Mighty Digests # 12: MAZUR, Conversation Analysis for Educational Technologists

Article: Mazur, J. (2004). Conversation analysis for educational technologists: theoretical and methodological issues for researching the structures, processes and meaning of on-line talk. In D. H. Jonassen (Ed.), Handbook for Research in Educational Communications and Technology, 2nd Edition. Mahwah, NJ: Lawrence Erlbaum Associates. (pp. 1073-1098)

Article available online at: http://www.aect.org/edtech/ed1/40.pdf

I selected this article, because I would like to explore the potential of CA and DA in online asynchronous conversations (blogs/forums). This is one of the best introductory articles I have found on the topic.

This reading was also very useful to consolidate some of the key points presented in the books by ten Have and Wood/Kroger.

Through the following synthesis/analysis/critique I hope to “save the reading” for future reference.

Conversation Analysis (CA)

-       It is a qualitative approach

-       In the tradition of Discourse Analysis

The study of discourse is considered multidisciplinary (van Dijk, 1997)

Discourse studies focus on theory and analysis of text and talk in virtually all disciplines.

Discourse analyses tend to focus on several topics:

1. discourse as verbal structure
1. order and form
2. meaning (semantics): any proposition is influenced by what comes before or after it in the discourse. (van Dijk, 1997) Another key semantic concept is that of coherence (Tannen, 1986).
3. style (word choice)
4. rhetoric (figures of speech)
2. discourse as cognition (expression of the knowledge of the writer/speaker)
1. Symbolic and Connectionist Models of Discourse
3. discourse and society
1. Gender, Ethnicity, and Discourse Analysis
2. The Explanatory Elements of Cultural Context: The Critical Turn (Rules for politeness, topic changes, and giving commands change across cultures)
4. discourse as action and interaction

-       The concept of language as action was first defined by John Austen in 1962.

-       Speech act theory is that speech is action.

-       Pragmatics is the field of study focused on the study of language use as action in social context.

-       Interaction takes many forms: turn-taking in conversation, agreeing and disagreeing, questioning and answering, opening and closing conversation, preparing to engage in and enter conversation, developing persona in conversation, saving face, attacking or defending, and persuading or explaining. These interactions in their social context are the subject of conversation analysis, a subset of discourse studies of interaction.

Conversation = people talking with each other.

Conversation Analysis = any study of people talking together in oral communication or language use.

Talk-in-interaction = the phenomena of interest for Conversation Analysis (CA)

Conversation and interaction continue to be redefined and reshaped by computer-mediated communication technologies.

The seminal Conversation Analysis work by Sacks, Schegloff, and Jefferson (1974) articulated three basic facts about conversation:

1. turn-taking occurs
2. one speaker tends to speak at a time
3. turns are taken with as little overlap between them as possible

Turn-taking

Turn-taking is one of the central principles of Conversation Analysis:

1. turn form
2. turn content
3. turn length

These are affected by the formality or informality of a situation.

Turn construction units have two prominent properties:

Projectability = the speaker projects what kind of a unit it is and when it is likely to end.

Transition-relevant places = occur at the boundaries of the turn construction unit and make it possible for transition between speakers.

Turn distribution has some “simple rules” as articulated by Sacks et al. (1974) that occur at the initial transition-relevant place in a turn.

Turn-taking rules

Rule 1

1. If the current speaker designates the next speaker, that speaker should take the turn at that place.
2. If no such selection occurs, then any speaker can self-select, with the first volunteer having the right to speak first.
3. If no speaker is selected, the first speaker may (or may not) continue speaking with another turn construction unit, unless or until another speaker self-selects, at which point that speaker has the floor.

Rule 2

Rules 1a–1c are reiterated at the point of the next transition-relevant place.

These practices are in evidence in transcriptions of talk.

Traditional Conversation Analysis has used the focus on turns to develop insightful accounts of the structural organization of topic shifts (Jefferson, 1986), agreement and disagreement (Pomerantz, 1984), laughter (Jefferson, Sacks, & Schegloff, 1987), repair and correction (Schegloff, 1986), invitations (Drew, 1984), and overlapping talk (Jefferson, 1986).

Sequential Organization and Intersubjectivity / Strategies and Goals

The interest of Conversation Analysis (CA) in goals and strategies is how conversants show their understandings and orientations to each other using their talk as evidence.

Heritage (1997, p. 162) notes that there are currently two prevalent branches of analytic conversation research:

1. Pure CA (examines the institution of interaction as an entity with its own structural, social, and moral characteristics)
2. Applied CA (focuses on the management of social institutions in interaction)

Within the applied CA framework, CA is a systematic method to observe the production of intention or the achievement of understandings in the turns of talk between human speakers.

Until the proliferation of on-line exchanges in the form of typed “virtual” conversations, data in CA studies consisted primarily of audio recordings of talk in naturally occurring settings.

The relationship of the knower to the known in the CA analytic process is a mutually constitutive one. The CA analyst is “reading” the transcript and bringing her own interpretative frames and lenses to this task. The analytic task is recursive and folds back on itself.

Technology and conversation

It is important to distinguish in a general sense between two terms, often used interchangeably:

Computer Mediated Communication (CMC) = mediation by

Human–Computer Interaction (HCI) = interaction with

Affordances (Gibson 1979) = possibilities for action that are suggested by the physical features and inherent properties of objects. Gibson emphasized that affordances can enable or constrain based on their physical properties. These properties are not interpreted by human actors but, rather, are real, material dimensions of objects (the “grasp-ability” of the door knob does not depend on someone’s opening or closing a door).

Technologies are not neutral (Ellul, 1964). Communication technologies affect the quality and conduct of conversation and interaction.

Design of a CA study:

1. Formulating initial focus questions (related to talk-in-interaction).
2. Making a plan for obtaining or making recordings of naturally occurring interactions (transcribing or obtaining a transcript).
3. Developing inductive, analytic strategies.
4. elaborating the analysis in conclusions and implications.

The specimen approach

Using a specimen approach, drawn from techniques in naturalist observation and biology, the reality under study is directly observable in the particular individual.

The most important caveat for collecting and producing data for CA is rooted in the concept of naturally occurring conversation. Regardless of how it is produced or recorded, the data for CA are always in a sense “primary sources” of talk-in-interaction.

Collecting/Producing Records of Conversation:

-       Audio recordings

-       Video recordings (Goodwin, 1995; Heath,1997). Video is particularly appropriate when aspects of the physical setting of the conversations are intrinsic to the conversation such as engineers discussing design models.

-       Text Logs from On-Line Forums (Synchronous or Asynchronous). The text logs themselves contain “naturally occurring” conversant-generated indications of some of the sociolinguistic dimensions evinced in recordings of speech. The use of emoticons are an example of this phenomenon quite prevalent in text-based on-line conversations.

-       Digital Screen Recordings of On-Line Interactions (Screen Playback). Point-to-point videoconferencing, using a camera, conversants can engage in computer-mediated face-to-face talks. Using a screen recorder such as HyperCam that captures screen images and stores them as digital movies, a researcher could conduct CA on these types of conversations.

-       “Transcription” of Real-Time Log Data for Text-Based On-Line Conversation. The use of a conversation map, a technique that built on the work of Levin, Kim, and Riel (1990) using message maps and Herring’s (1999) multiple thread schematics, was helpful to orient spatially the various topic threads occurring over time in a synchronous chat (Mazur& Jones, 2002).

-       Documentation of the Affordances and Conventions of the CMC Forum Environment.

Steps for Conducting an Analysis of Conversation

In CA, the analysis proceeds from the perspective of what Psathias (1995) has referred to as “unmotivated looking” (p. 45), a way to achieve an “examination not prompted by pre-specified goals” (Schegloff, 1996).

Researchers in many contexts have offered suggestions for the task of systematically analyzing conversation (Pomerantze & Fehr, 1997; Schegloff, 1989; ten Have, 1999).

1. Select a Sequence.
2. Characterize the Sequence. Answer the question:
1. What is the speaker doing in this turn?
2. What is the topic of the conversation?
3. Is the person trying to initiate, repair (clarify, elaborate), or close an interaction?
4. What is the meaning of the interaction?
5. How is meaning conveyed, received, coconstructred through interaction?
6. What do participants talk about?
7. How do they signal topic changes or the need to stay on a certain point?
3. Consider the Rights, Obligations, and Expectations constituted in the talk.

It is important for the researcher to maintain an ethical posture toward informing participants that their work will be the subject of an analysis either by a “participant-observer” or by an external researcher.

“The character of the particular forum and the context in which it is situated should be overarching frames for deciding on the procedures that will yield the richest analysis.” (Mazur 2004)

Of course, any inquiry begins with a question.

A Case Example: Conducting a CA of an On-Line Chat:

Step 1: Obtain Required Permissions to Observe or Participate in the On-Line Forum

Step 2: Compile the Entire Record of the On-Line Talk (into a single text record: the transcript)

Step 3: Prepare the Transcript for Analysis

-       remove identifying data

-       remove electronic text “noise”

-       keep emoticons

-       change actual names to initials or nicknames

-       number the lines for the complete record

-       convert dates to numerics (e.g., January 3, 2001 to 010301) to denote sections of chat “sessions”.

Step 4: Read the Transcript

Step 5: Define the Sample “Specimen”

Step 6: Analyze the Specimen: Examine the CA Elements

Step 7: Contextualize the CA Theoretically

(Mazur 2004)

“Evidence began to accumulate that on-line forums might encourage broader and deeper participation in group activities (Kiesler, Siefel, & McGuire, 1984; Pullinger, 1986; Spitzer, 1989). CMC was seen to enable participation of handicapped students (Batson, 1988) and to encourage the participation of students often marginalized inface-to-face classroom settings such as women and minorities (Hiltz, 1986; Meeks, 1985).”

“Abdullah (1998) asserts that despite the fact that on-line conversation is written, it has nonetheless evolved to have a distinctly informal and conversational tone through the use of incomplete sentences, the use of lowercase letters to begin sentences, and uncorrected spelling.”

Gunawardena model:

Phase I: Sharing/comparison of information

Within this phase operations include:

(a) a statement of observation or opinion

(b) a statement of agreement from one or more other participants.

Phase II: Discovery and exploration of dissonance or inconsistency among ideas, concepts, or statements

Within this phase operations include:

(a) identifying and stating areas of disagreement

(b) asking and answering questions to clarify the source and extent of disagreement.

Phase III: Negotiation of meaning/coconstruction of knowledge

Within this phase operations include:

(a) negotiation or clarification of the meaning of terms

(b) negotiation of the relative weight to be assigned to types of argument.

Phase IV: Testing and modification of proposed synthesis or coconstruction.

Within this phase operations include:

(a) testing the proposed synthesis against “received fact” as shared by participants and/or their culture

(b) testing against existing cognitive schema.

Phase V: Agreement statement(s)/application of newly constructed meaning.

Within this phase operations include:

(a) summarization of agreement

(b) applications of new knowledge.

Social Network Analysis (SNA; Scott, 1991;Wasserman and Faust,1997)

A Typology of Exchanges Within a Virtual Community.

Another approach to the examination of interaction has been developed by Burnett (2002). At first glance this approach appears to draw on the tradition of discourse as cognition and takes a view of interaction as “information seeking,” and Burnett denotes information exchange as the phenomenon of interest in a virtual community. However, the typology developed is based conceptually on Savolainen’s (1995) research of nonwork, everyday life information seeking as “a natural component of everyday practices (p. 261; quoted in Burnett, 2002,p. 3). The typology focuses on the behaviors of participants as they interact, often within the dislocations of time and space inherent in on-line talk and interaction:

1. Noninteractive Behaviors

“Participation that seemingly does not exist”. Specifically this category refers to the activities of participants who do not type in talk but are actively following along as reader/listeners in the discussion.

This activity, often termed lurking in on-line forums, might be elaborated to include a more positive term with less voyeuristic connotations, listeners, for example.

2. Interactive Behaviors

“Posting or active message writing”. Interactive behaviors are further broken down into hostile and collaborative/positive behaviors.

1. Hostile Behaviors

-       impolite, uncivil, antisocial

-       flaming: CAPITALIZED TEXT

-       racial slurs

-       profanity

-       trolling (deliberate postings of inflammatory or provocative messages)

-       spamming

-       cyber-rape (explicit sexual verbal assaults)

2. Collaborative Interactive Behaviors

(1) behaviors not specifically oriented toward information

-       neutral behaviors such as pleasantries and gossip, humorous behaviors, and empathic behaviors that offer emotional or moral support.

(2) behaviors directly related to either information seeking or providing information to other community members.

-       announcements, queries or specific requests for information.

Persistent conversation

“The archive is a vehicle for so-called lurkers and nonconversants to participate as “readers” of the conversation.”

The primary modes of discourse are talk and text. (Mazur 2004)