[as published in TEXT 6:1 (1986), 53-71]

Abstract

Evaluations are the means by which tellers of conversational stories, and, by analogy, narrators of literary texts, convey their beliefs, values, and attitudes. From the listener's or the reader's point of view, evaluations are usually signalled by a sense of incongruity between a given element and the local norm of the text. We distinguish three types found in literary narratives: "discourse evaluations" which can occur when something is expressed in a distinctive way; "story evaluations" which can occur when something unexpected happens in the storyworld; "telling evaluations" which can occur when the fact that the narrator mentions something at all, or mentions it at a particular moment, is surprising. Texts appear to have redundant patterns of evaluation, and thus core perceptions tend to be shared among authors, narrators, and "point-driven" readers. A modern short story is used to illustrate some of these ideas.

When stories are told in conversation, we usually know how to hear them. Consider, for instance, this exchange between an adult female investigator and a five-year old girl:

Adult: Have you ever gotten jabbed with anything?
Child: Uh huh. I got jabbed by a bee.
Adult: By a bee. Oh, tell me about it.
Child: See, I got jabbed on my foot. I was barefooted. I screamed and I screamed and I cried and I cried. I screamed and I screamed. Until my next door neighbor came out and my Dad came out and my brother came out. And, they all carried me into the house but after that happened I got to sleep overnight with my neighbor. (Peterson and McCabe, 1983: 30).

The reading of literary texts can have important similarities to conversational story listening. Consider, for example, a passage in Hemingway's "Indian Camp." A doctor has been summoned to treat an Indian woman who is very sick. Two Indian men take the doctor, along with his brother George and his young son Nick, to the woman's cabin.

Inside on a wooden bunk lay a young Indian woman. She had been trying to have her baby for two days. All the old women in the camp had been helping her. The men had moved off up the road to sit in the dark and smoke out of range of the noise she made. She screamed just as Nick and the two Indians followed his father and Uncle George into the shanty (I 953: 92).

Thus, literary texts, like conversational stories, can be treated as if their tellers were extending invitations to share their attitudes, values, and beliefs. The consequence is that readers have the opportunity to construct a version of the narrator's "points," to develop a model of what the author might be "getting at." (Most readers of Indian Camp, for example, have understood Hemingway to be getting at something with more powerful implications than a Caesarean section performed under difficult conditions.)

In suggesting that there is a distinction to be made between, on the one hand, the story that is told, and, on the other, what its narrator's purpose in telling it apparently is -- whether we identify that purpose as "theme," "point," or "meaning" -- we are of course saying nothing that hasn't been said many times before. What has not been remarked upon so often, however, is that it is possible for a reader to learn how to read a passage, to understand what an author or a narrator is getting at, in ways that are directly parallel to the ways by which listeners know how to understand conversational stories.

We contend that narrators of literary stories use strategies similar to those of narrators of conversational stories to achieve their purposes. Further, we believe that a detailed consideration of just how they do so helps us to describe literary reading more exactly than has been possible in the past. In particular, it enables us to conceptualize more clearly some of the processes by which different readers regularly make similar sense of a literary text. Our purpose here is to explore in some detail how this might work.

The ways by which narrators of conversational stories attempt to share their attitudes, beliefs, and values -- in a word, their "points" -- have been termed evaluations by sociolinguists such as William Labov and Livia Polanyi. To demonstrate that this term is useful in understanding literary reading as well, we begin by examining the ways in which the concept is used in the study of conversational storytelling. We then present a typology of evaluations in literary texts, in which we distinguish "discourse," "story," and "telling" types. Finally, we give detailed examples of each of these types, as well as some problematic instances, by means of an analysis of a literary text, "The Day We Got Our Own Back," by the Irish writer Maeve Brennan. Our concern throughout is to show that the concept of evaluations is a productive one, whether one's ultimate concern is with textual or stylistic analysis or -- as ours is -- with reading.¹

2. Evaluations in conversational narratives

According to recent studies of conversational storytelling (Goffman, 1974, chap. 13; Labov, 1972; Polanyi, 1985; Schiffrin, 1984), when narrators tell a story in face-to-face interaction, they normally do at least three things. First, they recount events: they tell who did what. Second, they situate the events in spatio-temporal and motivational contexts: they explain where, when, and why. Third, they provide evaluation: they make clear the reasons why they are telling this story in the first place. By conveying their attitudes and beliefs, narrators thereby attempt to share them with, and to shape those of, their listeners.

It is easy to see why events and background material are necessary parts of narratives, but what about evaluation? The answer seems to be that without evaluation any sequence of events, even a contextualized sequence, is potentially understandable in many ways. In order to make it more likely that the audience will share his/her view of why the story is tellable, a narrator must provide cues as to how the story is to be taken, how it relates to the ongoing conversational and pragmatic situation of which it is a part. In this view, evaluation is one specific mechanism by which narrators convey their "points."

As William Labov has shown (1972: 370-393), narrators have available a variety of evaluative devices.. In external or overt evaluation, the narrator says more or less directly, from outside the storyworld, "This is it -- this is the important part." Less overtly, the narrator may attribute an evaluative comment or thought to a character, or, less directly still, convey evaluative force by means of an action (e.g. "I screamed"). Even suspension of action can be evaluative because it heightens the audience's anticipation, making the crucial action or "high point" seem even more striking when it does come. In general, Labov considers departures from normal narrative syntax to be evaluative. He lists, in fact, no fewer than twenty-nine separate syntactic devices (including intensifiers, negatives, comparatives, and so on) that narrators can use as evaluators (1972: 394). Thus, for Labov, whatever departs from "the unmarked case" -- that is, whatever is "marked," surprising, or unpredictable -- is potentially evaluative.

In a similar vein, Livia Polanyi (1985) notes that evaluations can occur at every level of linguistic structure: phonological, lexical, syntactic, and "discourse" (e.g. the use of flashbacks). She further distinguishes between "deictic" and "contential" evaluation, a distinction related to Labov's continuum of explicitness. In deictic evaluation, the evaluating device and the evaluated information are in separate clauses; for instance, when the narrator uses repetition, or comments from outside the storyworid. In the more implicit, contential evaluation, the evaluating device and the evaluated information are in the same clause; that is, the content itself "carries" the evaluation.

Polanyi's treatment of evaluation is important because it stresses that there are no absolute evaluative devices. Any linguistic or paralinguistic structure is potentially evaluative. A given device functions as an evaluation when it is distinctive; that is, when it departs from the local norm of the text.

Evaluation functions by departing from the locally established norm of the text -- it is a use to which any linguistic structure can be put and is not in itself a property of any structure. . . . To be "contentially" or "deictically" evaluative is not a property of specific devices but depends upon the use to which the device is put in the discourse. (p. 15n)

Although this treatment of evaluation by Labov and by Polanyi is useful, it does not directly address the question of how a listener (or reader) discerns the distinctiveness ("evaluativeness") of text elements. We believe that to understand the whole reading transaction it is important to view the issue of distinctiveness from the perspective of listeners and readers as well as from that of texts, speakers, and authors.

Comprehension in general, it is well known, relies heavily on predictive and anticipatory activities. Specifically, language understanders seem to prepare "mental spaces" into which incoming material will fit -- or not fit, in which case there is an experience of surprise, of expectations unfulfilled. Thus, for language understanders, what is distinctive -- what can have evaluative force -- is that which is not normally expectable; the unpredictable, surprising, inconsistent, or incongruous. ²

Subjectively, evaluative force seems to be a consequence of things not quite "fitting." Like two rock formations that have shifted slightly and are no longer perfectly lined up, the evaluative force of a phrase or an episode seems to depend on its being slightly out of alignment with what Polanyi calls the locally established norm of the text. Evaluations, then, may be thought of as the "fault lines" of a text.³ An important difference between geological and textual fault lines, however, is that only the latter can productively be assumed to be intentional. A reader can therefore take evaluations as being deliberate invitations to share a meaning with the storyteller.

There are many ways of categorizing evaluations. Labov classifies them according to their degree of explicitness; Polanyi identifies evaluations at various linguistic levels, and further divides them into deictic and contential classes. These divisions are helpful, but again they do not frame the issue from the perspective of the comprehender.

Let us, then, consider evaluations from the point of view of the reader. In reading, a reader reads a text in which some abstract "story" is put into language -- that is, is "discoursed" (Chatman, 1978) -- by some narrator, or "teller." Looked at in this way, it is apparent that the reader is able to notice and respond to evaluations because they are signalled in three rather different ways: they may be a function of the discourse that is used in the telling; they may be a function of the more abstract story that is told; or they may be a function of the telling itself. Consequently, we posit three broad categories of textual evaluations: discourse evaluations, story evaluations, and telling evaluations. We turn now to a discussion of each of these types.

3.1. Discourse evaluations

For many readers, what is most characteristic about "literary" texts is that they include discourse elements -- stylistic patterns, rhetorical figures, diction -- that are incongruous with respect to the local norm of the text. We refer to such elements as "discourse evaluations," because they operate, in Seymour Chatman's (1978) terms, at the discourse as opposed to the story level of narrative. Although "discourse evaluation" may be a new term, there is nothing new about the kinds of distinctive stylistic and rhetorical figures that often function in this way: metaphor, simile, metonymy, parallelism, hyperbole, and the like. None of these devices are evaluative per se, however: their evaluative force derives from the way they are used in a given context; that is, from their relationship to the local norm of the text.

3.2. Story evaluations

The term "story evaluation" is also derived from Chatman's story/discourse distinction. When an incongruous element in a narrative is part of the storyworld -- that is, an event, setting, or character -- we call it a "story evaluation." It is not that a given story element is expressed in a distinctive way; rather, it is the very occurrence of the event in the storyworld that is distinctive. What counts as distinctive, of course, is determined relative not only to "the norm of the text" but also to social and cultural frames, as they are invoked by readers and narrators.

3.3. Telling evaluations

Suppose a particular event is mentioned in a narrative. The event is neither expressed in a distinctive way (and therefore is not a discourse evaluation), nor is it an unpredictable story event (and therefore is not a story evaluation). Its existence in the narrative seems incongruous, but what is incongruous is that it should have been mentioned or "told" at all, or told at that particular moment. This is a "telling evaluation." We use the term also to refer to instances in which the narrator makes meta-narrative comments concerning the storyworld or the act of narrating. Such comments are made from outside the storyworld, and therefore, by definition, are incongruous with respect to the "normal" narrating of events and states.

4. Examples of the three types of evaluations in "The Day We Got Our Own Back"

Let us justify these somewhat abstract categories, and make them a bit more concrete, by identifying some examples of each in a specific text. The first part of Maeve Brennan's "The Day We Got Our Own Back" (1969) is reproduced in Table 1. Although this story is neither an acknowledged classic, nor the sort of "difficult" fiction often used to test the limits of readers' skills, it can be considered a prototypical member of the genre: the kind of literary text one would expect high school and college students to be able to read with understanding.⁵

There are numerous passages in the text of "The Day We Got Our Own Back" that seem to us to include "evaluation." Table 2 presents a list of these from the first part of the story. The first two columns are self-explanatory. In the third column each evaluation is classified as to type (discourse, story, or telling). The fourth column shows explanations of each; these are offered not as "interpretations" but merely as traces of what we, in various re-readings, have made of the text at that stage. The table, in other words, is meant to indicate not what the text "is" but rather -- in James J. Gibson's (1979) useful term -- what it "affords." We do not, of course, claim that any one reader would notice or respond to all these evaluations, or that any two readers, or readings, would notice exactly the same ones.⁶

4.1. Discourse evaluations

Discourse evaluations, we have said, are distinctive or marked ways of expression. As an example, consider: "they crowded into our narrow little hall, and tramped around" (lines 15-16). Although this could be paraphrased, "they came into our narrow little hall, and walked around," without changing the story's event structure, the original affords something the paraphrase does not; namely, an opportunity to share a view of the men as intruders, "crowding" the family out of its home, and moreover not just walking but "tramping" aggressively around the house. Thus the narrator invites us to see the men as she does, as insensitive bullies. As we have suggested, to read in a literary way is (among other things) to accept such invitations.

4.2. Story evaluations

Story evaluations refer to distinctive story events, characters, descriptions, or settings. Like discourse evaluations, story evaluations are noticed because they are incongruous or unpredictable, either against the local norm of text, or against the norm of more general knowledge frames that readers use in comprehension. As an example of the latter, consider the initial description of the men, "carrying revolvers" (line 2). What makes revolvers evaluative, we suggest, is the social (frame) knowledge that when men in civilian clothes come to one's house, they normally do not carry revolvers. Again we are invited to share the narrator's perception of the men as truly dangerous, and hence to see the situation as threatening; indeed, as a life-and-death one.

The difference between story and discourse evaluations can be seen most clearly when they are paraphrased. Discourse evaluations seem to lose evaluative force when paraphrased (e.g. tramped vs. walked), whereas story

Table 2. Some evaluations in the opening Fragment of "The Day We Got Our Own Back"

Evaluation	Lines	Type	What is afforded
unfriendly	1	telling	the narrator-as-child expects friendliness; hence, narrator as young, innocent
dressed	1	discourse	men as play-acting; wolves in sheep's clothing
civilian clothes	1	telling	narrator as inviting the reader to share the expectation that men who visit one's house are normally dressed differently -- e.g., as soldiers
carrying revolvers	2	story	men as truly dangerous
searching for my father	2	story	political situation as close to home; personally relevant for narrator
held out	7	discourse	father as courageous, an underdog
My father . . . to see us	7-10	story	father as noble underdog
one the run	9	discourse	father as hunted person; like a fox in a fox-hunt
stealing home	10	discourse	father and family as victims
I suppose . . . to go home [Paraphrase: even though my father was not at home, we kept in touch	11-14	story	family as close; father as noble underdog; child's POV
I suppose . . . Anyway,	11-14	telling	narrator as looking back on her experience; implication of narrative voice
They crowded into our narrow little hall, and tramped around the house.	15-16	discourse	the men as powerful, arrogant; the family as vulnerable, weak
Emer, my elder sister, and my mother's chief prop	18-19	discourse	? (see text)
I was five	20-21	telling	narrator was very young; tension between adult view and child view
they crowded into the room where I sat	23-24	discourse	men as arrogant
They camped around the room	25	discourse	men as relaxed, thereby indicating just how powerful they are relative to the family
talking idly among them- selves	26	discourse	men as powerful
she was . . . him trapped	27-29	story	mother as afraid; situation as unpleasant, dangerous; mother as seeing the situation from children's POV; a proper mother
that he would . . . him trapped	29	discourse	father as hunted, a victim; men as hunters
One of the men . . . he whispered [Paraphrase: one of the men tried to trick me into revealing my father's whereabouts]	30-35	story	men as devious, immoral
This exchange . . . very clever	33-34	story	narrator (as child) as innocent, naive
my mother flew across the room at him	36-37	story	mother as protector; maternal instinct; head of the family
flew across the room at him	36-37	discourse	mother as bird; valiant underdog, quick to defend her young

evaluations, apparently, can be paraphrased with no loss of force. Thus, in the example above, what is evaluative about carrying revolvers is not the way it is expressed but rather the idea that the men were armed, an idea that could of course be expressed in many different ways. A large segment of text (e.g., an episode) may thus serve as a single story evaluation. Several instances of this are noted in Table 2. For example, lines 30-35 can be paraphrased as one evaluative event: "one of the men tried to trick me into revealing my father's whereabouts."

4.3. Telling evaluations

As with the other two types, readers can notice and respond to telling evaluations because they are unexpected, but in this case the marking occurs neither because the event is unexpected nor because its manner of expression is distinctive. Instead, the incongruity is due to the narrator's mentioning something at all, or mentioning it at a particular moment. For instance, the narrator's choosing to inform the reader that the men were wearing "civilian clothes" -- hardly a remarkable fact in itself -- can be considered a telling evaluation: by mentioning this seemingly trivial detail, the narrator invites the reader to view it as noteworthy that the men are indeed dressed as civilians.

An example of an evaluation whose force depends on its timing occurs at the end of the first paragraph, where the narrator announces that, when the events in question occurred, she was five years old.⁷ It is not remarkable that we should be told such a fact, and had it been presented at the outset ('One afternoon when I was five some unfriendly men . . .') it would hardly have been a surprise. Coming where it does, though, it stands out: we don't expect this sort of background information to be offered at this point. Specifically, "I was five" seems to constitute an invitation to share with the (adult) narrator the implications of such a young child being in such a situation.⁸

Another form of telling evaluation is meta-narrative commentary. As an example, consider: "I suppose my mother . . . I only remember . . . I know I found it very odd . . ." (lines 11-13). In our scheme, these phrases are instances of telling evaluations because they do not refer to the storyworld, but rather to the teller's narrating activities. They have the effect of abruptly pulling the reader outside the storyworld, emphasizing that the text is not "really" a story but rather a "telling" of one. In these cases, they represent, in part at least, the narrator's attempt to share a distinction between what she can remember and what she must reconstruct. The reader is invited to hear the narrator as insisting on a distinction between herself as adult narrator, looking back, and as child, involved in the incidents.

The evaluations discussed above are unusual in being fairly straightforward. Often, of course, they are more complicated. For instance, one evaluation may be embedded within another, as in: "My mother flew across the room at him" (lines 36-37). This can be considered a story evaluation, because any paraphrase -- for instance. "my mother moved quickly across the room towards him" -- also affords a view of the mother as a properly maternal, protective mother. But, in addition, the metaphoric "flew" invites the reader to share a view of the mother as a bird protecting her nest, valiant in the face of a dangerous enemy. Thus a single phrase can carry evaluation in more than one way.

Similarly, a longer stretch of text that can be summarized as a single story evaluation may contain within it segments that represent discourse, telling, or other story evaluations. Examples of each of these are given in Table 2. For instance, the telling evaluations noted earlier (I suppose; I only remember, etc.) are embedded within a story evaluation that can be paraphrased, "even though my father was not at home, we kept in touch" (lines I 1-14).

Sometimes evaluations are embedded so tightly that it is difficult to imagine separating them. Consider, for instance, lines 9-10: "He was on the run, sleeping one night in one house and the next night in another, and sometimes stealing home to see us." Here the manner of expression is distinctive (on the run, stealing home), and yet it also seems that the ideas, the storyworld events, have evaluative force in their own right (that is, the very idea that the father is a fugitive, that he must "sneak" home). In such cases it is difficult to state a clear dividing line between what is expressed and how the narrator chooses to express it.

Occasionally there is a fairly obvious instance of evaluation, and yet what it affords is not clear at all. A case in point is "my mother's chief prop" (line 19). Chief prop is an unexpected expression -- a discourse evaluation -- but it can afford quite different things for different readers. A reader for whom "chief prop" sounds like a cliché may infer that the narrator is inviting us to share a slightly hostile view of the sister; another may think the narrator is trying to appear grown-up by using this term; another may feel that she is somewhat childishly repeating a phrase she has overheard others use; and so on.

5. Ways of reading "The Day We Got Our Own Back"

Evaluations, in short, are various and complicated; they are a narrator's invitations to a reader, not the imperatives of a sovereign text.⁹ This is one reason why it is impossible to predict the details of any particular reading. On the other hand, the preceding analysis makes it possible to characterize some patterns into which particular readings may fall, according to whether and how readers transact with textual evaluations.

First of all, readers might not respond to textual evaluations at all, because of private agendas governing their reading. One might read "The Day We Got Our Own Back" in order to find out about Irish history or housekeeping habits; to flesh out a biographical picture of Maeve Brennan; to gather evidence in support of a particular interpretation of the story, or to locate examples of evaluative language in literary texts. We term such readings "information-driven." This is a complex and valuable type of reading, but readers who are engaged in it do not accept invitations to share meanings, because, in Louise Rosenblatt's (1978) terms, their concern is not with the "lived-through experience" of the work, but rather with what can be "carried away" from it (p. 24).

Second, readers who are concerned with "lived-through experience" may notice evaluations, but treat them as "facts" true with respect to that storyworld, rather than as indications of, and invitations to share, the narrator's attitudes. They are reading primarily in order to find out what happens, to immerse themselves imaginatively in the storyworld. We term readings dominated by such motives and strategies "story-driven."

In a third type of reading, however, which may be termed "point-driven," readers assume that the text is the product of an intentional act.¹⁰ More specifically, as such a reader encounters the relatively large number of incongruous words, phrases, and events in the early part of the story, she hypothesizes that these divergences from expectability are deliberate. The fact that the divergences tend to be consistent with one another (see below) is powerful evidence that they are in fact the narrator's invitations to notice and share evaluations (rather than being indicators, say, of "bad writing"). In turn, because the hypothesis of intentionality has been supported in the early part of the text, this reader becomes more likely to notice and respond to subsequent evaluations.

6. Redundancy

We have said that no one reader could (or would want to) notice and respond to all the evaluations in a text. The fact that different readers will inevitably notice different evaluations, and that the same reader will respond to different ones on different occasions, in part accounts for the obvious and often-remarked variability of literary reading. On the other hand. what also needs to be accounted for is the much less frequently remarked phenomenon of consistency across instances of reading -- both those by different readers as well as multiple readings by the same reader.

We believe that these consistencies occur at least in part because textual evaluations are usually highly redundant. For example, we have suggested that the first invitation in "The Day We Got Our Own Back" to perceive the men as dangerous is given by the phrase "carrying revolvers" (line 2). It is important to note that, whether or not this particular invitation is accepted, the perception of the men as dangerous, powerful, and arrogant is afforded many times, by means of both discourse and story evaluations (this is the case even within the relatively short fragment reproduced here; see Table 2). Similarly, invitations to share a view of the family as powerless and vulnerable, yet noble, are offered frequently throughout the text.

Conversely, it may be a lack of redundancy that accounts for disagreements and uncertainties, such as the one noted earlier in connection with the phrase chief prop. Part of the reason why it is difficult to decide what the narrator might be getting at with chief prop is that this is virtually the only evaluative reference to the sister in the story. It is much easier to share the narrator's attitudes towards the men, the mother, and the father, because there are many evaluative references to each of these throughout the text. In conversational narratives, similarly, storytellers tend to encode their attitudes and beliefs redundantly, at all linguistic, paralinguistic, and behavioral levels. Communication is fallible, in art as in life, but redundant encoding makes it more likely that there will be basic agreement on the core perceptions. There is an obvious analogy here with information theory and its attention to the ways in which redundancy and overcoding can work to ensure that messages get through in spite of the inevitable "noise" in communication systems of various kinds, from military code to the genetic messages encoded in DNA (cf. Campbell, 1982).

In brief, even though each reader will notice and respond to a unique set of text features, those who are engaged with the text in the ways we have called "literary" reading will have some major evaluative perceptions of the story in common with one another. What is crucial about this argument is not, of course, our contention that readers tend to agree, but its power to explain in some detail why they agree, without assuming that literary reading is determined by the absolute dictates of the text.

7. Conclusion

We usually know how to hear stories told in everyday, face-to-face interaction. We would not, for instance, respond to a five-year old's vivid account of getting stung by a bee by saying "So what?" The reason seems to be that such narratives occur in social contexts where narrators have obvious motives. Such narratives are heavily evaluated, and thus provide clues as to the narrator's beliefs, attitudes, and values: evaluations are, in fact, invitations to construct and share a version of what a teller might be "getting at." Now, it is perfectly clear that literary texts -- "Indian Camp,"among many others -- are usually less heavily and obviously evaluated than conversational stories, and clear also that most readers do not perceive them as occurring in conversational contexts. Even so, as we have tried to demonstrate in this paper, the processes of conversational story understanding and literary reading may be more radically similar than has previously been suspected -- linked as they are through the common concept of evaluations.

Notes

* This work was supported by grant 410-85-0612 from the Social Sciences and Humanities Research Council of Canada. We are grateful to our colleagues James A. Reither, Thom Parkhill, and especially Alan W. Mason for thoughtful critiques of previous drafts. Please send requests for reprints to Dr. Russell A. Hunt, Department of English, St. Thomas University, Fredericton, New Brunswick, Canada, E3B 5G3.

1. We are not the first, of course, to suggest an analogy between conversational and literary storytelling. The most notable precedent is the work of Pratt (1977). Pratt, however, focuses on narrators and authors, whereas our main concern is with readers. Also, the scope of her treatment is broader than ours: she demonstrates a host of structural similarities between the telling and the writing of narratives, whereas our concern here is restricted to the question of evaluations.

2. In the remainder of this paper we use these terms more or less interchangeably. It is worth emphasizing, however, that the feeling of "surprise" is not always (although it is often) associated with evaluation.

3. Analogously, historians, anthropologists, and others sometimes examine what might be termed "cultural fault lines" -- inconsistencies, say, between what is said and what is done -- because these are places that yield a particularly rich understanding of a given culture. Jonathan Smith (1982), writing of Northern hemisphere bear-hunting peoples, says: "There appears to be a gap, an incongruity between the hunters" ideological statements of how they ought to hunt and their actual behavior while hunting. For me, it is far more important and interesting that they say this is the way they bunt than they actually do so . . . . It is here, as they face the gap, that any society's genius and creativity, as well as its ordinary and understandable humanity, is to be located. It is its skill at rationalization, accommodation, and adjustment' (p. 62). We thank Thom Parkhill for bringing Smith's work to our attention.

4. "Discourse evaluations" are akin to Riffaterre's (1959) "stylistic devices." Although we would not go as far as Riffatterre when he suggests that the text acts as a computer program that prescribes its own decoding (1979/1983: 6), we do agree that stylistic devices attract attention because they are unpredictable with respect to their immediate context. However, it is worth stressing that we regard such devices, as narrator-based invitations, whereas Riffaterre appears to see them as objective text properties.

5. Some background information: "The Day We Got Our Own Back" was first published in The New Yorker 29: 120-123 (24 October 1953). Maeve Brennan moved from Dublin to the U.S. in 1934 at the age of 17. During the 1950s and 1960s she was a staff writer (the "Long-Winded Lady") on The New Yorker. She has published one other collection of short stories.

It might be wondered why we use this text instead of some better-known member of the literary canon. The immediate reason for using the Brennan story is that we have also used it in a number of empirical studies (Hunt and Vipond, 1985). In any case, our purpose here is not to elucidate any given work, but rather to illustrate our typology of evaluations -- a purpose for which the Brennan text is well suited, since it has a range of evaluation types and a fairly straightforward narrative structure. We have not found any literary texts that are unevaluated, although of course the proportions of the different evaluation types vary from text to text. For instance, in "Indian Camp" there are a large number of story and telling evaluations, but relatively few "literary language' type discourse evaluations (i.e. metaphors, similes, etc.): most of the discourse evaluations are understatements. Thus it would be quite possible to read Hemingway and be unaware of his pointedness, and, indeed, many of our students read him in this way.

6. Reading, Scholes (1985) reminds us, is always conditioned by history and ideology. How we read this story depends in unknown ways on the fact that we are North American white males, professors of English and psychology, writing in the mid-1980s. Other persons, in other historical contexts, with other ideological presuppositions, may take the story differently -- how differently, is an empirical question. This is just another way of saying that our reading is not, and in principle cannot be, the reading.

7. To say that some telling evaluations are due to their "timing" is to adopt a temporal or linear perspective on the reading process. As argued by Phelps (1985), this temporal or "flow" perspective, in which reading is viewed as on-line activity (i.e. as energy), can be set against a spatial or "design" perspective, in which reading is contemplated after the event (i.e. as matter). From the standpoint of design, then, these "timing" evaluations could perhaps be more accurately categorized as organizational evaluations -- that is, their effectiveness is due to the way in which they are juxtaposed against the rest of the picture. We owe the concept of "organizational evaluation" to Alan W. Mason.

8. Part of the distinctiveness of this item may also be due to the simple way it is expressed; thus it may be considered a discourse as well as a telling evaluation.

9. Critical theory in the twentieth century has been characterized by a dialectic between the position that texts determine, or should determine, readings and the contention that readings are essentially under the control of readers. Recently there has been more support for views similar to that first outlined by Louise Rosenblatt (1938/1976, 1978), that reading represents a transaction between text and reader to which both contribute. For a useful summary of this process, see Weaver (1985).

10. For more extended treatments of these three modes of reading, see Hunt and Vipond, 1985; Vipond and Hunt, 1984. It is worth pointing out that these modes exist in pure form only in theory; in practice they are fuzzy and overlapping.

References

Brennan, M. (1969). In and Out of Never-Never Land: Twenty-Two Stories. New York: Scribner.

Campbell, J. (1982). Grammatical Mart: Information, Entropy. Language, and Life. New York: Simon and Schuster.

Chatman, S. (1978). Story and Discourse: Narrative Structure in Fiction and Film. Ithaca: Cornell University Press.

Gibson, J. J. (1979). The Ecological Approach to Visual Perception. Boston: Houghton Mifflin.

Goffman, E. (1974). Frame Analysis: An Essay on the Organization of Experience. Cambridge: Harvard University Press.

Hemingway, E. (1953). The Short Stories of Ernest Hemingway. New York: Scribner.

Hunt, R. A. and Vipond, D. (1985). Crash-testing a transactional model of literary reading. Reader: Essays in Reader-Oriented Theory, Criticism and Pedagogy 14: 23-39.

Labov, W. (1972). Language in the Inner City: Studies in the Black English Vernacular. Philadelphia: University of Pennsylvania Press.

Peterson, C., and McCabe, A. (1983). Developmental Psycholinguistics: Three Ways of Looking at a Child's Narrative. New York: Plenum.

Phelps, L. W. (1985). Dialectics of coherence: Toward an integrative theory. College English 47: 12-29.

Polanyi, L. (1985). Telling the American Story: A Structural and Cultural Analysis of Conversational Storytelling. Norwood, NJ: Ablex.

Pratt, M. L. (1977). Toward a Speech Act Theory of Literary Discourse. Bloomington: Indiana University Press.

Riffaterre, M. (1979). Criteria for style analysis. Word 15: 154-174.

----- (1983). Text Production. (Transl. by Terese Lyons). New York: Columbia University Press. (Original work published 1979)

Rosenblatt, L. M. (1976). Literature as Exploration (3rd. ed.). New York: Barnes and Noble. (Original work published 1938)

----- (1978). The Reader, the Text, the Poem: The Transactional Theory of the Literary Work. Carbondale: Southern Illinois University Press.

Schiffrin, D. (1984). How a story says what it means and does. TEXT 4 (4): 313-346.

Scholes, R. (1985). Textual Power: Literary Theory and the Teaching of English. New Haven: Yale University Press.

Smith, J. Z. (1982). Imagining Religion: From Babylon to Jonestown. Chicago: University of Chicago Press.

Vipond, D., and Hunt, R. A. (1984). Point-driven understanding: Pragmatic and cognitive dimensions of literary reading. Poetics 13: 261-277.

Weaver, C. (1985). Parallels between new paradigms in science and in reading and literary theories: An essay review. Research in the Teaching of English 19: 298-316.

Russell A. Hunt teaches literary theory and eighteenth-century English literature, and Douglas Vipond teaches psychology, at St. Thomas University, Fredericton, New Brunswick, Canada. They are currently undertaking theoretical and empirical studies of the cognitive and social process of literary reading.