who visit Hollywood on business but a woman is a street girl. However, these differences are an important factor contributing the similarity of the basic structure of two movies which tells us a love story that is hard to exist in real world. In both movies main characters who are different in social class and living circumstance meet accidentally and fall in love with each other. Though, in the middle of the stories of both movies, they broke up unfortunately because of misunderstandings, conflicts, and obstacles which are caused by very different living circumstances and social class, their relationship turns out to be true love at the end. They certified their own love. Besides this structural similarity that tells us unrealistic but romantic love story, the same leading actress, Julia Roberts, was very proper and well matched to the both movies. Her understated, unanced performance gave great contribution for both movies, especially in “Notting Hill”, playing a role not too far removed from her own existence. “Notting Hill” and “Pretty woman” are different in character’s social class and the living circumstance but similar in telling us romantic love story. In this version of the text, the majority of the textual revisions appear to be motivated 19 chiefly by the replacement of the unmarked/marked ordering structure of the original text by a marked/unmarked one in the second. As a result, the second text is somewhat longer and probably less efficient. The text-maker’s substitution of the marked/unmarked order has involved reworking somewhat haphazardly certain of the clause elements: “Besides this structural similarity that tells us unrealistic but romantic love story, the same leading actress, Julia Roberts, was very proper and well matched to the both movies. Her understated, unanced performance gave great contribution for both movies, especially in “Notting Hill”, playing a role not too far removed from her own existence”. This reordering does not represent an improvement. A second rewording occurs as part of the description of the plot. Here, the major changes are the slight thematic colouring and the attempt to add a more definite resolution: “Though, in the middle of the stories of both movies, they broke up unfortunately because of misunderstandings, conflicts, and obstacles which are caused by very different living circumstances and social class, their relationship turns out to be true love at the end. They certified their own love.” Once again, this rearrangement is not noticeably better than the wording employed by the original. What these ineffectual revisions indicate is something profoundly important for second language studies: a learner text-maker is simply unable to perceive the ineffective choices in texts at levels of sophistication higher than those he or she is capable of creating. It is for this reason that this learner text-maker responds to the task of optimizing inefficient language choices in another text by merely redistributing the original inefficiencies. This insight is of crucial importance to the theory of second language textual ability. Among other things, the idea that learner text-makers can only redistribute the inefficiencies in texts composed at a higher level of language 20 development provides a secure means for gauging many aspects of second language development. Here now is the same text rewritten by a second language text-maker of mid-level ability: The movie "Notting Hill" which was released in this summer can be compared to the movie "Pretty Woman" which was made in 1990. First, the main actress is the same person, Julia Roberts. She performs her part with brilliance in both movies, even though the roles of her characters are totally different; one is a Hollywood star as she is in real world, and the other is a prostitute. She acts so well that she looks proper and natural in both movies. Secondly, the basic structure of the plot is similar. The two movies are modern versions of Cinderella love story, which is hard to exist in the real world. Main characters who are different in social class and living circumstance meet accidently and fall in love with each other. They have misunderstandings and conflicts at first because they have lived in very different ways. In spite of that, they realize that they have found their true love after all in each other. Although the basic plots are similar, the details of the movies are different in terms of occupations of characters and backgrounds. In "Notting Hill", the main female character, who is a world famous movie star, visits Notting Hill in London where the main male character runs a small travel bookstore. In "Pretty Woman", the man is a rich businessman who visits Hollywood on business whereas the woman is a street girl. Another different thing between the movies is who plays the Cinderella role. In "Pretty Woman", Julia Roberts meets her perfect prince who she has dreamt of whereas in "Notting Hill" the male character, who is a 35-year- 21 old ordinary divorcee, falls in love with the great actress, who is Roberts. For these reasons, the two movies can be seen as similar and different. At the level of grammar, the choices demonstrate a marked improvement. For example, “the main female character”, “a world famous movie star”, “the main male character”, “a small travel bookstore”, “the man”, “a rich businessman”, “the woman”, and “a street girl” are all correctly chosen. Nonetheless, there are still a couple of inappropriate choices: for example, “in [the] real world”, “[the] Cinderella love story” and “[The] main characters”. At the level of lexis and collocation too, a number of the choices have now been optimized. Instead of “She was very proper and well-matched to the movies”, the text now reads: “she looks proper and natural in both movies”; while instead of “certify their own love”, this text now reads “found their true love … in each other”. However, the non-standard collocation “which is hard to exist in the real world” is incorrectly retained. At the level of logico-rhetorical structure, this text correctly fronts the description of the plot structure with its Orientation: in other words, the idea that these two movies are Cinderella stories now precedes the description of the two movies’ common plot structure. This second text also eliminates the inappropriately end-focused subordinate clause “So they break up” in order to prevent its possible misreading as the climax of the plot. However, the logico-rhetorical structuring “at first” which has been added is positioned after the complement “misunderstandings and conflicts”. In other words, the text-maker has not recognized that the optimal position would be as a marked Theme to guide the reader’s processing of the entire clause: “At first, they have misunderstandings and conflicts because they have lived in very different ways”. Finally, we can compare both re-textualizations with a fourth version written by a student with a very high level of second language sophistication: 22 "Notting Hill" and "Pretty Woman" are modern day versions of 'Cinderalla'. Both movies feature Julia Roberts as a starring role in a love story which is hard to exist in the real world. The basic structure of the two movies are very similar. Characters coming from entirely different social class and culture meet each other by chance and fall in love. Since their lives have been completely different, lovers find themselves in a quagmire of misunderstandings, conflicts, and unforeseen obstacles. Although they break up, they unite only to certify their love and to live happily ever after. However, there are noticeable differences between the two movies as well; the roles are reversed. Whereas in 'Pretty Woman', it is the man who is rich and wealthy, the woman in 'Notting Hill' is the one with wealth and fame. While "Pretty woman" is a love story between a hooker and a businessman, an actress and a 35-year-old bookstore owner falls in love in "Notting Hill". Nevertheless, the two movies tell a tale which is hard to find in real life. Very much like 'Cinderalla', 'Notting Hill' and 'Pretty Woman' are modern day fairy tales about unlikely couples finding themselves in love. In this fourth version, the majority of the problems relating to the grammar have been successfully resolved. The major improvement, however, stems from the text-maker’s inspired decision to place the Orientation at the very beginning. This allows the reader to process the entire text coherently in terms of the Cinderella story. In other words, the text-maker has made a substantial improvement to the logico-rhetorical structure of the text. In this version too, the text-maker clearly understands the need to make main information truly main and subordinate information truly subordinate. This explains why the information about the year of release has been eliminated and why the entire text segment about Julia Roberts has been subordinated to the idea of the two movies as 23 improbable love stories. As a result, the text-maker has eliminated entirely the disconnected sequence “In first, … Besides, … Someone say that … Though entire structure is very similar,…” and substituted instead the well-formed sequence “"Notting Hill" and "Pretty Woman" are modern day versions of 'Cinderalla'… Both movies … The basic structure of the two movies… However,… Nevertheless,… Very much like 'Cinderalla',…”. As a result, the choices now correctly guide the reader through the main transitions in the overall “point” of the text. Of course, there are still occasional problems with the grammar (“as a starring role”), the lexical collocations (“hard to exist in the real world” and “certify their love”); and the logico-rhetorical structure (“there are noticeable differences”), but in other respects, this text clearly approaches the sophistication of a first language text-maker. Conclusion: Computers and Inefficiently Self-Regulating Learner Texts Situationally fine-tuned as they are, written texts, including the written texts of second language learners, are among the most difficult of human undertakings. They present immense problems for both the writer and for the analyst who wishes to say intelligent things about them. The situational fine-tuning of all genuinely creative texts is the reason that the investigation of ESL written texts must be extremely sensitive. The analysis presented here of the multiple textualizations of the same basic information should alert the researcher to the obvious limitations of simple word frequency counts in attempting to gauge second language development accurately. The future investigation of this recent field requires the active intelligence of the researcher performing qualitative analysis on the inefficiently self-regulating nature of a variety of learner texts. In other words, computer techniques may be used to facilitate certain tedious mental 24 operations, but number crunching and simple lexical frequency profiles are unlikely to reveal much that is of genuine interest about learner textualization. The analysis of any written text involves the active consideration of the text-maker’s motivation for the selection of language items at a variety of finely gradated levels. In the case of the learner text, motivation is primarily determined by the text-maker’s uneven awareness of the distinct but interlocking levels of grammar, lexical collocation, and logicorhetorical structure. 25 Bibliography Beaugrande, R. de. (1980). Text, discourse and process. New Jersey: Ablex Publishing Corporation. Beaugrande, R. de, & Dressler, W. (1981). Introduction to text linguistics. London and New York: Longman. Beaugrande, R. de. (1997). New foundations for a science of text and discourse. Norwood, NJ: Ablex Publishing Company. Clark, H. & Clark, E. (1978). Universals, relativity and language processing. In J. Greenberg (Ed.), Method and theory. (pp. 235-277) Universals of Human Language Vol. 1. Stanford: Stanford University Press. Coulthard, M. (1994). “On analyzing and evaluating written text.” In Advances in written text analysis. Ed. M. Coulthard. London and New York: Routledge: 1-11. Firth, J. R. (1968). Selected Papers of J. R. Firth, 1952-59. Ed. F. R. Palmer. London: Longman. Fries, P. (1994). “On theme, rheme and discourse goals.” In Advances in written text analysis. Ed. M. Coulthard. London and New York: Routledge: 229-249. Freedman, A., Pringle, I., & Yalden, J. (Eds.). (1979). Learning to write: First language/second language. London and New York: Longman. Granger, S., Hung, J. & Petch-Tyson, S. (Eds.). (2002). Computer learner corpora, second language acquisition, and foreign language teaching. Amsterdam and Philadelphia: John Benjamins Publishing Company. Granger, S. (2002). “A bird’s eye view of learner corpus research.” In Granger, S., Hung, J. & Petch-Tyson, S. Eds. Computer learner corpora, second language 26 acquisition, and foreign language teaching. Amsterdam and Philadelphia: John Benjamins Publishing Company. Greenberg, J. (1966). Language universals. The Hague: Mouton. Halliday, M., McIntosh, A. and Strevens, P. (1964). The linguistic sciences and language teaching. London: The English Language Book Society and Longman Group Ltd. Halliday, Michael. (1966). “Lexis as a linguistic level” in Bazzell, C.E., J.C. Catford, M.A.K. Halliday and R.H. Robins. Eds. In memory of J.R. Firth. London: Longman. Halliday, M. & Hasan, R. (1976). Cohesion in English. London and New York: Longman. Halliday, M. (1994a.). “The construction of knowledge and value in the grammar of scientific discourse, with reference to Charles Darwin’s The origin of species.” In Advances in written text analysis. Ed. M. Coulthard. London and New York: Routledge, 136-156. Halliday, M. (1994b). An introduction to functional grammar. 2nd Edition London: Edward Arnold. Halliday, M. (2004). The language of science. Vol 5. Collected Works. Ed. J. J. Webster. London and New York: Continuum. Hwang, Son-min. (2003). “Cohesion in interlanguage texts: An analysis of the use of synonyms and reference for achieving cohesion in English interlanguage texts”. Term paper for “Research Paper Writing”, Graduate English Program, Spring Semester, Yonsei University. Jo, Soo-jin. (2003). “Substitution in the interlanguage writing development: An analysis of the functions of substitution in an English interlanguage corpus”. Term paper for “Research Paper Writing”, Graduate English Program, Spring Semester, Yonsei University. 27 Iser, W. (1978). The act of reading: A theory of aesthetic response. Baltimore and London: The Johns Hopkins University Press. Kroll, B. (Ed.). (1990). Second language writing: Research insights for the classroom. Cambridge: Cambridge University Press. Laufer, B. & Nation, P. (1998). Vocabulary size and use: Lexical richness in L2 written production. Applied linguistics, 19.2: 225-254. Laufer, B. (2005) “Lexical frequency profiles: From Monte Carlo to the real world: A response to Meara.” Applied Linguistics 26.2: 582-588. Meara, P. (2005) “Lexical frequency profiles: A Monte Carlo analysis.” Applied Linguistics, 26.1: 32-47. Murphy, T. (2001). “The emergence of texture: An analysis of the functions of the nominal demonstratives in an English language interlanguage corpus.” Language Learning and Technology 5.3: 152-173. Murphy, T. P. (2004). “James Joyce and narrative territory: The distinct function of lost time in “An Encounter” and “The Sisters”, Journal of Literary Semantics 33.2 (Fall): 131-154. Murphy, T. P. (2005a). “The uncertainties of conversational exchange: Dialogue monitoring as a function of the narrative voice”, Style 39.4: 396-411. Murphy, T. P. (2005b). “Interpreting marked order narration: The case of James Joyce’s ‘Eveline’”, Journal of Literary Semantics 34.2: 107-124. Reid, J. (1992). “A computer text analysis of four cohesion devices in English discourse by native and nonnative writers”. Journal of second language writing 1.2: 79-107. Rohde, D. (2003) “Another attack in central Iraq kills another U.S. soldier.” New York Times June 6. http://www.nytimes.com/2003/06/06/international/worldspecial/06FIGH.html 28 Shaw, P. & Liu, E. (1998). “What develops in the development of second-language writing?” Applied linguistics, 19.2: 225-254. Sinclair, John. (1991). Corpus, concordance, collocation. Oxford: Oxford University Press. Wolfe-Quintero, K., Inagaki, S. and Kim, H-K. (1998). Second Language Development in Writing: Measures of Fluency, Accuracy and Complexity. Technical Report 17. Second Language Teaching and Curriculum Center: University of Hawai‘i at Manoa. Wray, Alison. Formulaic Language and the Lexicon. Cambridge: Cambridge University Press, 2002. Yoon, Hyun-sook and Hirvela, A. (2004). “ESL student attitudes towards corpus use in L2 writing” Journal of second language writing 13: 257-283. 29 Keywords 1. Comment. A sentence that picks up the Given information of the previous sentence and elaborates it in some way. 2. Emergent Texture. In Murphy (2001), the concept of emergent texture is defined as “the manner in which interlanguage texts gradually extend their use and control of the grammatical means used to establish lexical and textual cohesion” (154). The more refined definition utilized in this chapter is that emergent texture refers to a given text’s inefficient utilization of the full set of grammatical, lexical, and logico-rhetorical ties. In this sense, an analysis of the state of the emergent texture of a second language text is one measure of its distance from a reconfigured first language textualization of the same information. 3. Inefficiently Self-Regulating System. A text is an inefficiently self-regulating system when its grammatical, lexical and logico-rhetorical ties are improperly configured. As a result of this, the textual directives lead the reader into the discovery of ambiguities, discrepancies, paradoxes, contradictions and redundancies. While the reader’s first language knowledge may sometimes be enough to recover the intended meaning, some of these misused or missing directives will result in the reader’s failure to understand (portions of) the text. All poorly written texts, including the specific subset of second language texts, may be regarded as inefficiently self-regulating systems. 4. Logico-Rhetorical Structure. In functional grammar, the logico-rhetorical structuring of a text refers to the intermediate range of choice situated between the grammar and the lexis. In conjunction with the punctuation, it is this intermediate range of choice that is used when configuring the text’s logico-rhetorical structure. The choice of different logico-rhetorical structuring allows the same textual information to be presented in a definite (but not unlimited) range of alternative textualizations. In the terms of Robert de Beaugrande, the logico-rhetorical structure of the text is a form of procedural knowledge. In well-organized texts, this procedural knowledge is “formatted as programs designed to run in specifically anticipated ways” (1980 65). One example of a program designed to run in an anticipated way is the Situation-Problem-SolutionEvaluation structure. It follows that an inefficient text runs in unanticipated or 30 unpredictable ways. 5. Markedness. According to Roman Jakobson, “the general meaning of a marked category states the presence of a certain property A; the general meaning of the corresponding unmarked category states nothing about the presence of A and is used chiefly but not exclusively to indicate the absence of A (quoted in Greenberg, 25). For example, in some environments, actor is to actress as “male thespian” is to “female thespian”. However, in other environments, actress is neutralized by the term actor because actress can only refer to female thespians. In addition, actress is morphologically the more complex of the two terms, requiring the addition of an extra morpheme. For this reason, within the terms of the unmarked/marked distinction, actor is unmarked, whereas actress is marked (Clark and Clark 231; Greenberg 26). In narrative fictions, an extremely important distinction may be made between chronologically ordinary narratives such as romances and marked order narratives such as detective fictions and Gothic horror stories. There is also the important secondary distinction between the marked character and the other characters, who are all unmarked (Murphy 2004; 2005b). 6. Move. A sentence that picks up the New information of the previous sentence and extends it in some way. 7. Self-Regulating System. First proposed by Wolfgang Iser in The Act of Reading (1978), the concept of the text as a self-regulating cybernetic system was more precisely formulated by Robert de Beaugrande in Text, Discourse and Process (1980). According to Beaugrande, “The stability of the text as a cybernetic system … is characterized by its connectivities, i.e. unbroken access among the occurring elements of the participating language systems”. In other words, a text will contain “sequential connectivity of grammatical dependencies in the surface text”, “conceptual connectivity” and “planning connectivity” (17). In a study of Charles Dickens’ Oliver Twist, Murphy (2005a) extended the concept of the self-regulating textual system to the nineteenth century novel by explaining how a reader might process the discrepancies discovered in the clash of directly quoted character speech. These discrepancies are resolved by means of the conversation monitoring of the narrative voice. 8. Texture. In Halliday and Hasan’s Cohesion in English (1976), “a text has texture, and this is what distinguishes it from something that is not a text. It derives this texture from 31 that fact that it functions as a unity with respect to its environment” (2). The texture of any text is constituted by the five major cohesive ties: those of reference, substitution, ellipsis, conjunction and lexis. Index Reference List Indexer Reference List for: Handbook of Research on Computer Enhanced Language Acquisition and Learning Editor: Beth Barber and Felicia Zhang Chapter Title: The Texture of Inefficiently Self-Regulating ESL Systems Author: Terence Patrick Murphy Term 1: Lexis Also known as: word choice, vocabulary Similar to: lexical richness, lexical sophistication Associated in the manuscript with: grammar, logical-rhetorical structure, textual development Notable appearances of this term can be found on: Page 3: Argument against an exclusive focus on lexis in understanding ESL development. Page 4-5: The different range of choice possible at the level of grammar and lexis. Page 8: The idea that ESL texts are lexically inefficient is advanced. Term 2: Grammar Also known as: Similar to: cohesion Associated in the manuscript with: lexis, logico-rhetorical structure, texture Notable appearances of this term can be found on: Page 5: Grammar involved in the formation of texture as a more sensitive indicator of ESL development than lexis. Page 5: The different kinds of difficulty posed to the ESL text-maker by grammar, lexis, and logico-rhetorical structure. Term 3: Logico-rhetorical structure Also known as: argument structure, semantic structure Similar to: coherence Associated in the manuscript with: lexis, logico-rhetorical structure, coherence 32 Notable appearances of this term can be found on: Page 5: The different kinds of difficulty posed to the ESL text-maker by grammar, lexis, and logico-rhetorical structure. Page 10: Logico-rhetorical structure as the work of the punctuation and word items distinct from both grammar and lexis. Term 5: Written text Also known as: composition, text Similar to: Associated in the manuscript with: well-written texts, poorly written texts, the dialectic of Moves and Comments Notable appearances of this term can be found on: Page 11: Against the idea of the written text as a system of fixed sentence choices. Page 11: In support of the idea of the written text as a potential system for optimal information clausing. Term 6: Move Also known as: Similar to: New information Associated in the manuscript with: New information, Comments Notable appearances of this term can be found on: Page 11: Definition of Move sentence. Term 7: Comment Also known as: Similar to: Given information Associated in the manuscript with: Notable appearances of this term can be found on: Page 11: definition of Comment sentence. Term 8: Cohesion Also known as: Similar to: grammar Associated in the manuscript with: texture Notable appearances of this term can be found on: Page 5: Definition of the first four sets of cohesive ties. 33 Page 7: Inefficient cohesion associated with the choice of unmarked language items. Page 8. Definition of the final set of cohesive ties. 34