(Updated 4/3/13 with correction regarding Schank and Abelson story about Dr. Schank’s daughter.)
This is the third post in a series exploring a possible cognitive theory of genre. The first, appearing here, covered some basic concepts from contemporary cognitive science. The second focused on prototype theory, script theory, and relevance theory. You’ll find this post easier if you read the previous ones.
I should note at the outset that what I propose here as a “cognitive theory of genre” is not fully theorized or set into the context of all the genre-related research that has preceded it. (That’s a euphemistic way of saying that this is all half-baked.) Indeed, these are really just my first musings on what cognitive genre would look like and how it fits in with the small number of texts on my reading list. I also do not intend cognitive genre to replace or supplant any of the existing traditions of genre study. Instead, it is a call for an additional focus on the cognitive aspects of genre knowledge.
In keeping with the tentative nature of this exposition of cognitive genre theory, I’m offering my suggestions and observations in the form of a potpourri of large-scale issues. Thus, I will take up a focus on agents and texts, arguing that a cognitive theory of genre will focus on individual agents and away from social constructionism, and acknowledging the focus on texts that genre studies have always exhibited. We will consider whether cognitive science and relevance theory can offer experimentally verifiably explanations of key concepts that genre theory has relied on since its inception: typification and the stability/dynamism balance. And we will consider how genres function in meaning-making and how cognitive theory might secure a role for genre theory in writing pedagogy.
By the end, it will be clear that I’ve left an elephant in the room, which I will treat in the section on de-reifying genre.
Focusing on agents and texts
The cognitive theory of genre focuses on individual agents and the texts (broadly construed) and contexts with which they interact. It emphasizes efforts to understand cognitive processes that allow agents to identify situation or episode “tokens” that can or should be treated as “types” of situation; to identify characteristics of communicative tokens by others as belonging to generic types; to apply knowledge of generic types to generate tokens that are suitable for the type of situation or episode in which the agent finds herself; etc. Cognitive genre also recognizes the importance of formal characteristics of genre; if there were not common formal elements in communicative tokens, we would have no occasion to describe types of them as “genres.”
(An aside: I’m tentatively using the token/type distinction here; it comes from the linguistic and philosophical traditions. The “token” is an instance of the “type.” Thus, a particular communicative performance by an agent in a real situation is a token. So, for example, this prelim exam answer is a token of the type “prelim exam answer.” Swales refers to genre “exemplars,” which would generally correspond to tokens, but in my approach, each copy of a text would function as a separate token with potentially different cognitive semiotic effects. Note, therefore, that “text type” here is not a reference to earlier conceptions of that term in its relation to register and genre. Here “text type” just means a category of text tokens.)
The turn to the individual is not unexpected. Existing scholarship acknowledges that individuals are the sources of the communicative tokens that represent and reinforce genres. Devitt acknowledges, for example, that “individual perception must be the source of recurrence, for discourse exists only through the actions of individuals” (20-21). However, despite my proposed focus on the individual agent, I do not favor the approach, embraced by Swales (1990) and Devitt (2004), that privileges the genre nomenclature of genre users. Devitt writes, “the most significant genre labels for a rhetorical definition of genre—and the classifications of most concern to rhetorical genre scholars—are the labels given by the people who use the genres” (8), the implication being that users recognize genres as such. My intuition is that this is not true, in the same way that the average native speaker of a language is not the best informant regarding its grammatical categories and structure. Of course, the researcher needs to understand how users of a genre interpret it as a genre; but the researcher’s obligation in my view is to understand what the agents are doing, not just what they think or say they are doing.
The focus on text has always been there, too, but some genre theorists appear to have unease about addressing texts’ formal characteristics. This is more overtly pronounced in the research of those with a strongly socio-rhetorical emphasis (Miller, Devitt, etc.) and less so in the research of those with an overtly linguistic background or focus (Swales, Bhatia, Berkenkotter and Huckin). Nevertheless, form appears to matter—a lot. Devitt (2004) identifies what she calls “subgenres” in the work of an accounting firm based on formal characteristics that the users of the genres themselves do not recognize. Even Miller (Miller and Shepherd, 2004) falls prey to excessive formalism, classing “blogs” as a genre based largely on formal characteristics and some conclusory cultural criticism. She later recants (Miller and Shepherd, 2009).
Among the theorists on this reading list, only Bhatia (1993) and Berkenkotter and Huckin (1994) plausibly claim an individual cognitive orientation to genre. But Bhatia does not identify cognitive techniques, and though Berkenkotter and Huckin refer loosely to situated cognition throughout their work, they reserve the most detailed account of it for the discussion of genre theory’s use in pedagogy. (More on that later.)
I expect to pursue this effort with the methods previously favored by genre theorists: expert interviews; participant interviews, including discourse-based or modified discourse-based interviews; think aloud protocols; in situ observations, whether properly characterized as ethnography or not; and linguistic and rhetorical analysis of text tokens. Many of these methods respect the social and situated nature of communicative behavior. Some of them have a long tradition in cognitive science. For example, Newell and Simon (discussed above) invented the think-aloud protocol as an effort to elicit cognitive process data from participants, and it is still used in cognitive studies today. However, in addition to these methods, I would propose to use experimental and quasi-experimental approaches to test whether certain theoretical claims have cognitive correlates. I would also use corpus linguistic techniques to understand tendencies in the communicative tokens that represent genres. Corpus methods might allow the researcher to inductively identify genres where even the users/agents do not recognize them; they may also prove valuable for identifying formal characteristics that are effective across different communicative activities.
Units of analysis that would be deemphasized in cognitive genre are “groups,” “communities,” “societies,” etc. To start with, these concepts are methodologically complicated in any but the simplest circumstances. For example, Devitt (2004) defines a “community” as a group “of people who share substantial amounts of time together in common endeavors” (p. 42; for her, this entails Swales’, 1990, “discourse communities”); but she acknowledges that porous membranes bound communities, so it’s not always clear who is in them and who is not. She also defines a “collective” as lacking the “frequency or intensity of contact of a community” (44), but she does not articulate any way of operationalizing such a distinction. Berkenkotter and Huckin (1994) too acknowledge that discourse community is “a slippery proposition because neither concept refers to a static entity” (21).
It is possible to operationalize the concept of “group” cognitively, according to the perceptions and mental representations of group members. Thus, each putative member of a group has his/her own sense about who is in-group and who is not. That knowledge is important from a cognitive perspective, because strong in-group/out-group heuristics bias group members’ thinking about other individuals (see Gigerenzer & Brighton). These understandings can relatively easily be documented by the researcher using social network analysis (see, for example, Hansen et al., 2010; Kaufmann, et al., 1976). (This claim greatly oversimplifies the process of using network analysis in a case like this. But I don’t have the time or space to elaborate on the issues here.)
Situated (re)cognition: Explaining typification
Genre theory has so far done too little to understand the workings of “typification.” Miller’s (1984) offering adopts the theoretical (and not empirically grounded) typification theory of German sociologist Alfred Schütz. This theory emphasizes the potency of language and social practices in typification. We can illustrate this with an anecdote from Bawarshi and Reiff: at a music festival, Bawarshi’s daughter sees a little boy in a princess costume; she refuses to accept that the boy is not a girl, because of the gender convention (princess costume) that he has adopted. Children clearly see gender conventions enacted every day, but does this illustrate how knowledge of communicative types arises in individuals? Consider a contrasting anecdote about a car purchase from Schank and Abelson (1975), recounted by Roger Schank
Robert Abelson(see Dr. Schank’s comment below, which motivated this correction):
My daughter Hana (age 4) was with me when we bought it and asked if I was going to get a new key chain. I asked her what she meant. She replied that when we had gotten our old car in Rhode Island (where it had arrived off the boat 2 years earlier) I had bought a new key chain. This was her only experience with getting a car and already the events in it were a script for her. (p. 68)
What cognitive processes resulted in Hana typifying the one experience she had previously had of buying a new car? What processes resulted in her picking out the corresponding acquisition of a new key chain as part of the script for getting a new car? There are, of course, analogous questions in genre theory, genre theory should seek the answers to these questions, and cognitive methods should be able to help find them.
For example, Kirsh’s (2009) conception of situated cognition may be important here. He discusses affordances, noted in a previous post, which are characteristics of the environment that an agent actually perceives as giving her a “handle” on the environment. In other words, her notions of the possibility of use of the environmental characteristic affect her perception of the characteristic. He says that a situational theory of affordances should explain why some people see affordances in a situation where others do not and why. In the case of typifying social interactions, like purchases, it should be possible to identify those features of an interaction most likely to be recorded as part of a script. Kirsh’s ideas may also have importance for genre pedagogy, as discussed below.
We should also study how agents assess the typicality of text tokens, that is, (in Swales’ terms) how they recognize exemplars of a genre. Formal characteristics play a role here, but which formal characteristics play which roles? Consider Figure 1, the “caption” from a memorandum submitted by one party to litigation opposing a motion filed by the other party to the litigation. Almost any document filed with the court (and many that the court will never see) in this litigation will include a caption in almost identical form (except the title of the document, appearing in the panel on the right). This caption is a token of a type of document component but is it a generic convention, and if so, how do users recognize it as such?
Figure 1: Caption from court memorandum
Rosch and Mervis (1975) offer a processing model they call cue validity “in which the validity of a cue [i.e., a characteristic’s comparative value in distinguishing types or categories] is defined in terms of its total frequency within a category and its proportional frequency in that category relative to contrasting categories” (575). In other words, the important characteristics for distinguishing types are those that appear frequently within a category and infrequently outside of it. They acknowledge that cue validity is not the only factor in forming a prototype: “the frequency of items and the salience of particular attributes or particular members of the categories (perceptual, social, or memorial salience) as well as the as yet undefined gestalt properties of stimuli and stimulus combinations, undoubtedly contribute to prototype formation” (599). But research of this kind provides the outline of a framework for studies that should be done in the genre-theory context to understand the process of typification.
The next post will consider a couple other regions of genre theory that should receive the attention of a cognitive perspective.