Move Analysis: A Corpus-Based Approach
Discourse on the Move: Corpus Analysis and Discourse Structure
Introduction to Move Analysis
This chapter introduces move analysis, a top-down approach to corpus-based discourse analysis. It details what move analysis is, its applications, and how to conduct one using corpora.
Background
Move analysis, developed by Swales (1981), examines the rhetorical organization of texts, specifically research articles, based on communicative purposes. A move is a text section with a specific communicative function contributing to the genre's overall purpose. This framework allows researchers to dissect complex texts into manageable, purposeful segments, offering insights into how writers achieve specific communicative goals.
The purposes define the genre and influence content and style, leading to structural, stylistic, and content similarities among texts within a genre (Swales, 1990, p. 58). Genre, in this context, can be understood as a conventionalized form of communication that responds to recurrent social needs or purposes. These conventions guide both the production and interpretation of texts.
Genre analysis grew in the 1970s and 1980s, focusing on discourse organization. Bhatia (2004) notes a shift from lexico-grammatical features to structural concerns, which were further linked to communicative functions. This evolution marked a significant move from merely describing language elements to understanding how these elements function within a broader communicative context.
Expert members of a discourse community recognize genre purposes, shaping conventions. Research papers exemplify this, as novice writers are taught paper-writing conventions. However, some genres, like fundraising letters, have weaker community influence, allowing for deviation from norms. The strength of community influence impacts the rigidity of genre conventions. In highly regulated genres, deviations can lead to miscommunication or failure to achieve the intended purpose.
Move analysis describes texts as a series of functional units (moves) fulfilling the genre's communicative purpose (Connor, Davis, & De Rycker, 1995). Moves vary in length but contain at least one proposition (Connor & Mauranen, 1999). Conventional moves occur frequently, while optional ones are less frequent. Moves include ‘steps’ (Swales, 1990) or ‘strategies’ (Bhatia, 1993a) to achieve the move's purpose. These steps or strategies are specific linguistic or rhetorical choices made by the writer to effectively realize the communicative goal of the move.
Swales’ Move Analysis of Research Articles
Swales (1981) developed move analysis for English for Specific Purposes (ESP), revised in Swales (1990), to aid non-native English speakers (NNSs) in reading and writing research articles. His analysis led to a series of moves defining the rhetorical structure of research article introductions. The model helps NNSs navigate the complexities of academic writing by providing a clear framework for understanding and producing research articles.
The CARS Model
Swales’ three-move schema, known as the Create a Research Space (CARS) model, is described in Table 2.1. It specifies the preferred sequences of moves and steps. The CARS model is a foundational concept in move analysis, providing a structured approach to understanding the introductory sections of research articles. The model suggests that effective introductions typically establish a context, identify a gap or problem, and then present the current research as a means to address that gap.
Table 2.1 CARS model for research article introductions, adapted from Swales (1990, p. 141)
Move 1: Establishing a territory
Step 1: Claiming centrality and/or
Step 2: Making topic generalization(s) and/or
Step 3: Reviewing items of previous research
Move 2: Establishing a niche
Step 1A: Counter-claiming or
Step 1B: Indicating a gap or
Step 1C: Question raising or
Step 1D: Continuing a tradition
Move 3: Occupying the niche
Step 1A: Outlining purposes or
Step 1B: Announcing present research
Step 2: Announcing principal findings
Step 3: Indicating RA structure
Move 1 introduces the general research topic. It provides background information and establishes the importance of the research area. This move aims to contextualize the study within the broader field.
Move 2 identifies specific areas needing investigation. By pointing out gaps in existing research, raising questions, or challenging previous claims, Move 2 justifies the need for the current study. Identifying a niche is crucial for positioning the research as relevant and necessary.
Move 3 introduces the current study in the context of prior research. Move 3 bridges the gap between the established context and the present study. It highlights how the research contributes to addressing the identified niche.
Steps within Move 1: Establishing a Territory
Move 1 has up to three steps. Step 1, Claiming centrality, asserts the importance of the research:
Example: “The study of…has become an important aspect of…” (Swales, 1990, p. 144). Claiming centrality emphasizes the relevance and significance of the research topic, capturing the reader's attention and highlighting the need for investigation.
Step 2, Making topic generalizations, offers neutral statements about knowledge or phenomena:
Example: “The aetiology and pathology …is well known.” (Swales, 1990, p. 146). Making topic generalizations provides a broad overview of the subject matter, setting the stage for more specific discussions and establishing a common ground with the audience.
Step 3, Reviewing items of previous literature, reviews relevant past research:
Example: “X Was found by Sang et al. (1972) to be impaired.” (Swales, 1990, p. 150). Reviewing previous literature situates the current study within the existing body of knowledge, demonstrating the researcher's familiarity with the field and highlighting how the new research builds upon or diverges from prior work.
Move 2: Establishing a Niche
Move 2 connects Move 1 to Move 3, articulating the need for current research, with four options:
Step 1A: Counter claiming. Challenging or questioning existing research findings to create space for new perspectives.
Step 1B: Indicating a gap. Identifying areas where current research is lacking or incomplete, thus justifying the need for further investigation.
Step 1C: Question raising. Posing questions that have not yet been answered by previous research, stimulating curiosity and setting the stage for the present study.
Step 1D: Continuing a tradition. Building upon established lines of inquiry, demonstrating how the current research contributes to an ongoing scholarly conversation.
Move 3: Occupying the Niche
Move 3 is distinct, with authors actively presenting their research:
Step 1A: Stating research purpose(s). Clearly articulating the goals and objectives of the study, providing a roadmap for the reader.
Step 1B: Describing main features of the research. Outlining the key aspects of the study, such as the methodology, data sources, and analytical techniques.
Step 2: Announcing the principal findings. Highlighting the main results and discoveries of the research, offering a preview of the study's contributions.
Step 3: Indicating the research article structure. Providing a brief overview of how the article is organized, helping the reader navigate the content.
Revisions to the CARS Model
The CARS model has been validated and studied, demonstrating recursion (Swales, 1990, p. 140), where moves or steps recur. Recursion involves the repetition of moves or steps within the introduction, allowing writers to reinforce key points or provide further elaboration. This adds depth and complexity to the rhetorical structure.
Bunton (2002) found that Ph.D. thesis introductions include ‘Defining terms’ in Move 1, unlike research articles. The inclusion of definitional elements reflects the need to establish a clear and precise understanding of key concepts for the thesis's specific audience.
Swales (2004) revised his model to reflect variability across research fields as seen in Table 2.2. The original “one size fits all” perspective evolved into a discipline-specific understanding. This revision acknowledges that different academic disciplines may have unique conventions and expectations for research article introductions.
Table 2.2 Swales’ revised model for research article Introductions (2004, pp. 230, 232)
Move 1: Establishing a territory (citations required) via Topic generalizations of increasing specificity
Move 2: Establishing a niche (citations possible) via:
Step 1A: Indicating a gap, or
Step 1B: Adding to what is known
Step 2: Presenting positive justification (optional)
Move 3: Presenting the present work via:
Step 1: Announcing present research descriptively and/or purposively (obligatory)
Step 2: Presenting research questions or hypotheses* (optional)
Step 3: Definitional clarifications* (optional)
Step 4: Summarizing methods* (optional)
Step 5: Announcing principal outcomes (optional)**
Step 6: Stating the value of the present research (optional)**
Step 7: Outlining the structure of the paper (optional)**
Related genres share move types, but each has unique structural characteristics reflecting specific communicative functions. While different genres may employ similar rhetorical moves, the specific ways these moves are realized and organized can vary significantly depending on the genre's purpose and audience.
Move Analysis Across Genres
Move analysis extends to English for Business and Technology (Bhatia, 1993a, 1997a) and English for Professional Communication (Flowerdew, 1993). It has been applied to biochemistry (Kanoksilapatham, 2005; D. Thompson, 1993), biology (Samraj, 2002), computer science (Posteguillo, 1999), and medicine (Nwogu, 1997; Williams, 1999), university lectures (S. Thompson, 1994), dissertations (Hopkins & Dudley-Evans, 1988), and textbooks (Nwogu, 1991). The broad applicability of move analysis across diverse fields underscores its versatility as a framework for analyzing and understanding textual structures.
Within research articles, studies focus on specific sections like Introductions (Crookes, 1986), Methods (Wood, 1982), Results (D. Thompson, 1993; Williams, 1999), and Discussion (Peng, 1987). Some explore moves across multiple sections (Posteguillo, 1999; Nwogu, 1997; Kanoksilapatham, 2005). This targeted approach allows researchers to gain in-depth insights into the rhetorical strategies employed in different parts of a research article.
Recently, professional discourse has been examined, including legal discourse (Bhatia, 1993b), philanthropic discourse (Upton, 2002; Upton & Connor, 2001; Connor, 2000; Connor & Mauranen, 1999; Connor & Upton, 2004a), and movie reviews (Pang, 2002). The expansion of move analysis into professional contexts highlights its relevance for understanding communication practices in various real-world settings.
Example: Job Application Letters
A move analysis of job application letters (Connor, Precht, & Upton, 2002) illustrates differing genre move types.
The letters from the Indianapolis Business Learner Corpus (IBLC) were written by business students:
Move 1: Identify the source of information.
Move 2: Apply for the position.
Move 3: Provide arguments for the job application.
Step 1: Implicit arguments with neutral info.
Step 2: Arguments based on benefit to the company.
Step 3: Arguments based on benefit to the applicant.
Move 4: Indicate desire for contact.
Move 5: Express pleasantries.
Move 6: Offer more information.
Move 7: Reference resume.
The key difference between research article introductions (3 moves) and job application letters (7 moves) is notable because introductions are lengthier. This contrast illustrates how genre conventions shape the structure and content of different types of texts.
Moves are identified by communicative purpose and vary in length. The length and complexity of moves can depend on the specific goals and context of the communication.
Genres differ in move structure complexity; some moves have multiple steps, others do not. The presence or absence of steps within a move reflects the varying levels of detail and elaboration required by different genres.
Some move types are common (obligatory), others optional (Lewin, Fine, and Young (2001); Bhatia, 1993a). Obligatory moves are those that are essential for fulfilling the genre's communicative purpose, while optional moves provide additional information or nuance.
Move elements may or may not appear regularly. The regularity with which move elements appear can indicate their significance and conventionality within a particular genre.
Moves can recur cyclically (Swales, 2004) or be interrupted (Upton, 2002), mainly in less constrained genres. Cyclic recurrence and interruption of moves add flexibility to the structure of texts, especially in genres with less rigid conventions.
Summary of previous research on move analysis
Genres have definable, predictable moves. The identification of these moves aids in understanding and producing effective communication within those genres.
Article introductions have three rhetorical moves: establishing territory, establishing a niche, and occupying the niche. These moves provide a structured framework for introducing research and situating it within the existing scholarly landscape.
Letters of application also have distinguishable move types. These move types enable job seekers to present their qualifications and express their interest in a clear, persuasive manner.
Bhatia (1993a) argues move structure is the genre's property controlled by communicative purpose. The communicative purpose of a genre shapes its move structure, influencing how information is presented and organized.
Moves are building blocks for teaching novice writers (Dudley-Evans, 1995). By understanding move structures, novice writers can develop the skills to produce effective and appropriate texts in various genres.
Methods for Move Analysis
Kwan (2006) introduces functional-semantic methods for identifying moves. These methods emphasize the importance of understanding both the function and meaning of text segments in identifying moves.
This requires cognitive judgment to identify text intention and boundaries (Bhatia, 1993a; Paltridge, 1994). Identifying moves requires careful cognitive analysis to determine the writer's intention and to delineate the boundaries between different moves.
General Steps
There are no strict 'rules,’ but common steps include:
Understanding overall rhetorical purpose. Grasping the overarching goal of the text is essential for identifying the specific moves that contribute to that goal.
Evaluating each segment's function. Analyzing the role each segment plays within the text helps to identify its communicative purpose and potential move type.
Identifying common functional/semantic themes and grouping them into steps. Grouping segments with similar functions and meanings helps to reveal underlying move structures and patterns.
Swales’ CARS model includes establishing a territory because introductions often begin with context. Providing context is a common strategy for engaging the reader and situating the research within a broader framework.
Coding begins with a pilot, involving at least two coders. Conducting a pilot coding session helps to refine the coding protocol and ensure consistency among coders.
Initial analyses are discussed until there's agreement on purposes realized by segments, creating a protocol. Establishing a clear coding protocol is crucial for ensuring inter-rater reliability and the accuracy of the move analysis.
For corpus-based analysis, the protocol is applied to all texts. Applying the coding protocol consistently across the entire corpus ensures that the analysis is systematic and comprehensive.
Inter-rater reliability is checked. Assessing inter-rater reliability is essential for validating the coding process and ensuring that the identified moves are robust and reliable.
Discrepancies are resolved and texts re-coded. Addressing discrepancies and re-coding texts helps to improve the accuracy and consistency of the move analysis.
Additional steps/moves may emerge during analysis. Remaining open to the possibility of identifying new steps or moves during the analysis allows for a more nuanced and comprehensive understanding of the text.
Fundraising discourse may have optional move elements and unfixed order. The flexible nature of fundraising discourse highlights the importance of adapting move analysis techniques to suit different genres and contexts.
The typical process is summarized in Table 2.3.
Table 2.3 General steps often used to conduct a corpus-based move analysis
Determine rhetorical purposes of the genre
Determine rhetorical function of each text segment in its local context; identify the possible move types of the genre
Group functional and/or semantic themes that are either in relative proximity to each other or often occur in similar locations in representative texts. These reflect the specific steps that can be used to realize a broader move.
Conduct pilot-coding to test and fine-tune definitions of move purposes.
Develop coding protocol with clear definitions and examples of move types and steps.
Code full set of texts, with inter-rater reliability check to confirm that there is clear understanding of move definitions and how moves/steps are realized in texts.
Add any additional steps and/or moves that are revealed in the full analysis.
Revise coding protocol to resolve any discrepancies revealed by the inter-rater reliability check or by newly ‘discovered’ moves/steps, and re-code problematic areas.
Conduct linguistic analysis of move features and/or other corpus-facilitated analyses.
Describe corpus of texts in terms of typical and alternate move structures and linguistic characteristics
Inter-Rater Reliability
Top-down analysis involves human judgments to code discourse components. Human judgment is crucial in move analysis, as it enables researchers to interpret the nuances of language and identify the underlying communicative purposes of text segments.
A detailed coding rubric explicitly defines discourse components. A well-defined coding rubric ensures that all raters have a clear understanding of the move types and their characteristics.
Inter-rater reliability determines if raters understand the coding definitions consistently. Assessing inter-rater reliability is essential for validating the coding process and ensuring that the identified moves are robust and reliable.
Percent agreement measures agreements per coding decisions but doesn't account for chance. While percent agreement is a simple measure, it does not account for the possibility that raters may agree by chance, rather than due to a shared understanding of the coding definitions.
Cohen’s kappa (k) is a chance-corrected measure. Cohen's kappa provides a more accurate measure of inter-rater reliability by accounting for the level of agreement that would be expected by chance.
Training improves inter-rater reliability. Providing training to raters helps to ensure that they have a clear understanding of the coding definitions and are able to apply them consistently.
Inter-rater reliability measures consistency, not objectivity or validity. While inter-rater reliability indicates the consistency of coding, it does not guarantee that the coding is objective or valid. It is possible for raters to consistently apply an incorrect coding definition.
Variation among raters is expected. Some degree of variation among raters is inevitable, as different individuals may interpret text segments in slightly different ways.
An inter-rater reliability of 0.80 is deemed to be acceptable. An inter-rater reliability of 0.80 or higher is generally considered to be acceptable in move analysis, indicating a high level of agreement among raters.
Discrepancies increase reliability; moves perform communicative functions, but raters can differ in their understanding of a text portion. Examining discrepancies can provide valuable insights into the nuances of the text and the ways in which different raters interpret its communicative functions.
Using a Corpus-Based Approach
Corpus-Based Move Analysis
Typical discourse analysis is qualitative, focusing on few texts. Traditional discourse analysis often relies on the close reading of a small number of texts, which can limit the generalizability of the findings.
Corpus-based analysis requires a representative collection of texts encoded electronically for complex research findings (Baker, 2006, p. 2). A corpus-based approach enables researchers to analyze large amounts of text in a systematic and comprehensive way, leading to more robust and generalizable findings.
Corpus-based analysis uses both quantitative and qualitative techniques. The integration of quantitative and qualitative methods allows researchers to combine statistical analysis with in-depth interpretation of the text.
Researchers qualitatively identify and tag moves and types, and quantitative data must still be interpreted functionally (Biber et al., 1998, p. 4). The qualitative identification and tagging of moves is essential for capturing the nuances of language and the communicative purposes of text segments, the quantitative data must be interpreted.
A corpus-based approach to move analysis includes:
a) Analyses on a large representative text collection.
b) Electronic encoding for computerized counts.
c) Coding all texts using a move type rubric.
d) Analyzing linguistic characteristics of moves.
e) Quantitative counts to discuss general trends.
Advantages of Corpus-Based Approaches
Baker (2006) outlines four advantages:
Reduces Researcher Bias: Corpus-based approach places restrictions on cognitive biases (p. 12). By analyzing large amounts of text in a systematic way, corpus-based approaches can help to minimize the influence of researcher bias.
Incremental Effect of Discourse: Helps understand language use, revealing patterns in words and structures (p. 13). Corpus-based analysis can reveal subtle patterns in language use that may not be apparent through traditional methods, providing insights into the gradual evolution of discourse conventions.
Easier Counter-Example Identification: Facilitates the identification of counter-examples (p. 14). By allowing researchers to quickly search and analyze large amounts of text, corpus-based approaches make it easier to identify counter-examples that challenge existing theories or assumptions.
Combines with Methodologies: Reinforces and strengthens analysis via 'triangulation.' The ability to integrate corpus-based methods with other research methodologies, such as qualitative interviews or surveys, allows for a more comprehensive and robust analysis of discourse.
Specific Advantages for Move analysis
Identifying Linguistic Features: Allows descriptions of typical linguistic characteristics of move types. By analyzing the linguistic features associated with different move types, researchers can gain insights into how language is used to achieve specific communicative purposes.
Move Frequencies and Lengths: Enables description of distributional characteristics, such as whether a move is obligatory or optional. Analyzing the frequencies and lengths of different moves can reveal important information about the structure and conventions of a particular genre.
Mapping Move Use and Locations: Observations permit the analysis of relationships, such as text position of moves. By mapping the use and location of different moves within a text, researchers can gain insights into the overall rhetorical structure and flow of the discourse.
Genre Prototypes: With statistics on move frequencies and lengths, prototypes are valuable in educational contexts. Creating genre prototypes based on corpus analysis can provide valuable guidance for students and novice writers, helping them to understand the conventions of different genres.
Summary
Move analysis is a text-focused tool for analyzing discourse structure. It provides a systematic approach to understanding how texts are organized and how different segments contribute to the overall communicative purpose.
It has theoretical/empirical underpinnings. Move analysis is grounded in both theoretical frameworks and empirical research, providing a robust and evidence-based approach to discourse analysis.
Corpus-based move analysis follows guidelines for top-down analysis. By adhering to the principles of top-down analysis, corpus-based move analysis ensures that the analysis is driven by the overall communicative purpose of the text, rather than by individual linguistic features.
Advantages of corpus-based approach:
Ease of identifying moves' linguistic characteristics.
Understanding their frequencies and lengths.
Mapping their use and location