New coding systemTo compare the nature of the interactions that occur in face-to-face and computer-mediated communications, I adapted Bellack, Kliebard, Hyman and Smith's (1966) model of pedagogical moves. However, rather than grouping phrases as they did to view primarily the back-and-forth interactions, I shall examine the content of these moves on a sentence-by-sentence basis, since the nature of asynchronous communication makes it impossible to define moves the same way.
My reasoning was that it would be unrealistic to presume that the same number of social acts would occur in a class taught by asynchronous communication as in FTF. Lacking the give and take of conversation, one is able to make uninterrupted statements. The function, method, and ratio of their contents, however, could be compared fairly. I needed a system that would allow analysis of the nature of the interaction occurring within both FTF and CMC learning environments, and thus permit an examination of oral and written interaction. This is particularly important for examining modes of classroom interaction in which non-verbal behaviours are less apparent or non-existent, such as in the mediated-communications of distance learning.
Influenced by the study of language and meaning by Wittgenstein (1958), Feigl (1949) and Brown (1958), Bellack et al. viewed classroom interactions as a sort of game in which there were specific rules and modes of conduct. They imagined classroom interaction as a series of reciprocal relationships: "Questions are asked to be answered; assignments are made to be carried out; explanations are made to be understood" (p. 2). They believed that classroom talk has a characteristic underlying structure and pattern from which teachers and students deviate little (Stubbs, 1983b).
Each of these relationships was called a "pedagogical move," the idea being that in the classroom people play different yet complementary roles in this game of language. For instance, one person may ask a question, and the other person (or a representative from a group of people) is obliged to answer it. In a similar way in which playing a game with a ball involves more than merely throwing, kicking, and catching, the game requires one to do one or the other to certain other people when the game has reached a certain state. In the classroom, for example, one may solicit a response of someone. According to the "rules" of the game, the other person must then respond.
To this end, Bellack et al. performed an experiment in which 345 16-to-18-year-old students in New York City were taught classes in economics by 15 different teachers. The classroom discussion was recorded, coded, and analysed.
Bellack et al. devised their coding system to determine the nature of these classroom interactions. They wanted to learn what teachers and students spoke about, as well as "how much, when, under what conditions, and with what effect" (p. 2). These moves provide a way to define how student-teacher interaction works by means of sequences.
Each pedagogical move was coded on eight indices: (1) the move's speaker, (2) the type of pedagogical move, (3) the move's Substantive meaning, (4) the move's Substantive-logical meaning, (5) the number of lines in parts three and four, (6) the move's Instructional meaning, (7) the move's Instructional-logical meaning, and (8) the number of lines in parts six and seven. Each move was coded in the format 1/2/3/4/5/6/7/8. An example of a coded pedagogical move, T/STR/IMX/XPL/4/PRC/FAC/2, would be interpreted as: "a teacher makes a structuring move in which he explains something about imports and exports for four lines of transcript and also states facts about class procedures for two lines of transcript" (italics in original, p. 16).
Bellack initially codes a pedagogical move by its "source of utterance" (p. 38), the speaker. The possibilities allowed consist of Teacher, Pupil, or Audio-Visual Device. Although this is a rather straightforward means of analysis, I found it paradoxically both overly broad and severely limited, as I shall discuss in my modification of the system later.
Bellack's system describes the basic verbal turn-taking of classroom discourse in terms of four basic pedagogical moves, a reference to their perceived function. These moves are divided by purpose in terms of being "initiatory" or "reflexive."
An initiatory move is one which begins interaction, either by setting its context or direct solicitation. The first initiatory move, structuring, sets the "context for subsequent behavior by launching or halting-excluding interaction" (p. 38) as well as "indicating the nature of the interaction" (p. 16f). Structuring moves indicate the direction the speaker thinks the discussion should take (Stubbs, 1983b). In this move, the speaker indicates what is going to be said in terms of "dimensions of time, agent, activity, topic and cognitive process, regulations, reasons, and instructional aids" (p. 17).
Although it is possible for these moves to be very short, for example, "Tomorrow we are going to discuss foreign investments" (p. 27), a move continues, regardless of its length, until its function changes. As a result, a move can contain more than one sentence. The following, for example, is coded as a single move:
Last night I was sitting out on the porch trying to cool off and also trying to ask myself how to best begin this unit on international economics, international trade, world trade, or whatever else one wants to call it. And after thinking over a number of approaches, I wrote down on paper a number of items that are concerned with our story. Some of them are with your experience. Some of them are found within your pamphlet (p. 17).
Bellack et al. also differentiate between structuring moves which occur in the normal classroom interaction and structuring moves which occur "as a result of an assignment, such as student reports or debates" (p. 17).
The second initiatory move, soliciting, is similar to the interrogative Elicitive mode of linguistic function used by (Mabrito, 1991) in his examination of e-mail as a vehicle for peer response. However, unlike CMC interaction, the soliciting move "directly elicits verbal, physical, or mental response" (p. 38). In keeping with this, the following sentences are all coded as soliciting moves:
What are the factors of production?May we keep our books open?
Turn the lights out, Bobby!
Pay attention to this! (p. 18)
Reflexive moves complete the reciprocal action of initiating moves. In contrast to an initiating move, a reflexive move is designed to indicate a response to previously initiated interaction. These also take two forms. The first reflexive move, Responding, "fulfills expectation of solicitation; bears reciprocal relation only to solicitation" (p. 38). This serves as the complement to the Soliciting move and may take the form of a verbal, physical, or cognitive response. In the following examples, a teacher solicits a response and the student responds verbally:
Teacher: "What are the factors of production?"The second reflexive move, Reacting, serves to modify or clarify a preceding utterance (Stubbs, 1983b). Unlike a Responding move, however, Reacting moves may be given in response to Structuring, Soliciting, Responding, or prior Reacting moves. The essential difference between these two moves is that Responding moves are always given in response to Soliciting moves, whereas Reacting may be in response to any previously given move. Bellack et al. give the illustration that "the rating by a teacher of a student's response is designated a responding move; that is, the student's response is the occasion for the teacher's rating reaction but does not actively elicit it" (p. 19).Student: "Land, labor, and capital."
Teacher: "Why didn't you do the assignment?"
Student: "I was absent yesterday."
Teacher: "What is exchange control?"
Student: "I don't know" (p. 18).
Bellack et al. also use a variant of this code to indicate when the response is the result of more than one move, for example to tie together two threads of discussion. Examples of this variation include:
"Now Kathy has mentioned natural resources and you raise the problem of human skills. Both are factors of production.""All of the instances of foreign investment that we have discussed here can be classified as either direct or portfolio types of investment" (p. 19).
In addition to the type of pedagogical move, Bellack et al. code moves in terms of four different kinds of meaning: substantive with associated substantive-logical meanings, and instructional with associated instructional-logical meanings.
The first of these, Substantive meanings, refer to the subject matter under study by the class. The topics for Bellack et al.'s experiment were taken from readings in international trade, and included categories of Trade, Imports, and/or Exports and Promoting Free Trade. These categories also included Relevant to Trade, used to code a "topic relevant to trade or economics but not considered" in the course text; and Not Trade, in which the discussion was "not about trade or economics and not related to either" (p. 22).
Substantive-logical meanings refer to the cognitive processes involved in dealing with the subject matter of the course being examined. Substantive-logical processes are categorised under three general types: Analytic, Empirical, and Evaluative processes.
Analytic processes consist of statements about the proposed use of language. Bellack et al. code these as Defining, (General, Denotative, or Connotative), or Interpreting.
Empirical processes give information about the world, based on one's experiences of it. In other words, one makes observations and decides whether the statement is true or false. As such, these types of moves are coded as Fact-Stating or Explaining.
Evaluative processes are statements that grade, praise, blame, commend, or criticise. These statements are verified by a set of criteria or principles of judgement. These include the codes Opining and Justifying.
Instructional meanings of moves describe matters pertaining to classroom management, assignments, and procedures that are part of the instructional process. Bellack et al. indicate these by the codes of Assignment, Material, Person, Procedure, Statement, or Logical Process (discussion of the way language is used). These also include codes for talk about Actions (General, Vocal, Physical, Cognitive, Emotional) and Language Mechanics (discussion of language use or grammar).
Instructional-logical meanings refer to the instructional moves such as those involved in positive and negative rating and giving instructions. These use the same codes described in the Analytic and Empirical processes used in substantive-logical meanings. In addition, the codes of Rating (Positive, Admitting, Repeating, Qualifying, Not Admitting, Negative, Positive or Negative, Admitting or Not Admitting) are used, and the processes of Performing and Directing.
Although Bellack et al.'s system was an excellent starting point for a coding system that could be used for analysing both FTF and CMC, there were still a few points which needed to be refined. To this end, I followed the steps described by Weber (1985) to classify text into categories of content. First, I defined the recording units, "the most basic unit of text to be defined," as sentences.
I opted to examine discourse by sentence for it permits an examination of purpose in both synchronous and asynchronous forms of communication. Atkinson (1981) explains that the ability to respond and participate in interaction in synchronous communication is made possible by the "typing" of utterances to the immediately preceding one. A basic form of this is described by Sacks, Schegloff and Jefferson (1974) as a minimal "adjacency pair." In other words, what one will say will most likely be related to what someone else has just said, such as a greeting and a reply. Asynchronous communications, however, lack the immediate feedback of FTF interaction. As a result, the pattern of interaction is going to change. One may still preface communications with a greeting, but one is unlikely to wait around for an acknowledgement before continuing.
The assumption in Bellack et al's system, however, is that interaction is synchronous, without delays in interaction. As a result, one would find it difficult to use their coding system asynchronously. Bellack et al. assume that large amounts of time do not pass between moves, and as such coders are able to draw distinctions between Reacting moves and Responding moves, for example. However, in asynchronous communications the delays between an initiation and a response distort time, making the experience of it different for each participant. Thus, one cannot be sure of interpreting reflexive moves correctly. This can only be solved by examining each sentence in terms of its purpose in the overall discourse, which allows one to go beyond the role of its move and focus on its purpose in the communication process.
There is, however, a minor difficulty to working with a system based upon sentences rather than moves. The problem lies in defining what the sentences are in the first place. Within conventional written language, sentences are easy to determine because the author has inserted punctuation or used space to indicate their beginning and endings. In this respect, defining the sentences from the CMC transcripts was simply a matter of using the sentences as they were written by the CMC participants.
The FTF courses, however, were more difficult. People do not always speak grammatically, and so their utterances do not always take the form of a proper sentence, containing at least one subject with its predicate and a finite verb or verb phrase. For better or worse, the syntax of spoken language is different from the syntax of formal written language (Crystal, 1980). Spoken language tends to take the form of loosely co-ordinated clause complexes.
Stubbs (1983) notes that an attempt to divide an oral transcript into sentences "involves making arbitrary decisions about sentence boundaries, due to the large number of clauses coordinated with and, and a few other items such as but and then" (p. 35). I postulate that the determination of sentence boundaries, like word boundaries (Pinker, 1994), is a learned communication skill. As such, it would be best to have those with the most skill interpret these sentence boundaries. As a result, these were defined by a professional transcriptionist, someone for whom recognising sentence boundaries is an essential qualification for his or her job.
Having decided to examine discourse by sentence, I then defined the categories of my system (Weber, 1985). Initially, I began with Bellack et al.'s system and tried coding with that, but I found several problems with it.
The first problem I had with Bellack et al.'s system is its process of quantifying the size of each move by counting its lines [33]. Bellack quantifies the amount of interaction by counting the number of lines spoken: "All complete utterances of less than one line are counted as one line" (p. 255). Although this makes sense in that any utterance requires some transcript real estate to exist, it fails to differentiate between a one-word response and a one-line response. An analysis of Bellack et al.'s sample transcript (pp. 267-269) reveals that complete lines from the transcript contain an average of 8.7 words.
This give unfair weight to utterances that are less than one line-length in length, which is particularly important if one is to analyse differences between teachers and students. Of the nine teacher utterances in Bellack et al.'s transcript, only one consisted of a single word. Of the eight student utterances in the transcript, five consisted of one word. If this pattern of one-word turns is extrapolated to its lowest-common denominator, a teacher would utter one-word lines eight times in 72 utterances compared to the students' 45.
Admittedly, this is a small sample. However, if it is representative of the transcripts it indicates an overestimation of the amount of student utterances in Bellack et al.'s analysis. In the same 72 lines, a teacher would have said about 565 words:
= (64 full lines @ 8.7 words/line) + 8 one-word turns= (64 x 8.7) + 8
= 556.8 + 8
= 564.8
more than twice as many as the students' 280:
= (27 full lines @ 8.7 words/line) + 45 one-word turns= (27 x 8.7) + 45
= 234.9 + 45
= 279.9
Although this is irrelevant in terms of examining the nature of turn taking, which was Bellack et al's aim, one could not accurately use this method to describe the amount of interaction which occurs in a classroom. Student talk would be overestimated, and differences in the quantity of utterances by teachers and students would appear smaller than they were.
A second problem with Bellack et al.'s coding system is that the initiatory moves, Structuring and Soliciting, are insufficient for coding the moves of a lecture (as opposed to recitation or discussion) environment. Structuring is used for setting up context and Soliciting is used to elicit a response. Unfortunately, neither of them may be used to code utterances which neither set up context nor request interaction, as would be evidenced in a lecture.
In Bellack et al.'s examples, few of the teachers deliver monologues. This may not have been a problem in other academic studies using this system, for one rarely lectures to children aged nine to 13 (Bergen, 1984; Edwards, 1980; MacDonald, 1981; Polzella, 1988; Scanlan, 1988; Webber, 1980; Wilson, 1984), or 15 to 18 (Sapiens, 1982; Wing, 1980). In fact, few studies using Bellack et al.'s system have been conducted with adult students (Malcomson, 1990; Rossignol, 1992) in which lecturing may have occurred.
At best, one could rationalise that Soliciting would be the default move, in that it elicits one to listen and think about what is being said. This default mode of coding, however, invalidates moves by failing to distinguish between moves which are truly requesting interaction from others and those in which it is assumed. It would be difficult, for example, to find a teacher who doesn't believe that his or her students should always be listening and thinking about what is being said.
A third deficiency with Bellack et al.'s system is the distinction between the two reflexive moves, Responding and Reacting. Both serve the purpose of acknowledging the stimulus of a previous move. Responding fulfils the expectations of Soliciting moves only. Reacting modifies or rates what was said in any preceding moves. The difference is, briefly, that Responding serves as a direct reply to a specific elicitation whereas Reacting is a response to a previous move, but is not specifically elicited by that move. As I described earlier, in asynchronous communications it would be difficult to determine the context, and consequently the nature of the response. However, even in synchronous communication the differences are somewhat arbitrary. Often the functions of Reacting and Responding moves are combined within the same sentence and there is a point at which they can be so intertwined within a sentence as to make drawing a distinction between them useless. Viewed in a larger context, for example, the entire course could be full of Reacting moves to the teacher's initial Solicitation, "What is this course?"
Bellack et al. submit the following as an example of a teacher's Reacting move following a student's Reacting move:
"Good. It limits specifically the number of items of one type or another which can come into the country. For example, we might decide that no more than one thousand of German automobiles will be imported in any calendar year. This is a specific quota which the government checks" (p. 19).
I maintain that this is an arbitrary distinction and a poor example. The first sentence of the example, "Good." fits squarely within the definition of a Reacting move for it rates the submitted response, most likely a student's response. Unfortunately, it could also fit within the definition of a Responding move in that it could answer a previous Soliciting move, such as a spontaneous insight from a student concluded by a request for the appropriateness of the insight. (This, however, is less likely, given the power relationships in the classroom.)
The second sentence is Reacting since it serves to clarify or expand the student's answer. The third sentence, however, is doing more than merely acknowledging, repeating, or expanding. True, it is expanding, but it is no longer expanding directly as a result of anything the student has answered. In other words, within the context of the paragraph, there is nothing uniquely Reacting-like about that third sentence. It does not serve to answer the student's question. Rather, it is used as an opportunity for the teacher to lecture and give an example. This sentence could fit just as well under the label of a Structuring move. As Bellack et al. have set it up, the differences between Responding and Reacting moves are specious.
A fourth problem with the coding system is that Bellack et al.'s Substantive Meaning code does not account for progressive levels of meaning, which makes it difficult to use the coding system with any subject in which the same ideas or procedures are used at higher levels. For example, in a course of introductory algebra one would continue to use procedures learned earlier, such as multiplication and division, but instead as being seen as ends in themselves, they would be seen as part of a larger process. For example, suppose I were to teach someone how to determine the value of y in the following equation:
After explaining the idea of isolating the variable on one side of an equation, I would then describe the specific process of subtracting one from each side to achieve this end. According to Bellack et al., the substantive meaning would be "solving for single variables." Unfortunately, although this gives a gross view of what is happening, it does not take into account that I might actually be explaining that the opposite of addition is subtraction.
One could devise a modified coding system in which one used the highest-level process available. However, this would still require the coder(s) to determine the highest-applicable level for each move. And, if this were done what one was actually taught would not be conveyed, but rather the desired result of what one was teaching.
A fifth deficiency with Bellack et al.'s system again features their Substantive Meaning codes. These codes were specially designed for each course based on course material used in the experiment, and could not be used to compare data without being modified. More precisely, since the categories "were developed by means of a content analysis of the pamphlet" on which the course was taught (p. 21), even if only the text for the course changed, these categories would have to be altered.
However, the problem doesn't stop there. These categories also prevent the use of this system to compare the language of the classroom with the same course that uses a different textbook, or across different courses. Bellack treats this bug as a feature, noting that with "appropriate modifications" (p. 250) this provides a potentially useful technique for the examination of subjects of the secondary-school curriculum. Such modification would, however, render comparisons between courses impossible.
The sixth pitfall with using Bellack et al.'s coding system is that the codes for Instructional Meaning are too limited, specifically in terms of available codes for responses. For example, "yes," as one might use for the acknowledgement "Yes, I heard you" cannot be coded because, according to the system, there is no Instructional Meaning. At best, it could be coded as a Statement, but this is inaccurate. An acknowledgement is not a statement; at best it is a small subset of all possible statements. Understandably, Bellack et al. were examining the pedagogic functions of discourse rather than the social maintenance of personal relationships and interpersonal dynamics. Nevertheless, this is inadequate, and becomes worse the more closely a course approaches the characteristics of a discussion.
The seventh problem with the coding system is Bellack et al.'s inclusion of Audio-Visual Device as a valid source of utterance. Content is an important part of the educational process, as I have discussed earlier in this work. Unfortunately, including Audio-Visual Device as a source of utterance inflates its importance in the context of classroom interaction to that of a participant. Although educational contributions can be obtained from such materials, they simulate the pseudo-interactional aspects of the lecture.
Furthermore, if one were to include the text of a film, for example, with the reasoning that it is part of the educational experience, one would be similarly obligated to include the text of the assigned readings or videotapes observed in class. Although the student may be absorbing what is presented in a film, for example, one could not claim that the student is interacting with the content in an observable manner. For the purposes of this study I shall examine only the overt, interpersonal, verbal classroom interaction.
An eighth shortfall of the coding system developed by Bellack et al. is the single category for all students. Since the 1966 publication of their coding system, there has been much research examining classroom interaction as a function of student sex. Their system does not allow examination of this element, and it is time to correct this oversight. By permitting these differences to be acknowledged, comparisons can be performed. If required, however, these categories can also be combined to recreate the original.
A final problem with Bellack et al.'s system is the specious distinction drawn between discussion that occurs "as the result of an assignment" and intra-classroom discourse. Defining a Structuring-Assignment pedagogical move is particularly difficult in courses where the teacher guides discussions which make use of knowledge gleaned from the assignments, but doesn't discuss them directly.
In addition, this code fails to take into account topics which may have initially been assigned as homework but are then built upon in class. For example, in the FTF Systems Analysis courses the teacher explained how to prepare for a group project. The bulk of the group work was performed as homework, but the assignment was to be presented to the class. Although the final presentation, the finished assignment, was a result of homework, it was also a reflection of things which were discussed within the classroom. Trying to code this would require determining whether the presentation is a result of the homework or the class. Perhaps the prepared portions of the presentation would be coded as related to homework, but answers to the teacher's questions would not be. On the other hand, what if a student answered the teacher's question with a prepared demonstration or statement? Such a distinction is an invitation to confusion.
Having discovered these problems with Bellack et al.'s coding system, I then continued following Weber's (1985) steps. I condensed several categories, created several more, and revised the overall structure to create a more flexible coding system. I pilot-tested this system on a sample of 200 randomly selected sentences from the FTF transcripts (see Appendix H). My wife coded the same 200 sentences, and we compared results when we were done. Our codes agreed on 127 sentences.
I examined where we had disagreed and made some changes to the coding system and definitions of the categories. Having done this, I again tested my system on a pseudo-randomly-selected [34] sample of text from the transcripts. My wife and I each coded 500 sentences and compared our results. Our codes agreed on 437 sentences. I then added Figures Figure 9 and Figure 10 (see pages 64 and 72 respectively) to aid in decision-making and refined the definitions of the codes on how they were to be distinguished from other, similar, codes of the same tier.
Finally, my wife coded the remaining 51,581 sentences, using the coding system described here.
Purpose, as defined in this work with an upper-case P, is similar to the pedagogical functions of Bellack et al. which describe "the verbal maneuvers of both teachers and students in classroom discourse and set the framework for the analysis of meanings communicated in the classroom" (p. 15).
At first glance, it may seem odd to assume that every sentence one utters has a purpose. Of course not all purposes are conscious. For example, in FTF interaction, people do not consciously intend to say "uh" or "um" 1.4% of the time [35] and yet such utterances do serve a purpose in speech, even if only to demonstrate that the speaker is inarticulate. To this end, the categories for Purpose are defined as Organising, Lecturing, Eliciting, Responding, Idling, and Not Clear.
Organising is based on Bellack et al.'s definition of Structuring, which serves the pedagogical function of "setting the context for subsequent behavior by either launching or halting-excluding interaction between students and teachers." For instance, teachers use Organising to "focus attention on the topic or problem to be disclosed during that session," a sort of advance organiser (p. 4). Bellack et al. emphasise the use of their Structuring move as a means of transition, a means of setting "the context for subsequent behavior or performance" and, consequently, a means to direct the educational process in a certain direction "by focusing on topics, subjects or problems to be discussed, or procedures to be followed" (p. 17).
Like Structuring, Organising sentences are not designed to elicit a response, nor are they responses in and of themselves. Organising is used to set up an agenda, to organise a discussion or recitation, and as a means to get to other Purposes. Organising is like a sliproad or on-ramp to the discussion, helping the speaker describe upcoming ideas, behaviours, or events. Examples of Organising include:
"In a minute I'll be handing you an overview of the course as well as handouts for the first session." [Organising/Fact-Stating/Procedure]"We'll talk a little bit about that at the end of the session." [Organising/Fact-Stating/Content]
"I will give out the exam after the break." [Organising/Fact-Stating/Procedure]
Eliciting moves consist of solicitations or directions. Like Bellack et al.'s definitions of Soliciting and Directing, they "encourage persons addressed to attend to something" or solicit a physical response. All questions, commands, imperatives, and requisitions are Eliciting moves. Eliciting moves are specifically designed to cause interaction. This includes asking a question giving directions or instructions. Examples of Eliciting moves include:
"Send me a Response to Response if you have any questions concerning the basic forms creation process." [Eliciting/Performing/Procedure]"And what is the data warehouse?" [Eliciting/Fact-Stating/Content]
"By the way, if anybody has anything to add from their definition, please, you know, uh, jump in." [Eliciting/Performing/Content]
Responding moves combine Bellack et al.'s categories of Responding and Reacting. As such they form a reciprocal relationship to any sort of previously uttered move. It is important to note, specifically in CMC courses, that one may respond to a single word of a post just as easily as one may respond to its entire content. As a result, a Responding move may be given in response to any previously uttered or posted move or moves.
A Responding move has concluded when sentences cease to serve the function of directly responding to the previous moves. For example, if one were to mention Lotus' intention to include HTML tags in its next version of Notes, another participant's Responding move may be to talk about the HTML implementation, such as would it include the latest non-standard tags. However, if the second participant were then to go off on a tangent about World Wide Web server software, this would no longer be a response to the original discussion of Lotus Notes.
Examples of responding moves include:
"Yeah I hear, I hear." [Responding/Rating/Person]"All right." [Responding/Rating/Person]
"As Group D did not provide Response or Response to Response forms for any of their three forms, Group C may assume these two forms will be identical to the Response and Response to Response forms in the Analysis.NSF database." [Responding/Explaining/Procedure]
Lecturing consists of talk about the course content that is neither explaining a change in topic (Organising), soliciting a response (Eliciting), nor Responding. Lecturing is differentiated from Responding in that Responding is directly applicable to an Eliciting purpose. When the Responding move has moved away from the purpose of merely answering the Eliciting, it is then Lecturing. For example, suppose a student asked a teacher what colour fire engines were. The immediate answer, "red" (or "fluorescent yellow-green") would be a Responding sentence, but anything beyond that, such as explaining why so many fire engines are red, would be Lecturing.
When Lecturing one is not engaging others in dialogue. The lecturer is talking or writing away and continues to do so until he or she stops to solicit, is interrupted, or abruptly changes the direction of the lecture via an Organising move. Some examples of Lecturing include:
"We went a little over tonight, but that's all right." [Lecturing/Fact-Stating/Content]"Here, um, student one and student two- ah, the address here represents, one case is a home address." [Lecturing/Explaining/Content]
Traditionally identified with distance education, Humanising is used to create an atmosphere conducive to interaction by means of making students feel welcome with jokes, comments about the weather, or other light banter (Monson, 1978). Relevant personal vignettes, anecdotes, and experiences encourage trust among participants and reduce anxiety (Burge & Howard, 1990). This fosters a receptive learning environment, enhancing the climate for motivation, creativity, brainstorming, and risk taking (Bruce & Shade, 1994; Holmberg, 1988).
Bellack et al's coding scheme does not consider interaction which is free of pedagogical meaning, but is intended simply to facilitate the learning environment. Humanising differs from other codes by means of intent -- the purpose of the utterance is to make someone feel at ease or maintain the relationship. The idea is that getting students involved will aid the educational process. Asking questions is not enough; students also need to feel engaged on a more personal level. Although Humanising is technically free from any pedagogical content, it aids in creating an enjoyable and involved classroom environment. Some examples of this include:
"So, will the jury please disregard that [36]." [Humanising/Fact-Stating/Content]"You don't prefer to be called Jill?" [Humanising/Fact-Stating/Person]
"Let me check for my wallet when I talk to you." [Humanising/Fact-Stating/Content]
At this juncture, special mention goes to the use of emoticoms. Although they do serve as signposts, warning the unwary reader of the hazards of sarcasm and irony in the treacherous waters of discourse, emoticoms "quickly can become overworked clichés and of limited value" (Bellman, 1992). In some USENET newsgroups, posters are flamed for using them. Nevertheless, in interaction, emoticoms are used for a variety of Purposes, including Eliciting "Get it? It's funny. Get it?", Humanising "I'm such a merry dog; see, I'm making a joke.", and Responding "I thought that it was so amusing I chuckled, grinned, and chortled." As such, each emoticom must be interpreted in context to determine its function.
For example, in a typical instance, ":)" used at the end of a CMC post to explain that the participant was joking was coded as:
<smile> [Humanising/Explaining/Person]
A second category of Purpose which Bellack et al. do not include is that of Idling -- content-free, yet recognisable discourse. Idling sentences are those in which every word is understandable, but together they do not seem to serve any pedagogical function, and, unlike Humanising, they aren't designed to create an enjoyable classroom environment.
Although pedagogically the speaker may appear to be blathering, Idling sentences serve as a sort of place-holder, allowing the speaker time to collect his or her thoughts while maintaining control of dialogue. Some examples of Idling include:
"Okay, let's take ah let's see." [Idling/Filler/Not Clear]"You know, that, that kind of thing, you know." [Idling/Filler/Not Clear]
"That you, you know when you, oh no, no, no." [Idling/Filler/Not Clear]
The final category of Purpose is that of Not Clear. This is used when one is unable to determine the Purpose of the utterance due to missing or unintelligible words. This is similar in function to Bellack et al.'s Not Codable category, which served to indicate "function uncertain because tape is inaudible" (p. 38).
As a way to avoid confusion in determining the purpose of utterances in an educational context, Figure 9 was devised. It uses the metaphor of a motorway as the course. The idea of a course is to go forward, much in the way that one travels along a road, to arrive at an end or mastery of the course content.
Figure 9: Purpose as a motorway
To begin this journey, one must get to the motorway by means of directions called Organising. At the very least, Organising utterances would explain the name of the motorway, the length of the journey, what one might expect to see, and the destination.
Once on the motorway, one proceeds forward by means of Lecturing. Lecturing is the most direct route to the end, provided one's purpose is to broadcast declarative knowledge rather than negotiate its meaning.
Occasionally, one is required to exit the motorway for food, fuel, and stretching one's legs. This process of Humanising is necessary for continuing the journey, but does not in itself get one any further along in terms of transferring declarative knowledge. It is, however, necessary for building relationships, motivation, and interest.
Eliciting is a detour from Lecturing in that it lacks the speed and efficiency of the direct route. Yet, similar to a scenic route, an Eliciting utterance is designed to help people appreciate not merely where they're going but also to understand the beauty of the journey itself, the collaborative process.
A Responding utterance is one which brings one back to the Lecturing motorway.
An utterance devoid of Humanising or course content is Idling -- the participant is sitting in the car with the radio on, making noise but not going anywhere.
The second tier of the coding system is similar to Bellack et al.'s instructional-logical meaning. This part of the coding system describes the means or agency by which the meaning of the sentence takes place -- how the subject of the sentence is being discussed.
Fact-Stating includes Bellack et al.'s definition of Fact-Stating, "what is, was, or will be without explanation or evaluation; account, report, description, statement of event or state of affairs." In Fact-Stating sentences, statements are made which describe the definition or nature of things. Some examples from the transcripts include:
"This is of course, an class evaluation questionnaire." [Organising/Fact-Stating/Supplies]"The overhead projector is on fire." [Eliciting/Fact-Stating/Action]
"Ah, JAD which is Joint Application Development, ah, allows you to conduct many meetings at once." [Lecturing/Fact-Stating/Content]
The Mechanism of Explaining is a combination of Bellack et al.'s definitions of interpreting, "verbal equivalent of a statement, slogan, aphorism, or proverb," explaining, "relation between objects, events, principles; conditional inference; cause-effect; explicit comparison-contrast; statement of principles, theories, or laws," and justifying "reasons or argument for or against opinion or judgment."
Explaining sentences differ from Fact-Stating ones in that one can mentally insert the word "because" or "since" in the sentence. Pedagogically, this code is used for sentences in which a clarification, definition, or rationale is given. For example, a Fact-Stating sentence might be, "Your aunt is wicked." whereas an Explaining sentence would be "Your aunt is wicked because she deliberately missed your mother's funeral." Some examples from the transcripts include:
"I was absent last week, I'm not sure of exams." [Responding/Explaining/Procedure]"In our example form, you would then enter six lines of static text before reaching the next field creation position (for StudentID)." [Lecturing/Explaining/Procedure-Supplies]
"It's a tool, and just like other tools (say automobiles, guns, and chain saws), people can use it constructively and destructively, wisely and wastefully." [Responding/Explaining/Content]
Opining is similar to the definition presented by Bellack et al., which includes "personal values for statement of policy, judgment or evaluation of event, idea, state of affairs; direct and indirect evaluation included." In Opining sentences, the participant gives his or her thoughts of what should be done or evaluates the fairness, importance, or merit of an action, person, procedure, supply, or course content. Some examples from the transcripts include:
"That's the way I would do it." [Responding/Opining/Content]"As many large companies conform to a team work group concept, I feel there is a growing need for educated managers in this field." [Lecturing/Opining/Person]
"I think that one of the most important skills that a systems analyst must have is personal skills." [Lecturing/Opining/Content]
The Mechanism of Performing is similar to Bellack et al.'s instructional-logical meanings of Performing and Directing, in which one requests or expects an action to occur. Quite simply, Performing is the process of telling someone to do something. Some examples from the transcripts include:
"When you are finished with the student biography, pass them towards the center, please." [Eliciting/Performing/Action]"Okay, this is the, ah, basic model, ah, let's talk about variations or any questions on it, if you have questions." [Organising/Performing/Procedure]
"Okay, let's just do a couple more and, are you finding this exercise gives you a feel for the real practical basis?" [Eliciting/Performing/Content]
The definition of Repeating is similar to that of Bellack et al. in which one (in their case, presumably the teacher) repeats or rephrases what is said (presumably by a student) as a way to indicate "an implicit admitting" that the offered response was correct. An illustration from the FTF transcripts:
Teacher: Now remember, the output of analysis is the input to-- [Eliciting/Performing/Content]Student: To design. [Responding/Fact-Stating/Content]
Teacher: To design. What else do you need to know? [Responding/Repeating/Content, Eliciting/Fact-Stating/Content]
My coding system, however, has broadened this definition to also include repetition used for the purposes of setting context, which is not necessarily an indication of agreement. In this example, the use of repeating isn't merely to show one agrees (in the sense that the student is correct), but it is also used for getting attention -- as a point of focus -- to provide context for actions to follow. In this case, the context is "You say 'design.' I acknowledge your answer, and use it as a point of focus or context for my next move." This is also used in asynchronous communications, such as USENET newsgroups or courses delivered via CMC, in which text is quoted so that other participants will understand what the respondent is talking about.
The Mechanism of Rating combines Bellack et al.'s instructional-logical categories of Rating with Acknowledgment to include sentences regardless of the nature of their "reference to metacommunication; usually an evaluative reaction." Pedagogically, Rating serves to appraise (either positively or negatively) or acknowledge a participant's move. Following are some examples from the transcripts:
"Good." [Responding/Rating/Content]"He's right." [Responding/Rating/Person]
"Okay, that's one." [Responding/Rating/Content]
A Rhetorical Device is an Eliciting move that is used not to solicit a response, but rather to serve as an illustration. The Rhetorical Device gives the appearance of interaction without the risk of going off topic or derailing the speaker's train of thought with potentially irrelevant matters.
The Mechanism of Rhetorical Device differs from the other Mechanisms. If used in conjunction with an Eliciting Purpose, no answer is expected or only one answer may be made. The teacher may thus use a Rhetorical Device in the middle of a Lecturing/Fact-Stating series to make a point or simulate interaction with other participants. For example, in one of the FTF courses, the teacher referred to an often-used diagram from the textbook: "This, for the 937th time, is the conceptual model for uh Pine Valley. How many of you are tired of Pine Valley? How many of you are more tired of Mountain View? Uh, this is the relationship, this is the logical model that uh relate them." An analysis of these sentences revealed the context for the Rhetorical Device:
| Sentence | Purpose | Mechanism | Subject |
| This, for the 937th time, is the conceptual model for uh Pine Valley. | Lecturing | Fact-Stating | Content-Supplies |
| How many of you are tired of Pine Valley? | Eliciting | Rhetorical Device | Person |
| How many of you are more tired of Mountain View? | Eliciting | Rhetorical Device | Person |
| Uh, this is the relationship, this is the logical model that uh relate them. | Lecturing | Fact-Stating | Content-Supplies |
The two middle sentences made up the Rhetorical Device. Examined individually, the Lecturing/Fact-Stating sentences are broken up by Eliciting/Rhetorical Device sentences. In addition, the Subject of the sentences changes from Content-Supplies (the diagram) to Person (the students).
The Mechanism of Filler is similar to the Purpose of Idling in that it was possible to record what the participant said, but in context the statement didn't make any sense. Pedagogically, it is content-free. This sort of redundancy is necessary within this tier of the coding system to account for sentences that have a clear Purpose but lack an intelligible Mechanism. For example, in the context of a transcript a sentence may obviously be Responding, but the Mechanism by which it is doing so may be unclear. Sentences with a Mechanism of Filler from the transcripts include:
"Okay, let's take ah let's see." [Idling/Filler/Not Clear]"But uh, I'll tell you." [Lecturing/Filler/Not Clear]
"You know something." [Responding/Filler/Person]
The final Mechanism category, Not Clear, is used for sentences in which the words are unintelligible or a Mechanism cannot be determined. For example, in a Not Clear sentence from the transcripts, "It is so, is that what." [Responding/Not Clear/Not Clear] the Purpose is clear -- in context it's a response -- but how the speaker is going about this is a mystery.
The Subject tier of the coding system refers to what is being discussed in a sentence, the content being considered or statements about something. This is somewhat similar to Bellack et al.'s Instructional Meaning in which factors related to classroom management are discussed.
The most basic level of the Subject coding is Person. Any sentence in collaboration or even a lecture, is, if nothing else, addressed to a person. These sentences are used to describe one's dress, occupation, a personal experience, or to make personal remarks. Examples from the transcripts include:
"I was pleased to have the opportunity to get to know you a bit through my role as Client in the Systems Analysis course Case Study." [Lecturing/Opining/Person]"I'm going to the convention next week, so I'm going to be out Monday through, won't be returning until Saturday." [Responding/Fact-Stating/Person]
"Ah, you're from human resources right?" [Eliciting/Fact-Stating/Person]
The next level of Subject coding is Action, in which a Person or an object does something. This definition includes Bellack et al.'s definitions of Action -- General, Vocal, Physical, Cognitive, and Emotional. Examples from the transcripts include:
"Those of you who do not have books, look on." [Eliciting/Performing/Action]"How many of you have backed up files on, uh, your PC?" [Eliciting/Fact-Stating/Action]
"Just keep in the back of your mind." [Eliciting/Performing/Action]
A Procedure is a subset of Action in which one is told how to do something, rather than merely to do something. My definition of Procedure is similar to that of Bellack et al.'s "Discussion of any course of action or set of activities, continuing activity, or future activity. Includes references to how class is to be conducted and what regulations are to be observed. Refers also to specific instances of class regulations or practice." My definition also includes Bellack et al.'s "assignment" category, for the reasons I explained earlier. Together, these provide a category for coding explanations of how one does something. Some examples from the transcripts include:
"We'll take a short break and then--" [Organising/Fact-Stating/Procedure]"If you've already made a new replica of the User's Guide, please do not replicate it further until the above mentioned posting." [Eliciting/Performing/Procedure]
"What I'm going to is, is, have the groups, let's say the A group over here, the B group over here, and the C group in the back." [Lecturing/Fact-Stating/Procedure]
The Content code is used for sentences relating to the course content. For example, in the instances of the courses observed in this study, content would refer to sentences regarding Database Management or Systems Analysis. This would be similar to Bellack et al.'s Substantive Meaning, except rather than defining it in terms of the specific content of a subject, Content is defined as being on-topic. If a sentence refers to the subject of the course, it is on-topic, and would therefore be Content. Examples from the transcripts include:
"With relational, you basically retrieve multiple records at a single time and the system decides how to access based on your call." [Lecturing/Explaining/Content]"Our system is going to produce the report for the data load that goes to ADP." [Responding/Fact-Stating/Content]
"Logistically speaking, this would require students and instructor to have access to a computer 'connected' to a central database." [Lecturing/Fact-Stating/Content]
The category of Supplies encompasses Bellack et al.'s definition of Material, which includes "teaching aids and instructional devices." The classroom, forms, books, projectors, VCRs, blackboards, and computers are all Supplies. Some examples from the transcripts include:
"This tape is about three years old." [Lecturing/Fact-Stating/Supplies]"This will allow Notes data to enter the mainstream and, in the last demo I went to at Lotus (Whitehall St) last week, synchronize databases on an incremental basis." [Responding/Explaining/Supplies]
"The book is an excellent book." [Lecturing/Opining/Supplies]
The Not Clear code is used when the Content of the sentence is unclear or the topic of discussion cannot be deduced.
Figure 10: Venn diagram of the subject hierarchy
In order to simplify the coding process, Figure 10 was devised to illustrate the following analysis involved in determining the Subject of a sentence. Consider it an illustration of several transparent blocks placed in a pile. The more specialised the Subject of the sentence, the higher and smaller the block which is used to represent the Subject.
Part of the surface on which these blocks rest is painted grey, for Supplies. The blocks are arranged so that only half of them are actually above this grey part. This is to show that although each of the higher blocks can involve Supplies, i.e., a sentence could be coded as Person-Supplies, Supplies are not necessarily a subject in all sentences. Supplies may be used with another Subject or alone.
The next level of Subject is Person. Philosophically, it is possible to have a course with no people in it. However, without students there is no classroom discourse. In a similar manner in which Thorpe (1979) declares that a pre-packaged course is not a course until it is engaged with by a student, I shall take as an initial premise that a course contains one or more students. Therefore, since the most basic requirement of a course is students, the most basic subject of an utterance is Person, talk to and about people.
The second most basic Subject of a sentence would be an Action. An Action would be at a higher level -- more selective -- than Person because one cannot take an Action without people, but people can talk without taking any Action. Action is in this instance a subset of Person.
The next level of Subject is Procedure. Procedure is a subset of Action because explaining how to do something in particular is a subset of a large range of possible actions one could perform.
The highest level of sentence Subject is that of Content. Content is a subset of Procedure because in view of the large number of things one could potentially explain how to do in a course, not all of them are related to the course content.
To put it another way, at the lowest level, the teacher is talking about what the student (Person) should or could do (Action). In some cases, the teacher could ask the student to do something specifically related to the supplies (Supplies). More specifically, the teacher directs the student to act in a certain way (Procedure). If this is part of the course content, it is coded as Content; if not, we code this as Procedure.
If a sentence encompasses more than one subject, it is coded as the highest appropriate level. Thus, to use Bellack et al.'s example, a teacher directing students to engage in some classroom procedure which required the use of supplies would be coded as Procedure-Supplies. The rationale is that to have the course, one must have students (Person). These students may be directed (Action) to engage in some classroom Procedure (Procedure) which required the use of Supplies (Supplies).
To determine the reliability of the coding, I randomly selected 1,000 records from the 52,081 records in the database and coded them myself. The degree to which my coding agreed with my wife's was the correlation of reliability (Carmines & Zeller, 1979; Traub, 1994).
To select the random records, I added fields to the FileMaker database to permit the marking of selected records and wrote an AppleScript which would select records and mark the new fields in the database. (For a complete explanation of this process, see Appendix H.)
Having selected the records, I then examined each sentence in context and coded it myself using the system described earlier. Whilst I coded the records, the database automatically summed the number of times my code agreed with that of my wife. When completed, the mean inter-coder reliability was .952, with strong agreements within Purpose (.969), Mechanism (.949), and Subject (.938).
| hillman@cantab.net | 1997 |