Textual Patterns: Key words and corpus analysis in language education
Mike Scott
University of Liverpool
Christopher Tribble
King's College, London University
John Benjamins Publishing Company
Amsterdam l Philadelphia
This volume is divided in to two major sections, the first being more resource and theory-based, the second largely applied to a set of distinct areas of knowledge.In the first, in Chapters I to 5, Mike Scott introduces and explains software resources which are relevant to those undertaking corpus-based research or teaching preparation, and some problematic aspects of the underlying theories.These are illustrated with a series of studies. In Part 2 Chris Tribble has written a series of research articles which provide further examples of how language teachers and researchers can draw on these approaches to develop pedagogically valuable insights into language in use.
Scott's section starts off with a chapter which does not involve any studies but quest ions and theories. Chapter 1 therefore considers some of the characteristics of corpus-based research, in terms of a quest for underlying pattern .Throughout the book we are identifying patterns in text . There are also four different starting-points for the corpus-based researcher, the text, the language,the culture, the brain. Finally there is the question of scope: what sort of context are we dealing with? The amount of text within the researcher 's scope can go from the smallest context up to the whole text and beyond; Chapter 3,for example, mostly operates at a narrow scope, while Chapter 5, at the other extreme, considers pattern linkages across groups of texts .
Chapter 2 is about word-lists. Through transformation of a text into a word-list, and selection of the words which go in to it, further questions are raised concerning the nature of the category "word" and "text". The issue of closed-set and open-set lexical items is linked here to word frequency, so we study the nature of high-, medium- and low-frequency items in a word- list.The chapter ends with a study which shows that word-lists are not randomly gene rated oddities but instead obey underlying laws.
The overall theme of Chapter 3 is concordancing, but first the nature of co-occurrence has to be clarified . words are like people, their relationships go beyond their neighbourhood and there can be one-way or two-way attraction between them. It is useful to distinguish between mental and textual co-occurrence and so a study is carried out to find out whether word associations match up with textual linkage. Scott also looks at further patterns which can be revealed through resources such as clusters and plots, and ends the chapter with a further study considering the potential linkages of a given word with its sentence-, paragraph- and text-positions.
The next chapter deals with key words (KWs), considering the nature of keyness and how this is dependent on repetition. Identifying keyness, in turn,is dependent on the use of filters and statistical tests, so the characteristics of a reference corpus are discussed too. There follow four studies, which concern first the key words of Romeo and Juliet, then the effect of varying the choice of reference corpus. Next comes the issue of plotting where the key words occur in texts and a distinction between global and local key words dispersion pattern s,which leads to a study of how key words are inter linked. The fourth study deals with the kinds of words which tend to be key; a lot of them seem to be nouns but then a lot of words in texts are nouns too, so the aim is to find out whether keywords are more likely than other words to be nouns.
Chapter 5 goes beyond the scope of the preceding one to consider the formal patterns of linkage and to analyse these patterns as they obtain between texts, not just within them. There follow two studies, the first of which aims to pinpoint the characteristics of "key key words'; to see whether these share some of the qualities of wordlists identified in Chapter 2, and in particular to find out whether key key words reflect aboutness. The study is done first with a general corpus and the n with corpora differentiated by domain. The final study looks at associates, defining these and studying the academic and non-academic associates of key key words in the Humanities, in Law and in Technology.
Chapter 6 introduces the reader to a simple corpus analysis methodology(drawing on word -lists and key word lists ) which can be used to find out the main lexico-gramrnatical features of a collections of similar texts. Using microcorpora extracted from the British National Corpus, Tribble shows how pedagogically important features of written and spoken production can be identified. The chapter makes explicit many of the procedures which will be used in those which follow and offers practical guidelines on the preparation and subsequent analysis of text data using simple Excel spreadsheets.
In Chapter 7 a small collect ion of business correspondence provides the basis for a consideration of how reader writer relationships are constructed in professional settings . An important feature of this chapter is the way in which Tribble shows how one can combine genre analysis with corpus and K\V analysis to build comprehensive accounts of language use in specific contexts. While a genre analytic reading of the texts allows the identification of a set of relationship s between readers and writer s along axes of distance and proximity,as well as the development of a move analysis which accounts for the major structural components of specific categories of correspondence, K\V and Plot analysis make it possible to identify the linguistic exponents which are used to realise these relationships.
Chapter 8 draws on a different kind of analysis to investigate contrasts between expert and apprentice performances in academic writing. In this case the focus is on what Scott calls clusters, otherwise known as lexical bundles in Biber and Conrad (1999 ). For Biber and Conrad, lexical bundles are four word combinations which occur with a threshold frequency per million, although for Scott, clusters are any multi-word combination which occurs within frequency thresholds set by the researcher. The interest is that by studying how clusters are used in contrasting text collections, it is possible to isolate the prefabricated units which different categories of writers depend on for building arguments, offering examples, referring and the like. The study in Chapter 8shows how the performances of advanced apprentice undergraduate writers inliterary studies (Polish 11A English Philology students ) contrast with those of authors of published research articles in the same field.
Chapter 9 demonstrates the use of KW analysis in a diachronic study. Taking a large collection of texts from the Guardian Weekly, Tribble shows how K\Vs can give in sights not only into the "aboutness" of texts, but also into the ways in which news reporters treat issues. In a series of studies, the chapter shows how KW analysis is able to give in sights into how contrasting political fortunes of Messrs. Blair, Clinton and Milosevic were reported; how certain topics are consistently linked with the negative prefix anti or the more positive prefix pro; how gender bias in news reporting can be tracked through the analysis of personal pronouns and titles; an d how the key themes for each year of the Guardian Weekly can be identified.
In complete contrast to this large scale study of many thousands of individual texts, Chapter 10 demonstrates how corpus too ls can be used in the study of a single, very short, short story. The analysis of a short story by the 20thCentury Irish writer Samuel Beckett works at two levels. In th e first part of the chapter, Tribb le shows how a simple statistical analysis of sentence length combined with a study of th e collocations of high frequency items can reveal the structure of a narrative text. In the second part, th is time using the concordancerto identify pattern ing around clause boundaries, Tribble is able to explicate the contadictory and unreliable nature of the text, and the tension between the reader's desire to make sense of things and the ultimate uncertaintyof this unreliable narrator.
All of these chapters are linked by what is the recurring theme of this book,that an attention to texts is the best starting point for a corpus informed language pedagogy. Rather than seeing corpus data as entirely abstracted from its linguistic and social context, the studies stress the obligation on the researcher to re-connect with the text (and, where possible, with its context of production) in order to build accounts of language in use which will have value for teachers and students of language alike.
Table of contents
Preface
Par t I
CHAPTER 1
Texts in language study and language education
CHAPTER 2
Word-lists: Approaching texts
CHAPTER 3
Concordances : The immedi ate context
CHAPTER 4
Key wo rds of individual texts: Aboutness and style
CHAPTER 5
Key words and genres
Part II
CHAPTER 6
General English language teaching: Grammar and lexis in spoken and written texts
CHAPTER 7
Business and professional com mu nicatio n: Managing relationships in professional writing
CHAPTER 8
English for academic purposes: Build ing an accou nt of expert and apprentice performances in literary criticism
CHAPTER 9
What counts in current journalism: Keywords in newspaper reporting
CHAPTER 10
Counting things in texts you can't count on: A study of Samuel Beckett's Texts for Nothing,
References
Name index
Subject index
Mike Scott
University of Liverpool
Christopher Tribble
King's College, London University
John Benjamins Publishing Company
Amsterdam l Philadelphia
This volume is divided in to two major sections, the first being more resource and theory-based, the second largely applied to a set of distinct areas of knowledge.In the first, in Chapters I to 5, Mike Scott introduces and explains software resources which are relevant to those undertaking corpus-based research or teaching preparation, and some problematic aspects of the underlying theories.These are illustrated with a series of studies. In Part 2 Chris Tribble has written a series of research articles which provide further examples of how language teachers and researchers can draw on these approaches to develop pedagogically valuable insights into language in use.
Scott's section starts off with a chapter which does not involve any studies but quest ions and theories. Chapter 1 therefore considers some of the characteristics of corpus-based research, in terms of a quest for underlying pattern .Throughout the book we are identifying patterns in text . There are also four different starting-points for the corpus-based researcher, the text, the language,the culture, the brain. Finally there is the question of scope: what sort of context are we dealing with? The amount of text within the researcher 's scope can go from the smallest context up to the whole text and beyond; Chapter 3,for example, mostly operates at a narrow scope, while Chapter 5, at the other extreme, considers pattern linkages across groups of texts .
Chapter 2 is about word-lists. Through transformation of a text into a word-list, and selection of the words which go in to it, further questions are raised concerning the nature of the category "word" and "text". The issue of closed-set and open-set lexical items is linked here to word frequency, so we study the nature of high-, medium- and low-frequency items in a word- list.The chapter ends with a study which shows that word-lists are not randomly gene rated oddities but instead obey underlying laws.
The overall theme of Chapter 3 is concordancing, but first the nature of co-occurrence has to be clarified . words are like people, their relationships go beyond their neighbourhood and there can be one-way or two-way attraction between them. It is useful to distinguish between mental and textual co-occurrence and so a study is carried out to find out whether word associations match up with textual linkage. Scott also looks at further patterns which can be revealed through resources such as clusters and plots, and ends the chapter with a further study considering the potential linkages of a given word with its sentence-, paragraph- and text-positions.
The next chapter deals with key words (KWs), considering the nature of keyness and how this is dependent on repetition. Identifying keyness, in turn,is dependent on the use of filters and statistical tests, so the characteristics of a reference corpus are discussed too. There follow four studies, which concern first the key words of Romeo and Juliet, then the effect of varying the choice of reference corpus. Next comes the issue of plotting where the key words occur in texts and a distinction between global and local key words dispersion pattern s,which leads to a study of how key words are inter linked. The fourth study deals with the kinds of words which tend to be key; a lot of them seem to be nouns but then a lot of words in texts are nouns too, so the aim is to find out whether keywords are more likely than other words to be nouns.
Chapter 5 goes beyond the scope of the preceding one to consider the formal patterns of linkage and to analyse these patterns as they obtain between texts, not just within them. There follow two studies, the first of which aims to pinpoint the characteristics of "key key words'; to see whether these share some of the qualities of wordlists identified in Chapter 2, and in particular to find out whether key key words reflect aboutness. The study is done first with a general corpus and the n with corpora differentiated by domain. The final study looks at associates, defining these and studying the academic and non-academic associates of key key words in the Humanities, in Law and in Technology.
Chapter 6 introduces the reader to a simple corpus analysis methodology(drawing on word -lists and key word lists ) which can be used to find out the main lexico-gramrnatical features of a collections of similar texts. Using microcorpora extracted from the British National Corpus, Tribble shows how pedagogically important features of written and spoken production can be identified. The chapter makes explicit many of the procedures which will be used in those which follow and offers practical guidelines on the preparation and subsequent analysis of text data using simple Excel spreadsheets.
In Chapter 7 a small collect ion of business correspondence provides the basis for a consideration of how reader writer relationships are constructed in professional settings . An important feature of this chapter is the way in which Tribble shows how one can combine genre analysis with corpus and K\V analysis to build comprehensive accounts of language use in specific contexts. While a genre analytic reading of the texts allows the identification of a set of relationship s between readers and writer s along axes of distance and proximity,as well as the development of a move analysis which accounts for the major structural components of specific categories of correspondence, K\V and Plot analysis make it possible to identify the linguistic exponents which are used to realise these relationships.
Chapter 8 draws on a different kind of analysis to investigate contrasts between expert and apprentice performances in academic writing. In this case the focus is on what Scott calls clusters, otherwise known as lexical bundles in Biber and Conrad (1999 ). For Biber and Conrad, lexical bundles are four word combinations which occur with a threshold frequency per million, although for Scott, clusters are any multi-word combination which occurs within frequency thresholds set by the researcher. The interest is that by studying how clusters are used in contrasting text collections, it is possible to isolate the prefabricated units which different categories of writers depend on for building arguments, offering examples, referring and the like. The study in Chapter 8shows how the performances of advanced apprentice undergraduate writers inliterary studies (Polish 11A English Philology students ) contrast with those of authors of published research articles in the same field.
Chapter 9 demonstrates the use of KW analysis in a diachronic study. Taking a large collection of texts from the Guardian Weekly, Tribble shows how K\Vs can give in sights not only into the "aboutness" of texts, but also into the ways in which news reporters treat issues. In a series of studies, the chapter shows how KW analysis is able to give in sights into how contrasting political fortunes of Messrs. Blair, Clinton and Milosevic were reported; how certain topics are consistently linked with the negative prefix anti or the more positive prefix pro; how gender bias in news reporting can be tracked through the analysis of personal pronouns and titles; an d how the key themes for each year of the Guardian Weekly can be identified.
In complete contrast to this large scale study of many thousands of individual texts, Chapter 10 demonstrates how corpus too ls can be used in the study of a single, very short, short story. The analysis of a short story by the 20thCentury Irish writer Samuel Beckett works at two levels. In th e first part of the chapter, Tribb le shows how a simple statistical analysis of sentence length combined with a study of th e collocations of high frequency items can reveal the structure of a narrative text. In the second part, th is time using the concordancerto identify pattern ing around clause boundaries, Tribble is able to explicate the contadictory and unreliable nature of the text, and the tension between the reader's desire to make sense of things and the ultimate uncertaintyof this unreliable narrator.
All of these chapters are linked by what is the recurring theme of this book,that an attention to texts is the best starting point for a corpus informed language pedagogy. Rather than seeing corpus data as entirely abstracted from its linguistic and social context, the studies stress the obligation on the researcher to re-connect with the text (and, where possible, with its context of production) in order to build accounts of language in use which will have value for teachers and students of language alike.
Table of contents
Preface
Par t I
CHAPTER 1
Texts in language study and language education
CHAPTER 2
Word-lists: Approaching texts
CHAPTER 3
Concordances : The immedi ate context
CHAPTER 4
Key wo rds of individual texts: Aboutness and style
CHAPTER 5
Key words and genres
Part II
CHAPTER 6
General English language teaching: Grammar and lexis in spoken and written texts
CHAPTER 7
Business and professional com mu nicatio n: Managing relationships in professional writing
CHAPTER 8
English for academic purposes: Build ing an accou nt of expert and apprentice performances in literary criticism
CHAPTER 9
What counts in current journalism: Keywords in newspaper reporting
CHAPTER 10
Counting things in texts you can't count on: A study of Samuel Beckett's Texts for Nothing,
References
Name index
Subject index
