| Carl Chatfield home page | Thesis | Chapter I |
When a technical writer produces material for a reader, who is the writer really writing for? Traditional technical communication textbooks suggest that through quantitative, demographic audience analysis, the careful writer can craft his material for a specific target audience. Houp and Pearsall, authors of a popular undergraduate technical communication textbook, stress the quantitative nature of audience analysis:
You must understand not only the purpose but also the background of your audience. You must know who your readers are, what they already know, and what they don't know. You must know what your readers will understand without explanation and without definitions. (Houp and Pearsall, 1984, p. 20)
The authors of another popular textbook are even more specific about the need to quantitatively understand one's audience:
The more you know about the groups of people within your audience (usually you write for several groups), the more you can shape your prose so they understand you, and the easier it will be for you to organize your information so they can find what they need. Your understanding of your audience can determine whether your manual succeeds or fails…You need to consult with anyone who can tell you about your audience. Then you can prepare an audience profile. (Price and Korman, 1993 p. 29)
In the traditional view of technical communication theory, readers may be known through demographic details such as average years of experience with the subject product, or generalized into a composite reader, for example, the "intermediate" user of Microsoft Word, whoever that may be. This generalizing of potential readers into definable categories is a notion easily understood by most technical communicators, but often implemented with great difficulty. For all generalizations there are exceptions, and many readers find themselves not fitting within the categories intended by authors. To make matters worse, many technical communicators do not provide adequate cues to indicate the category of reader for which a specific text or passage was written.
A very different way to consider audience analysis, however, is in terms of rhetorical author and reader roles. Theories of author and reader roles from literary analysis (Gibson, 1950; Iser, 1974; Ong, 1975) are now having an influence on technical communication theory (Coney, 1984; Simpson, 1989; Goodwin, 1991). Rather than (or in addition to) gathering demographic information about actual readers, the author can develop rhetorical constructs, or roles, for both himself and his reader to assume within the text. For example the author can develop the author persona of the "helpful mentor" and the reader role of the "eager learner." The author can also provide cues for specific reader profiles within a manual. Unlike the quantitative approach to audience analysis, these reader profiles or roles are not descriptive of the actual reader, but of a role the author would like the actual reader to assume while reading the text.
Authorial personas and reader roles are one manifestation of the rhetorical differences between primary and secondary computer manuals. Authors of both types of manuals should consider these rhetorical differences carefully. Consider this: authors of primary manuals work within the software development company and therefore should have better access to software developers, product specifications, and early prototypes than would authors of secondary manuals. In the case of the two manuals examined in this thesis, The Microsoft Word User's Guide (a primary manual) and Word for Windows 6 for Dummies (a secondary manual), the primary manual is well over twice the length of the secondary manual. It generally provides more breadth and depth in documenting Word features and tasks than does Word for Windows 6 for Dummies. Despite these facts large numbers of Word users are willing to spend an additional $16.95 to purchase Word for Windows 6 for Dummies. In fact Word for Windows 6 for Dummies is the best-selling secondary manual about Word. In addition to the number of Word users that purchase Word for Windows 6 for Dummies, far more spend comparable amounts on other secondary manuals. What is the lure of these secondary manuals? I propose that in the case of Word for Windows 6 for Dummies at least, rhetorical style and not factuality or comprehensiveness is a major contributor to the success of this secondary manual.
One way to start defining primary and secondary manuals is to categorize them as instruments of discourse. A good starting point is offered by Kinneavy (1971). Kinneavy identifies four types of discourse: reference, persuasive, literary, and expressive. Books such as The Microsoft Word User's Guide and Word for Windows 6 for Dummies fall into the reference category of discourse in general, and into the informative reference category in particular. Kinneavy defines the basic constituents of informative discourse as factuality, surprise value, and comprehensiveness. Factuality and comprehensiveness are the most relevant constituents of software manuals.
Factuality describes the truthfulness of the discourse. Factuality can be verified by empirical methods or by logical methods. For example, the following procedure from The Microsoft Word User's Guide can be empirically verified by following the steps and observing the results:
To print a document to a file in Word for Windows
Or, the following passage from Word for Windows 6 for Dummies can be logically verified by evaluating its internal logic (that is, the passage does not contradict itself):
Finally, your printer must be on-line or selected before you can print anything. Somewhere on your printer is a button labeled On-line or Select, and it should have a corresponding light. You have to press that button to turn the option (and the light) "on." Although your printer may be plugged in, the power switch on, and doing its warm-up stretching exercises, it doesn't print unless it's online or selected. (Gookin, 1994, p. 92)
Factuality is so essential to informative discourse that it is heavily stressed in many basic academic books on technical writing. When referring to factuality of content, Price urges the prospective technical writer to "Check it again and again and again. Have experts check your facts" (Price, 1984, p. 9).
Comprehensiveness is also an essential element of informative discourse. Ideas and procedures must be fully developed to be useful. However determining a manual's level of comprehensiveness is a subjective act of the author. For example, the authors of The Microsoft Word User's Guide apparently had a different notion of comprehensiveness than did the author of Word for Windows 6 for Dummies, as their relative differences in size (830 pages versus 370 pages) reveal.
Manuals such as The Microsoft Word User's Guide and Word for Windows 6 for Dummies must be factual and comprehensive to be useful. However, factuality and comprehensiveness alone cannot account for the total reader experience with either manual, or with other task-oriented manuals. Nor can the goals of factuality and comprehensiveness fully inform the would-be author of such manuals. A guideline such as "check and recheck your facts" even when faithfully followed, can still result in a document that is not usable to many readers. Considering these seemingly informative manuals as instruments of persuasive discourse, however, may lead to a more useful understanding of such manuals.
Harris provides an excellent theoretical basis for understanding technical documents such as software manuals as rhetorical documents. By considering how-to discourse as persuasive in nature, Harris is able to apply techniques of rhetoric to what she calls "how-to texts." Harris notes that the three main techniques of persuasion are arguments from character, emotion, and logic (Harris, 1983). How-to discourse is a form of persuasive discourse that depends most on logical arguments, somewhat on the character of the author, and least of all on emotion. How-to texts argue by induction. In this essay I will consider how-to texts such as The Microsoft Word User's Guide and Word for Windows 6 for Dummies to be persuasive documents and make a comparative critique of them in terms of rhetorical factors.
How-to texts are often used in place of interaction with other human beings (Bolter, 1991). When a Word user consults Word for Windows 6 for Dummies, for example, one might think of the book as a proxy for Dan Gookin, the Word expert who wrote the book. If the Word user had direct access to Mr. Gookin, that might be the preferred means of obtaining information about Word (1). But access to experts is often limited or impossible, so the printed representation of Mr. Gookin's knowledge of Word serves the reader instead.
How-to texts exist to instruct their readers in the performance of some physical or intellectual task. An action on the part of the reader is implicit in every how-to text. While many types of texts might motivate a reader to attempt or accomplish a certain task, only how-to texts use specific rhetorical and structural elements to facilitate the reader's performance. For example a great work of literature such as "A Christmas Carol" might inspire a certain kind of behavior in a reader, generosity, for example. The text itself, however, does not require this behavior from the reader to be considered successful. The text can be read and enjoyed on a number of levels as literature. Asking the question "Is the text intended by its author to help the reader accomplish the task described?" is a litmus test that distinguishes any how-to text from all other texts. While other types of texts may help a reader accomplish a task, achieve certain goals, or change behavior, only how-to texts are explicitly designed to do so.
How-to texts are technical documents but also persuasive documents. Their very existence suggests that a reader faces a task that is beyond his unaided capabilities. The fact that the author chose to write the instructions itself implies some level of empathy. The writer has said, in effect, "here is a problem to be solved (by the reader), and I will help." No how-to text simply informs. It calls the reader to perform a certain action or hold a certain conviction. An action might be assembling computer components in a particular order. A conviction might be to hold a mental model of the mechanics of a spreadsheet grid. In either case, the reader is moved towards the author's perception of some particular aspect of the world: that a computer system should be assembled in a specific way; that one should understand (and therefore use) spreadsheets in a specific way. As instruments of persuasive discourse, how-to texts are intended by their authors to enable their readers to perform certain actions that they have already decided they would like to do, but lack the information or experience necessary to do.
Another distinguishing factor of how-to texts is that they are usually about other things. These other things may also be means to another end. For example, one might read a passage in The Microsoft Word User's Guide to successfully use Word's Mail Merge feature. But using the Mail Merge feature is itself a means to another end-sending form letters to ones' customers, for example. How-to texts are about something else; they are not self-referential.
To sum up, how-to texts have qualities that distinguish them from other informative texts. How-to texts:
In this section I will build a model of author/reader interaction that is based on work from literary analysis. A simple model of speaker/listener interaction can be represented like this:
Figure 1: A simple model of speaker/listener interaction.
This interaction model can be applied to both verbal and written communication. In the case of verbal communication, the listener is in front of the speaker. In the case of written communication, the listener is simply further away from the speaker in time, space or both.
This model forms the basis of positivist thought. For example, if nature is thought of as the speaker and a scientific observer as the listener, this model presents the most pristine model of observation. One would think such a straightforward representation of information transfer would also serve as a sound basis of technical communication: the author transmits information directly to the reader without embellishment or distractions. This is in fact the common wisdom of the field of technical communication.
Ong (1975) refers to this representation as an "alluring but deceptively neat and mechanistic mock-up" (p. 10). The shift from this simplistic understanding of reading to a more meaningful understanding has been referred to in the following way.
Change that goes from seeing reading as straight information retrieval to seeing it as a process dependent on and subject to readers' models of reality (in the technical jargon of reading, "schema"), which are mental configurations or maps built from prior knowledge, feelings, personality, and culture which readers then apply to, or impose on, new experiences. (Petrosky, 1982)
When properly handled by the author, rhetorical factors such as author and reader roles can help the reader to invoke pre-existing or develop new schemas.
In much the same way that written literature has replaced oral storytelling in the Western world, written how-to texts have replaced oral instructions that passed down information essential to the performance of many tasks. Harris (1983) notes that " 'How to' writing has replaced oral traditions about how to plant potatoes, staunch bleeding, and drive mules" (p. 140). Ong (1975) traces the history of rhetorical analysis of literary texts back to oral traditions. Ong observes that while the relationship between orator and listener (or more commonly an audience of listeners) is well-defined and real, the relationship between writer and reader is always abstract. Iser (1980), a literary critic and constructivist, notes that unlike a speaker, a text cannot adapt itself to each reader it comes into contact with. Although an orator can (Ong, Gibson, and Coney would argue always does) assume a rhetorical role or mask, the interaction between orator and listener takes place within a live setting. Spoken words exist always and only within a known or knowable context. "Context for the spoken word is simply present" (Ong, 1975, p. 10). The writer, on the other hand, often does not know who will read his work, when it will be read, or under what circumstances. To address this aspect of written discourse, Ong proposes that the writer can invent a fictitious reader role.
The fictitious reader role is created by two acts. First, the writer must create a role of some sort within a text through various structural and stylistic cues. But the writer's act alone does not create the fictitious reader role. The actual reader, in the act of reading the text, must then step into the role cast for him by the writer in much the same way an actor assumes a character role. Here is a graphic representation of Ong's model:
Figure 2: Ong's writer/reader interaction model.
Others have built on Ong's model. Park (1982) looked specifically at the meaning of "audience" for a writer. He concluded that "audience" refers to the perspective of knowledge and motivation the writer sees in readers, and therefore the embodiment of this perspective allows for a context of the discourse between writer and reader. The essential difference between the notions of audience and reader is that the audience is a set of suggested or evoked attitudes, interests, reactions and conditions of knowledge implied in the text, while the reader is the actual person external to the text.
The writer's intent is only part of the reader role, however. The actual reader brings the reader role to life. When referring to the reader role, Coney notes that "This reader differs from the actual audience in that it is a role that must be assumed, a set of values, attitudes, biases, even facts that must be known and accepted if the text is to be read and understood" (Coney, 1987, p. 323).
Ong (1975) identified the following reader roles after analyzing the style of Hemingway and Time:
| Reader role | Description | ||
| Hemingway reader | Close companion of the writer. | ||
| Time magazine reader | Omniscient observer of world events. |
Ong identifies the "close companion" role in the opening passage of Hemingway's A Farewell to Arms:
"The late summer of that year," the reader begins. What year? The reader gathered that there is no need to say. "Across the river." What river? The reader apparently is supposed to know. "And the plain." What plain? "The plain"-remember? "To the mountains." What mountains? Do I have to tell you? Of course not. The mountains-those mountains we know. We have somehow been there together. Who? You, my dear reader, and me. The reader-every reader-is being cast into the role of a close companion of the writer (Ong, 1975, p. 13).
Hemingway uses, among other cues, the definite article "the" and the demonstrative adjective "that" to include the reader not only in the scene but in the personal history of the narrator. Or, put more generally, Hemingway creates a reader role that is familiar with the surroundings and background of the narrator.
One can imagine a wide range of reader roles for other texts. Some examples are:
| Reader role | Description |
| The mystery novel reader | Crime-solving sleuth competing with the declared sleuth within the text to solve the mystery. |
| The New Yorker reader | Member of the social elite. |
| Fortune reader | Take-charge corporate mover and shaker who seeks to improve his position. |
According to Gibson (1950), a primary responsibility of any magazine editor is to define the magazine's mock reader role. Others have developed reader role taxonomies for readers of how-to texts, which I will explore shortly.
Ong refers to a rhetorical model established by the literary analyst Gibson. Gibson begins with the author and reader of a text, and inserts between them the speaker and mock reader roles (Gibson, 1950). The speaker is the narrator of the text, or in some cases, an additional voice within the text but separate from the overt narrator. The mock reader role is that role within the text that the speaker role addresses. Here is a graphic representation of my developing interaction model after considering Gibson's speaker and mock reader roles:
Figure 3: Gibson's speaker/mock reader interaction model.
Gibson defines the literary experience as the relationships between the actual writer, the speaker role, the actual reader, and the mock reader role. From the perspective of the reader, a bad literary experience can be thought of as a role one cannot or will not take on. Or put more broadly, "one of the ways to describe misunderstandings or misreadings is to say that the writer's view of the reader and the readers' view of self at any particular time or place in the reading process don't square with one another" (Coney, 1992, p. 62).
The value of Gibson's refinement to the author/reader model presented in Figures 1 and 2 is that he makes clear the distinction between the author of a literary text and the speaker within the text. While the author is a living and complex set of attitudes, beliefs, and intentions, the speaker is made up only of the descriptions and stylistic cues within the text. In fact the speaker comes to life only in the act of reading. In technical communication, the speaker is more often than not absent or minimized to a corporate "we." Some secondary texts such as Word for Windows 6 for Dummies provide exceptions to this general rule.
Other literary analysts have commented on the interaction between reader and text. Iser sees the sequence of sentences within a text as no more than component parts of the text (Iser, 1974). Iser argues that the implied reader role comes into play when "the reader is able to 'climb aboard' the text. He has to accept certain given perspectives, but in doing so he inevitably causes them to interact" (Iser, 1974, p. 277). As a constructivist, Iser argues that the text takes on its full existence only within the reader's mind.
One way to understand the implied reader role is as the mechanism through which the actual reader makes meaning of a text. "The fact that completely different readers can be differently affected by the 'reality' of a particular text is ample evidence of the degree to which literary texts transform reading into a creative process that is far above mere perception of what is written" (Iser, 1974, p. 279). Contrast this perspective with the simple model of author/reader interaction shown in Figure 1. Although Iser's work deals with literary texts, his perspective of interaction between text and reader is applicable to how-to texts. Iser sees the reader filling in "gaps" of imagination within literary texts. "It is the gaps, the fundamental asymmetry between text and reader, that give rise to communication in the reading process; the lack of a common situation and a common frame of reference corresponds to the 'no-thing,' which brings about the interaction between persons" (Iser, 1980, p. 109). It is within these gaps that the implied reader lives. Iser sees these gaps as the pivot points on which the whole text-reader relationship revolves. Iser more fully develops an explanation of what actually happens as one reads: "As we read, we oscillate to a greater or lesser degree between the building and the breaking of illusions. In a process of trial and error, we organize and reorganize the various data offered us by the text" (Iser, 1974, p. 288). This may be a very apt description of the mental process of a reader of an how-to text as well as a literary text.
The constructivist perspective such as Iser's is especially helpful when considering how-to texts because it focuses on the active role of the reader in realizing the meaning of a text. Perhaps in no other written genre do readers seek meaning from the text to the extent they do in how-to texts.
Researchers have applied literary analysis models of author/reader interaction to how-to discourse. For example, this interaction model has appeared in research that was not specifically looking for rhetorical roles within texts. One telling example comes from Walters and Beck (1992).
While comparing primary and secondary computer software manuals, Walters and Beck found rhetorical effects in secondary manuals that are similar to those described by Gibson and Ong. Walters and Beck found that while both primary and secondary manuals contained comparable information to support learning, the secondary manuals took a decidedly different rhetorical tone.
Walters and Beck found that authors of secondary texts used rhetorical devices and strategies to establish a different kind of relationship with the reader than did the authors of primary texts. Specifically, the secondary text authors attempted to create a stronger bond with the reader, which Walters and Beck argued creates a richer context for learning (Walters and Beck, 1992). Although Walters and Beck did not explicitly cite the work of Ong or Gibson, their findings fit very well into the author/reader role models Ong and Gibson have developed. I will review Walters' and Beck's rhetorical factors in Chapter II, and analyze them in two how-to texts in Chapter III.
An important aspect of this author/reader role model is its implications for authors. The speaker or implied author role allows an author to reinvent himself within a text and even within individual passages. "The author creates, in short, an image of himself and another image of his reader; he makes his reader, as he makes his second self, and the most successful reading is one in which the created selves, author and reader, can find complete agreement" (Booth, as quoted in Coney, 1987, p. 323). Although Booth was referring to literature, the value of an implied author for the real author can be applied to how-to texts.
When analyzing the content of early microcomputer hardware manuals, Halse (1986) found that technical writers make assumptions about the capabilities, values, and status of their readers. These assumptions were so consistent through the manuals Halse examined that he argues that the rhetorical handling of readers by technical writers is becoming formalized as a genre. Halse found that the writers of the manuals he examined were compelled to establish a strong ethos in the text. Halse writes that "overall, the writers assume that they should be a friend, or, perhaps, a trusted 'professional' guide. At times, the writer may take on the role of teacher or confidante or buddy; at other times, the writer may be the disciplinarian or lord" (Halse, 1986, p.114). Similarly, Goodwin (1991) argues that the authors of all technical writing strive to create authorial roles that are immediately appropriate, interesting and pertinent. Others have noted that the reader will judge the writer's character by the writer's knowledge of the task, and his empathy with the reader's problem of performing it (Harris 1983).
Coney finds a different kind of value in the ethos or implied author within texts, and especially within how-to texts. The implied author is able to convey the referent, or meaning of a text without calling linguistic attention to himself, as a real author must (Coney, 1984). Or, put another way, an implied author can keep the reader's focus on the subject and not on the author. Thus the implied author can serve a linguistic as well as rhetorical purpose within a text .The effect of the implied author and mock reader roles in the interaction model developed above can now be represented like this:
Figure 4: The implied author/mock reader interaction model.
In this interaction model, the implied author mediates between the author and the text linguistically, and between the author and the reader rhetorically.
Theorists and practitioners have developed various taxonomies of reader roles specific to technical content. I will summarize two such taxonomies here. The first taxonomy I will summarize is that of Simpson (1989). Simpson presents a list of reader roles based on meta-analysis of usability test data.
| Reader role | Description |
| Doer | Supports completion of tasks external to the text. |
| Learner | Supports concept building in the reader's mind. |
| Searcher | Supports reader activities to locate information within a text. |
Simpson identifies three types of cues that can invoke a specific reader role. These cues exist at the sentence level, the passage level, and the document level.
Sentence level cues identify the agents in a passage. For example, the following passage from The Microsoft Word User's Guide directly addresses the reader as the active agent in the text: "To adjust the margins in a document, you can use the rules or the Page Setup command on the File menu" (Microsoft Corporation, 1994, p. 231). In general, when "you" the reader is the agent, the sentence invokes the doer role. This is true whether the reader is directly addressed as in the above passage, or implied as is more common in how-to texts. When the computer system or other technology is the agent, the sentence invokes the learner role. For example, in the following passage Word is the agent: "Margins determine the distance between the text and the edge of the paper. Word normally prints text and graphics inside the margins, while headers, footers, and page numbers are printed in the margins" (Microsoft Corporation, 1994, p. 230).
Passage level cues are about organization of content of a text. Passage level cues can invoke either the learner or doer role. For example, The Microsoft Word User's Guide contains a fairly consistent structure of general concept and task descriptions followed by detailed descriptions and procedures. Here is how one section (pp. 230-234) is outlined, with the overall function of each component indicated in parentheses. This passage-level organization displays a conceptual-to-procedural pattern and suggests a learner role for the reader.
Simpson's final type of cue is found at the document level. Document level cues can invoke any role. Document level cues that invoke the searcher role include elements such as tables of contents, headings, indices, and running headers. The learner role is activated by blocks of text and most art, and the doer role by task-oriented text such as enumerated procedures.
Simpson notes that in the software documentation he analyzed, cues did not uniformly invoke the roles intended by their authors. Simpson identified two types of user behavior, which he called text-dominant and task-dominant. Text-dominant behavior follows the rhetorical theory of cues: the cues invoke intended reader roles. "Text-dominant" is an appropriate label for this behavior, as the text leads the user through a sequence of reader roles.
The more commonly observed behavior, however, was task-dominant behavior. In this case, the user's context determines how they respond to cues within the manual. Simpson found that readers sometimes circumvented the intentions of the authors. For example, readers correctly identified cues that invoked the learner role, and simply avoided those passages. Simpson concluded that users' tasks invoke specific roles, which the text supports after-the-fact. Simpson's work suggests that, with how-to documents at least, the reader's context and not the author's intent is paramount in deciding the success or failure of reader roles.
The second taxonomy I will summarize is from Coney (1992). Coney identifies five reader roles. They are:
The reader seeks knowledge, but has no influence on the substance of the message. This view of reader is derived from logical positivism, and requires the writer to gather measurable, observable data about his intended audience. The rhetorical relationship between writer and reader is irrelevant. This reader role is illustrated in Figure 1.
The reader is goal-driven and generally not willing to read conceptual or other information that does not directly support the task at hand. This view of reader prescribes minimal texts that serve only as a means to an end. The rhetorical relationship between writer and reader is one of tutor and pupil. The implicit reader role in most software manuals is reader as user.
The reader possesses a special code that allows him to decipher meaning from a coded message, for example in a technical field. The rhetorical relationship between writer and reader is one of mutual exclusivity. Both parties are fluent in the coded language of the field.
The reader is a member of a professional community that builds knowledge through debate and consensus. The rhetorical relationship between writer and reader is one of mutual respect or parity.
The reader determines the meaning of a text. The rhetorical relationship between writer and reader heavily favors the reader's interpretation over the author's intent.
Coney notes that a single reader may play a series of roles within a single text, and that this taxonomy is not exhaustive. Coney's taxonomy of reader roles is helpful because it reflects a range of philosophies from logical positivism (reader as receiver of information), to structuralism (reader as decoder), to the New Rhetoric (reader as professional colleague) and finally to constructivism (reader as maker of meaning). In the case of how-to texts such as The Microsoft Word User's Guide and Word for Windows 6 for Dummies the dominant reader role is clearly reader as user. In fact all three of Simpson's reader roles (doer, learner, searcher) really fall under the category of reader as user in Coney's taxonomy. This should be no surprise considering the very strong reader-as-user focus evident in nearly all modern software documentation. Not surprisingly, the technical writing textbooks focus on reader as user. The advice of one popular technical writing textbook is clear on this point: "Keep the focus on their goals-not yours…These people see the computer and your particular product as an appliance, not an opportunity for 18 hours of tinkering, troubleshooting, and debugging" (Price, 1984, p. 32).
The pragmatic value of reader role taxonomies to technical writers may be to help them expand the meaning of audience analysis. Empirical and quantitative data collection about real readers will no doubt continue to inform technical writers about appropriate scope, style, and other factors. However adding the dimension of authorial and reader roles to the technical writer's domain can help both reader and writer achieve their goals. This is evident in the two how-to texts Microsoft Word User's Guide and Word for Windows 6 for Dummies examined in Chapter III.