Empirical Studies of Literature --

Dixon & Bortolussi, 2001

Dixon, P. & Bortolussi, M. "Prolegomena for a Science of Psychonarratology". New Perspectives on Narrative Perspective. Willie van Peer and Seymour Chatman (Eds.). Albany, NY: State University of New York Press, 2001: 275-287.
index

Summary

In this article, Dixon and Bortolussi propose some parameters for a new field of psychonarratology, an interdisciplinary enterprise bringing together narratology and discourse processing in a mutually beneficial exchange. According to the authors, in the field of narratology, there is a shared assumption that the process of reading must be considered in understanding the way in which narratives function and that the narrator should be thought of as a mental representation in the mind of the reader, rather than a collection of textual signs. As a result, it is assumed that reader processing is an empirical question. The problem, according to them, is that the field of narratology is ill-equipped to conduct investigations on how readers process texts. On the one hand, such a field could, then, benefit from discourse processing and cognitive psychology, which offer many of the methodological and interpretive tools for investigating readers' processing. On the other hand, the field of discourse processing has generated various evidences and theorizing on how readers process texts, but the research, as stated by the authors, has been minimally concerned with literary processing, generally, or with narrative perspective, specifically. The authors suggest that the problem of processing the narrator and narratorial perspective is often not recognized as a separate problem. As a consequence, they argue that discourse processing can benefit from narratology as it does not have the conceptual distinctions necessary for understanding narrative texts.

Therefore, the authors suggest that psychonarratology should deal with the problem of narratorial processing. Distinguishing between text features and text effects is the crucial element here. The term text features (including narratorial focalization) is used to refer to anything that can be objectively and reliably identified in the text. Text features include distinctions from classical studies of narratology, such as mode, focalization, hetero- versus homodeigetic narration, and absent versus present narrator. In contrast, text effects (including narratorial stance) refer to events in the mind of reader. These effects include mental representations of various sorts, changes in attitude or belief, and affective reactions. This means that stance is a text effect and should be construed as existing in the mind of the reader, not the text. This need to distinguish between text features and effects is essential for allowing one to bring the right methods to bear on different aspects of the problem of understanding narrative processing. Psychonarratology is, then, described as an interdisciplinary investigation of narratorial processing, consisting of the empirical investigation of how the distinctions and variables identified by narratologists are processed and how they affect readers.

The authors' arguments are based on speech act theory. They assume that readers construct a mental representation of the narrator that is parallel to that constructed of the other participant in conversation, that is, one in the form of an anthropomorphic entity. The attribution of traits to the narrator is such that readers assume that the narrator follows the conversational conventions described by Grice (1975). In this case, the representation is that of the narrator and not of an actual individual, and it is inferred by the reader from text features. They propose that, just as in conversation, the representation of the narrator contains information concerning the narrator's knowledge, goals and perspective. Instead of describing narrative as "communication in a special sense" (cf. Lanser, 1981), they adopt the view that communication and the related assumptions of speech act theory only exist in the mind of the reader. In other words, "text is not communication, but it is often treated as if it were by readers" (p. 280). Furthermore, they suspect this is a natural process, and that it is probable that language is handled the same way.

Four implications, or maybe assumptions, of the proposal of a conversational narrator are discussed. Such assumptions are all justified and based on empirical evidences that the authors do not deny can be regarded as controversial. First, such a notion leads the reader to attribute to the narrator those properties that are necessary for conversational communication. In particular, the narrator would be assumed to share perceptual ground, language, and culture with the reader. When it does not occur, readers will have trouble understanding and remembering material. Second, if the narrator is assumed to be a conversational participant communicating about the described world, then the narrator must be part of that world. This means that the narrator must have a unitary location and focalization in that world, even when the narrator is technically absent from the narration. The implication is that when focalization changes, comprehension will be disrupted because some extra inference or processing is required. They assume that unitary focalization and position are the logical consequences of representing the narrator as a conversational participant and as part of the described world. Third, they predict that readers attempt to cooperate with the narrator and use their own knowledge and experience to fill in gaps or background as required, in an attempt to justify the narrator's stance. Fourth, just as in conversation, readers assume that the narrator is cooperating with them. They expect the narrator to provide necessary and sufficient information. If the text in fact provides extra, unnecessary information, readers will spend additional time and effort attempting to ascertain how the additional information is relevant or important. Alternatively, if the text provides insufficient information for comprehension, readers will assume that there is a plausible inference that could be made in order to render the narrative comprehensible, and that they are likely to spend time searching for such an inference.

In conclusion, the authors intend to introduce psychonarratology as an evolving theoretical framework and approach to empirical investigation of narratorial processing. According to them, its value stems from the division of the work of understanding narratorial perspective into a number of "more manageable subproblems." These subproblems are, in their words, amenable to systematic and straightforward empirical investigation.

Critique

Dixon and Bortolussi base their psychonarratology on narratology and discourse processing. These two fields are, however, not without problems. Questioning around such a combination formed, then, some of the main targets for class discussion. Issues debated concerned methodological procedures, the validity of the assumptions they make, including a critique of their proposal in general.

The authors claim that a crucial step in their proposal is to distinguish text features from text effects. According to them, psychonarratology should consist of the empirical investigation how the text features identified by narratologists are processed. They describe text features as "anything that can be objectively identified in a text … [and] that trained observers can reliably agree whether or not such features are present in a given text" (p.277). Although they provide well supported arguments for such a claim in one previous article (Dixon et al., 1993), the question whether it can be generalized to all text features continues without answer.

Issues concerning the field of discourse processing were also raised. The strongest critique was that it is based uniquely on cognitive models and appears to provide a poor account of literary reading. It does not account for affective reactions, for instance. The appeal to Grice's Cooperative Principles (Grice, 1975) makes the issue even more questionable as such principles are based on what does not work, that is, on the notion of implicatures. His maxims are idealizations and not necessarily what happens in the "real" world. Besides that, such appeal presupposes that literary experience is all about communication. The authors make such a claim explicit while stating that "text is not communication, but it is often treated as if it were by readers" (p. 280). Communication in the ways suggested by Grice seems to be a poor analogy and to confine it exclusively to cognitive responses seems an impoverished account of what happens in the process of response to literary texts. The experiential aspect of the process is disregarded.

The proposal that the narrator is regarded as a conversational participant was evaluated as interesting, but the support the authors provide for such an argument appears to be unsuccessful in some cases. Discussions emerged around the question why the authors regard the empirical studies they base their assumptions on to be evidence for considering the narrator a conversational participant. One example is the study by Kintsch and Greene (1978) presented as a demonstration supporting such an assumption. This study seems to provide no evidence for such a claim. Also, to what extent is it possible to claim that longer reading times may provide evidence that extra processing is required due to the fact that the narrator is understood as a conversational participant and not due to the use of more complex forms of discourse, such as the free indirect speech, for instance? The possible need for other better suited procedures to further examine this assumption deserves attention. Another problem detected is that they seem to conflate narrator and author in some parts of their commentaries, as may be noticed on p. 282 of this article. In addition, it is asserted by the authors that text features have an effect on readers' mental constructions. The question such a proposition raises is whether it is a one-sided effect. How far readers' mental constructions have an effect on the reading of text features seems to be another relevant empirical question.

In the same line, discussions regarding methodological procedures questioned the implications of working with text manipulation, a procedure the authors frequently make use of. Once you manipulate a text, other variables may be introduced rather than the ones originally considered. Such a procedure is imported from the biological sciences, but appears to be troublesome when applied in the literary domain as it influences the "ecology" of the text. The debate was extended when the idea of working with text extracts was also considered a kind of text manipulation, although it is less frequently criticized.

A final critique related to the terminology used to name their proposed enterprise: "psychonarratology". According to the etymology of the word, "psychological" relates to all of the aspects counteracting the physical. It contemplates both cognition and affect. Although the authors state they are interested in accessing text effects such as changes in belief and affective reactions, they still seem to be centered around the investigation of the formal features of narrative, with a focus on cognition. The term "cognitive" relates solely to the intellect, to the notions of knowledge, perception, mental construction. The way this theory can account for ideological stances and feelings needs further clarification. Also, the question whether their proposal seems to be still more a theory of "narrative reading" rather than a "psychonarratology", which would then involve the production of narratives as well, was raised. Such a question seems to be untouched so far. Whether this is a matter of a "psychonarratology" or a "cognitive narratology," the fact is that cognitive approaches still seem to provide a limited account for literary reading.

References

Dixon, P., Bortolussi, M., Twilley, L. C., & Leung, A. "Literary Processing and Interpretation: Towards Empirical Foundations." Poetics 22 (1993): 5-33.

Grice, H. P. "Logic and Conversation." Syntax and Semantics 3 (1975): 41-58.

Kintsch, W., & Greene, E. "The Role of Culture-Specific Schemata in the Comprehension and Recall of Stories." Discourse Processes 1 (1978):1-13.

Document created November 20th 2005

Dixon & Bortolussi, 2001
Dixon, P. & Bortolussi, M. "Prolegomena for a Science of Psychonarratology". New Perspectives on Narrative Perspective. Willie van Peer and Seymour Chatman (Eds.). Albany, NY: State University of New York Press, 2001: 275-287.	index
Summary
In this article, Dixon and Bortolussi propose some parameters for a new field of psychonarratology, an interdisciplinary enterprise bringing together narratology and discourse processing in a mutually beneficial exchange. According to the authors, in the field of narratology, there is a shared assumption that the process of reading must be considered in understanding the way in which narratives function and that the narrator should be thought of as a mental representation in the mind of the reader, rather than a collection of textual signs. As a result, it is assumed that reader processing is an empirical question. The problem, according to them, is that the field of narratology is ill-equipped to conduct investigations on how readers process texts. On the one hand, such a field could, then, benefit from discourse processing and cognitive psychology, which offer many of the methodological and interpretive tools for investigating readers' processing. On the other hand, the field of discourse processing has generated various evidences and theorizing on how readers process texts, but the research, as stated by the authors, has been minimally concerned with literary processing, generally, or with narrative perspective, specifically. The authors suggest that the problem of processing the narrator and narratorial perspective is often not recognized as a separate problem. As a consequence, they argue that discourse processing can benefit from narratology as it does not have the conceptual distinctions necessary for understanding narrative texts.
Therefore, the authors suggest that psychonarratology should deal with the problem of narratorial processing. Distinguishing between text features and text effects is the crucial element here. The term text features (including narratorial focalization) is used to refer to anything that can be objectively and reliably identified in the text. Text features include distinctions from classical studies of narratology, such as mode, focalization, hetero- versus homodeigetic narration, and absent versus present narrator. In contrast, text effects (including narratorial stance) refer to events in the mind of reader. These effects include mental representations of various sorts, changes in attitude or belief, and affective reactions. This means that stance is a text effect and should be construed as existing in the mind of the reader, not the text. This need to distinguish between text features and effects is essential for allowing one to bring the right methods to bear on different aspects of the problem of understanding narrative processing. Psychonarratology is, then, described as an interdisciplinary investigation of narratorial processing, consisting of the empirical investigation of how the distinctions and variables identified by narratologists are processed and how they affect readers.
The authors' arguments are based on speech act theory. They assume that readers construct a mental representation of the narrator that is parallel to that constructed of the other participant in conversation, that is, one in the form of an anthropomorphic entity. The attribution of traits to the narrator is such that readers assume that the narrator follows the conversational conventions described by Grice (1975). In this case, the representation is that of the narrator and not of an actual individual, and it is inferred by the reader from text features. They propose that, just as in conversation, the representation of the narrator contains information concerning the narrator's knowledge, goals and perspective. Instead of describing narrative as "communication in a special sense" (cf. Lanser, 1981), they adopt the view that communication and the related assumptions of speech act theory only exist in the mind of the reader. In other words, "text is not communication, but it is often treated as if it were by readers" (p. 280). Furthermore, they suspect this is a natural process, and that it is probable that language is handled the same way.
Four implications, or maybe assumptions, of the proposal of a conversational narrator are discussed. Such assumptions are all justified and based on empirical evidences that the authors do not deny can be regarded as controversial. First, such a notion leads the reader to attribute to the narrator those properties that are necessary for conversational communication. In particular, the narrator would be assumed to share perceptual ground, language, and culture with the reader. When it does not occur, readers will have trouble understanding and remembering material. Second, if the narrator is assumed to be a conversational participant communicating about the described world, then the narrator must be part of that world. This means that the narrator must have a unitary location and focalization in that world, even when the narrator is technically absent from the narration. The implication is that when focalization changes, comprehension will be disrupted because some extra inference or processing is required. They assume that unitary focalization and position are the logical consequences of representing the narrator as a conversational participant and as part of the described world. Third, they predict that readers attempt to cooperate with the narrator and use their own knowledge and experience to fill in gaps or background as required, in an attempt to justify the narrator's stance. Fourth, just as in conversation, readers assume that the narrator is cooperating with them. They expect the narrator to provide necessary and sufficient information. If the text in fact provides extra, unnecessary information, readers will spend additional time and effort attempting to ascertain how the additional information is relevant or important. Alternatively, if the text provides insufficient information for comprehension, readers will assume that there is a plausible inference that could be made in order to render the narrative comprehensible, and that they are likely to spend time searching for such an inference.
In conclusion, the authors intend to introduce psychonarratology as an evolving theoretical framework and approach to empirical investigation of narratorial processing. According to them, its value stems from the division of the work of understanding narratorial perspective into a number of "more manageable subproblems." These subproblems are, in their words, amenable to systematic and straightforward empirical investigation.
Critique
Dixon and Bortolussi base their psychonarratology on narratology and discourse processing. These two fields are, however, not without problems. Questioning around such a combination formed, then, some of the main targets for class discussion. Issues debated concerned methodological procedures, the validity of the assumptions they make, including a critique of their proposal in general.
The authors claim that a crucial step in their proposal is to distinguish text features from text effects. According to them, psychonarratology should consist of the empirical investigation how the text features identified by narratologists are processed. They describe text features as "anything that can be objectively identified in a text … [and] that trained observers can reliably agree whether or not such features are present in a given text" (p.277). Although they provide well supported arguments for such a claim in one previous article (Dixon et al., 1993), the question whether it can be generalized to all text features continues without answer.
Issues concerning the field of discourse processing were also raised. The strongest critique was that it is based uniquely on cognitive models and appears to provide a poor account of literary reading. It does not account for affective reactions, for instance. The appeal to Grice's Cooperative Principles (Grice, 1975) makes the issue even more questionable as such principles are based on what does not work, that is, on the notion of implicatures. His maxims are idealizations and not necessarily what happens in the "real" world. Besides that, such appeal presupposes that literary experience is all about communication. The authors make such a claim explicit while stating that "text is not communication, but it is often treated as if it were by readers" (p. 280). Communication in the ways suggested by Grice seems to be a poor analogy and to confine it exclusively to cognitive responses seems an impoverished account of what happens in the process of response to literary texts. The experiential aspect of the process is disregarded.
The proposal that the narrator is regarded as a conversational participant was evaluated as interesting, but the support the authors provide for such an argument appears to be unsuccessful in some cases. Discussions emerged around the question why the authors regard the empirical studies they base their assumptions on to be evidence for considering the narrator a conversational participant. One example is the study by Kintsch and Greene (1978) presented as a demonstration supporting such an assumption. This study seems to provide no evidence for such a claim. Also, to what extent is it possible to claim that longer reading times may provide evidence that extra processing is required due to the fact that the narrator is understood as a conversational participant and not due to the use of more complex forms of discourse, such as the free indirect speech, for instance? The possible need for other better suited procedures to further examine this assumption deserves attention. Another problem detected is that they seem to conflate narrator and author in some parts of their commentaries, as may be noticed on p. 282 of this article. In addition, it is asserted by the authors that text features have an effect on readers' mental constructions. The question such a proposition raises is whether it is a one-sided effect. How far readers' mental constructions have an effect on the reading of text features seems to be another relevant empirical question.
In the same line, discussions regarding methodological procedures questioned the implications of working with text manipulation, a procedure the authors frequently make use of. Once you manipulate a text, other variables may be introduced rather than the ones originally considered. Such a procedure is imported from the biological sciences, but appears to be troublesome when applied in the literary domain as it influences the "ecology" of the text. The debate was extended when the idea of working with text extracts was also considered a kind of text manipulation, although it is less frequently criticized.
A final critique related to the terminology used to name their proposed enterprise: "psychonarratology". According to the etymology of the word, "psychological" relates to all of the aspects counteracting the physical. It contemplates both cognition and affect. Although the authors state they are interested in accessing text effects such as changes in belief and affective reactions, they still seem to be centered around the investigation of the formal features of narrative, with a focus on cognition. The term "cognitive" relates solely to the intellect, to the notions of knowledge, perception, mental construction. The way this theory can account for ideological stances and feelings needs further clarification. Also, the question whether their proposal seems to be still more a theory of "narrative reading" rather than a "psychonarratology", which would then involve the production of narratives as well, was raised. Such a question seems to be untouched so far. Whether this is a matter of a "psychonarratology" or a "cognitive narratology," the fact is that cognitive approaches still seem to provide a limited account for literary reading.
References
Dixon, P., Bortolussi, M., Twilley, L. C., & Leung, A. "Literary Processing and Interpretation: Towards Empirical Foundations." Poetics 22 (1993): 5-33. Grice, H. P. "Logic and Conversation." Syntax and Semantics 3 (1975): 41-58. Kintsch, W., & Greene, E. "The Role of Culture-Specific Schemata in the Comprehension and Recall of Stories." Discourse Processes 1 (1978):1-13.