Forms of reading: Recovering the self-as-reader
Isn't this the most elusive and private of all conditions, that of the self suspended in the medium of language, the particles of identity wavering in the magnetic current of another's expression? How are we to talk about it? -- Sven Birkerts, The Gutenberg Elegies (1994), p. 78.
1. The dismissal of formalism
Literary scholars continue to produce readings of texts and elaborations of literary and cultural theory at an increasing rate, but this industry is directed almost exclusively at fellow scholars and senior students. Almost no professional attention is being paid to the ordinary reader, who continues to read for the pleasure of understanding the world of the text rather than for the development of a deconstructive or historicist perspective. The concerns that an ordinary reader seems likely to have about a literary text, such as its style, its narrative structure, or the reader's relation to the author, the impact on the reader's understanding or feelings -- such concerns now seem of little interest. Even the primers being produced for the beginning literature student often focus on theory rather than literature: e.g. Durant and Fabb (1990) devote the bulk of their nine chapters almost entirely to theoretical concerns. Students learn about Leavis and Lacan before they are invited to engage in any depth with a single literary text. Literary texts, indeed, as this book seems to imply, are arraigned at the bar of the Court of Theory, where constructivists and deconstructivists, new historicists and neo-marxists, are council for prosecution as well as judge, and the literary text is either found guilty of crimes of class, patriarchy, or race, or declared indifferent to human concerns as the inhabitant of a world of pure language.
This disjunction between professional concerns and the interests of the ordinary reader seems profound. We would also suggest that it is dangerous and ultimately self-defeating. If the gap is to be narrowed, it will be by focusing once again on the formal aspects of the literary text through which, we will propose, the ordinary reader's concerns can primarily be located. However, in contrast to earlier, now discredited versions of formalism that explicitly forbade interest in readers, we argue that the formalist dimension of reading can be examined effectively only in cooperation with actual readers. By studying readers' experiences of literary reading and its outcomes, we will begin to map the structures of interaction between reader and text and discover what formal structures are created in common among readers of a given text. From this perspective too, we will develop a more ecologically valid approach to understanding the role and functions of literature in general. Why have all cultures, as far we know, developed a literary culture, whether oral or print? Why do people seek out and read novels and poems, or go to watch plays? And what place is there for literature, in the late twentieth century, after all, when other forms of media have become so prevalent?
Recent literary theory, as everyone knows, has legislated against the assumptions of formalism, but this is only one plank of a general assault on a range of assumptions that have informed critical practice since the time of Dryden or Boileau, and in some cases, from Aristotle. The ascription of agency to author and, in some quarters, to the reader, has been abandoned; along with this has gone the sense of the poet or writer as a maker, one who creates: the standard poststructuralist view is that texts are made out of other texts (as Barthes puts it, "woven entirely with citations, references, echoes, cultural languages"), their mode of being lying in the "infinite deferment of the signified" (Barthes, 1977, 160; 158). In this perspective a literary work has no inherent structure or unity: "there is no autonomy proper to the text," asserts Harari, introducing the work of the poststructuralist theorists collected in his 1979 anthology Textual Strategies (1979; p. 70). Critical theory itself now claims a place alongside the literary texts it purports to discuss: as a mode of production it is said to equal the work of literary authors in importance and creativity, although it has to be said that outside the academy almost no one is reading it, unlike the continued wide readership for literary texts from Shakespeare to Seamus Heaney (as even a check of the stock in any airport bookshop will show).
Of course, it may be the case that the vision of today's avant garde theorists will become the standard wisdom of tomorrow, just as we all (or almost all) now live within a world view shaped by Darwin and Einstein. However, modern arguments about literary meaning lack one dimension that, in the end, made the arguments of Darwin and Einstein compelling: today's theorists produce no evidence to back up their claims other than their own experience and assumptions and their appeals to other theorists. The empirical dimension, in other words, is absent. While Darwin's views were supported, for example, by his observations on a series of finches, and Einstein's universe of time and space have been confirmed by measurements of the speed of light, the major claims of poststructuralist theory have not been supported, to our knowledge, by a single empirical study. Indeed, study of actual readers is ruled out of order by theorists such as Culler (1981, p. 121): what is of interest, he argues, are the conventions that determine reading, not the experience of real readers; what these conventions might be is, of course, decided a priori by Culler and his colleagues. This refusal to check theoretical presuppositions against the reading practices of actual readers calls into question the recent, almost universal dismissal of formalism. The claims of formalism, as proposed, for example, by critics from Coleridge to Wimsatt, have not been falsified by any investigation of the case, i.e., whether formalist dynamics underlie the literary reading of ordinary readers.
In fact, Culler's dismissal of real readers as worthy of study is founded on a serious misconception, in which Culler's antipathy to both formalism and empiricism are oddly yoked together. Speaking of Fish's shift of attention to the interpretive process, Culler notes this advantage:
it frees us from the notion that the poem is some kind of autonomous object which 'contains' its meaning as an inherent property. That notion has unfortunate consequences; it suggests that the reader, like a good empiricist, should approach the poem without any preconceptions so as to read only what is there in the text. The implication that the ideal reader is a tabula rasa on which the text inscribes itself not only makes nonsense of the whole process of literary education and conceals the conventions and norms which make possible the production of meaning but also ensures the bankruptcy of literary theory, whose speculations on the properties of literary texts become ancillary . . . (p. 121)
Not only do these remarks reveal an undue anxiety that literary theory might be ancillary; more seriously, they create an image of empirical procedure that is a caricature, an activity that somehow gets underway with no prior notions of what may be significant (ignoring in the process both scientific method and the findings of psychology). Culler, in other words, has already decided what data would be elicited by an empirical study of the formalist hypothesis, and declares this invalid and a danger to both literary education and literary theory.
But among literary theorists of his school, Culler's procedure here is not untypical: the image of formalism he sets up is a straw man, which he then has no difficulty in demolishing. Meanwhile, the significant questions posed by formalism are left unanswered. These include questions about which Culler appears anxious, although he formulates them in a way that precludes empirical study: to what extent is a formalist approach to a literary text the result of literary education; how much is a formalist reading due to interpretive conventions, how much to psychological processes independent of education; what is supplied by the text while reading, what by the reader. Culler's account suggests that all readers would arrive at the same reading of a given text, but this is clearly far from likely, as the most superficial understanding of cognitive psychology would show. A theory of formalism does not require uniformity of interpretion.
Perhaps the central issue of formalism is that of "literariness": the claim that literary texts possess certain distinctive forms and features not found in other types of text. If this claim can be invalidated, then other concomitant claims asserted by formalism become untenable. Thus it is this position that we will now examine in more detail, by reviewing several theoretical and empirical studies, including our own. After we have done this we will suggest some of the wider implications of the concept of "literariness" for the cultural place of reading.
2. Literariness
Along with formalism itself, the arguments for literariness have been almost universally rejected by mainstream literary academics. They have also been rejected by some of the more significant theorists in the empirical tradition. By looking briefly at two of these, we will outline the more prominent reasons why the concept of literariness fell from favour.
In his paper of 1979, "Advice on theoretical poetics," Van Dijk put forward arguments which are now almost universally accepted. There are, he said, no properties of discourse which are not common to literary and non-literary fields. For example, the principle of "coherence" is of interest equally to students of literary and non-literary discourse. In enumerating the features of so-called "poetic language," we find few that appear to be distinctive to literature (598-9). Thus, he concludes, "there is no serious way in which the notion of 'poetic language' could be defined: no language forms are exclusively used in literature, or not used in literature, or even in poetry" (p. 601). Of the socio-cultural system that involves writers, readers, critics, professors of literature, etc., he remarks, "there are no sufficient or necessary properties of 'literature' in any of these contexts".
"Literature" in his view is a set of social practices. Van Dijk would ask for "a more or less empirically warranted account of what people (in some context) call 'literature'" (602). More generally, he would discard literature as a defining context for research: "I . . . would suggest that a theory of literary texts should be based on a more general theory of discourse, as it is developed within the new interdiscipline of discourse studies." (p. 598)
The position elaborated here, which coincides in important respects with that of Culler, is based on a theoretical view of discourse in general. However, the position is framed in a way that precludes empirical investigation. It rules out the concept of "literariness" a priori, by defining it as a set of "poetic features." Reified in this way, we are prevented from asking whether, even if such features do occur in all types of discourse, particular occurrences or conjunctions of poetic features elicit a response distinctive to literature. We should also ask whether discourse theory has the appropriate tools for the empirical study of such responses. For example, does the linguistic principle of "coherence" apply with the same meaning in a literary as a non-literary context?
Siegried Schmidt (1982) asserts a similar position. The attempt to locate attributes of literariness in surface features of texts, he claims, is an "ontological fallacy." It is "the human processes performed on such features that define the attributes in question" (p. 92), and such processes are switched on, as it were, not by the text, but by adoption of what Schmidt terms the Aesthetic Convention. As readers we are told, this is a novel, or, this book is known to be by a poet, thus we adopt an attitude that impels us to identify features of the text as literary. The process of recognizing literary features is itself a product of acculturation and education. While this argument appears plausible, it runs into two difficulties: first, it overlooks the occurrence of aesthetic response outside any institutional context (e.g., the success of Shakespeare's plays in their time with the "groundlings"), and it denies the possibility that inherent properties of literary features, even when presented within an aesthetic context, switch on the aesthetic response, rather than the reverse (it is, as all teachers of literature know, quite possible to present a text as aesthetic; it is another matter to have the text read aesthetically).
Schmidt seems to find the following argument decisive: "If there were such a thing as a "poetic function" or a "poetic competence" divorced from actual use, then a given text would either be poetic or non-poetic for all members of a language community at every moment -- a conclusion that can easily be disproven" (p. 94). On the contrary, let us suppose we present a given text, such as Shakespeare's Sonnet no. 73, to a random selection of people in the street. How many would assert, this is not a poem? Studies that have set out to examine this proposition, in fact, appear to show just what Schmidt seems to deny. We will examine two of these studies now, those by Hoffstaedter (1987) and Hanauer (1996).
Echoing Van Dijk and Schmidt, Hoffstaedter (1987) declares her "basic assumption" to be "that poeticity . . . is a property of text processing rather than a property of texts." It "takes place," she says, "under specific text and context conditions and depends also on specific dispositions of the reader" (p. 75) Her study, which sets out to show this empirically, in fact achieves the opposite, pointing to the conclusion that poetic features are inherent to literary texts.
In her study, 24 texts ranging from a Celan poem to a passage from a history text book were presented in either a newspaper or poetry reading context. They were then rated on how poetic they were thought to be: only ten of the texts showed significantly different scores, according to the two conditions. In other words, a number of texts were rated as poetic regardless of context. Hoffstaedter concludes that "the influence of the context condition depends on whether certain text properties occur which may cause poetic text processing. If, for instance, there are many properties which potentially contribute to a poetic processing, then the context information has little influence, i.e. the text is processed poetically in any context." (p. 80) This finding does not support Schmidt's Aesthetic Convention.
Hoffstaedter also assumed that literary experience is a determining factor in whether texts are read "poetically" or not. Thus her study employed two groups of readers, students of literature and of engineering. One of her measures asked readers to underline words and phrases deemed poetic in the 24 texts. For the text she reports in detail, the engineers selected half the number selected by literature students, 33% vs. 67%, a very significant difference. The data are presented in Hoffstaedter's Figure 2 which shows the percentage of underlinings across 21 words in three lines of a Celan poem. At first sight, this seems strong evidence that literary training influences what will be found poetic. But look again: a noticeable feature of the chart is the similar proportion of underlinings made by the two groups: a word found highly poetic by the literary group also occurs in a high position relative to other words in the engineers' group. In fact, the percentages of the two groups are highly correlated, r(19) = .599, p < .01. Thus, while the engineers appear less confident or committed in their judgements, they seem able to discriminate between levels of "poeticity" in this text as accurately as the more highly trained literature students.
In both these studies, then, Hoffstaedter's data point rather firmly towards accepting the formalist hypothesis that literariness is a quality of texts, perceived independently of context or literary training. Recognition of poetic features appears to be based on understanding of the language, i.e., what is appropriate to "normal" usage and what to "poetic" usage. Interestingly, Hoffstaedter also studied the effect of linguistic competence. With learners of German, performing the same underlining task, a more striking set of differences emerged: the underlinings of the less competent speakers of German were fewer and corresponded far less clearly with those of the native speakers. As the formalist hypothesis would predict, linguistic competence appears a more significant influence on judgement of what is poetic than literary experience (p. 83).
In Hanauer's (1996) recently published paper, two main views on how judgements of what is poetic are made are described: the "traditional," which gives a central place to formal features of a text, and the "conventionalist," which emphasizes the conventions that are applied during reading, such as Schmidt's Aesthetic Convention. Hanauer sets out to test the conventionalist and formalist positions against empirical evidence, and for this purpose he sees the text categorization judgement as the critical issue. Incidentally, we would point out that a moment's reflection will suggest that this is to create a somewhat improbable situation: in real life we are rarely in doubt over what kind of a text we are reading. However, the evidence reported by Hanauer is suggestive.
Two texts were used by Hanauer. Each was a poem where he manipulated its "poetic" features in two ways: the phonetic and graphic information were rewritten to produce versions that were low, middle, or high (original version) on each of these two dimensions. Readers were then asked to rate each version on a continuous scale running from "clearly a poem" at one end to "clearly not a poem" at the other. Readers were of two kinds, novice (entry level literature students) and experienced (holding a degree in literature). The study was replicated with both poems and with both types of reader, with broadly similar results.
The ratings of the novice group, especially as the lower end of the
scale, were generally lower than the experienced group: as in Hoffstaedter's
study (and our own studies), the less experienced readers seem less committed
to the act of reading. Hanauer takes this as support for the conventionalist
position. It should be noted, however, that to the extent that more of
the poetic features are present the closer are the judgements of the two
groups of readers. 
Our
own graph based on Hanauer's data for the first poem (see Figure 2, left)
shows this more clearly than Hanauer's report. This suggests that the more
poetic the text, the more independent it is of literary experience, an
argument for the traditionalist view.
More striking, however, is the fact that the distribution of judgements across the different texts was largely the same in both groups, as our graph shows, as well as Hanauer's own graphs: in other words, novice readers were as competent as experienced readers in placing texts on a poetic scale according to the degree to which they possessed graphic or phonetic features, a finding that corresponds closely to our reading of Hoffstaedter's data. Like Hoffstaedter, the judgements of the novice readers occur at a lower level than those of the experienced readers, a finding that appears to support the conventionalist view. But the close correspondence of the judgements clearly speaks for the traditionalist view, as the high correlation between them shows: r(7) = .944, p < 001. As Hanauer notes: "both novice and expert literary readers were found to be sensitive to the use of graphic and phonetic information in making poetry categorization judgements and . . . these information sources were integrated in a similar way" (p. 371).
Hanauer concludes that formal features thus play an important role in categorizing texts. However, "While sensitivity to formal textual features and the way to integrate this information may stay constant, the value assigned to these textual features was seen to change according to literary educational background" (p. 374). This last comment seems apt: literary education, among other things, enables a reader to build interpretive strategies upon the textual features seen, that is, to assign them a value within the larger unfolding sense of meaning of the text as a whole. But this is not a strong argument for the conventionalist position. If we take the traditionalist position to argue for the initiation of poetic processing through readers' recognition of poetic features, then the weight of evidence in Hanauer's study is largely in favour of it.
Our own research, particularly the set of studies we reported two years ago in Poetics (Miall and Kuiken, 1994a), can also be seen as evidence for the traditionalist view, although our main purpose in conducting the studies was not to arbitrate between the two opposing views, as in Hanauer's work. We set out, rather, to ask the question, what is the purpose of poetic features? Or, more precisely, what is distinctive about readers' responses to foregrounding? One answer is provided by Shklovsky's (1964/1917) well-known formulation: the "making strange" of literature by poetic devices lengthens perception; it makes the stone more stony, that is, when reading literature we seem to see and hear more vividly. As László (1990) showed, one effect of reading literature is the formation of more vivid imagery than when reading newspaper reports; and he notes a comment of Pavio (1985) that emotions are associated with images rather than with propositions. Foregrounding, then, elicits a more immediate, vivid, and personal response from a reader. Needless to say, a reader need not be consciously aware of responding to foregrounding in order for this effect to occur. In our work on foregrounding we have attempted to study the process of response to such literary devices. What, psychologically, appears to signify the encounter with foregrounding?
We took three literary short stories by O'Faolain, Woolf, and Mansfield, and coded each segment (roughly one sentence) for foregrounded features at the phonetic, grammatical, and semantic levels. We then elicited several measures from readers, such as reading times per segment, and ratings for affect and strikingness. We also employed two types of readers: experienced students of literature (most in their third or fourth year of studies) and introductory psychology students, who generally had little experience of literature. With both groups we obtained significant correlations between foregrounding and the reading times and ratings data. Both groups read the Woolf and Mansfield stories, and it is notable that the level of correlations with foregrounding is almost the same with both groups. This finding lends support to the hypothesis that all readers appear sensitive to foregrounding, regardless of literary training.
It was also noticeable that the overall means of the ratings provided by the two groups differed consistently: our experienced readers gave higher affect and strikingness ratings than the psychology students (e.g., mean affect ratings for the Woolf story were 2.98 and 2.62 respectively; strikingness ratings were 3.04 and 2.79; these differences were significant on a t-test, p < .05; similar results were found with the Mansfield story.) This suggests that the inexperienced readers were less committed to the reading or less interested in it. Yet, both groups appear to have been almost equally responsive to the presence of foregrounding.
In addition, the foregrounding measure correlated significantly with ratings for uncertainty, which we also collected for several stories. This has led us to propose that foregrounding initiates interpretive activity in the reader, first by defamiliarizing (the strikingness rating provides one measure of this) and by arousing feeling; then, the resulting uncertainty causes the reader to search for a context in which the new material can be understood, a process in which feeling probably plays a key role. Feeling may be a route to relevant memories, experiences, or concepts that have not up to this point been applied to understanding the text. A study of Andringa (1990) seems to support this proposal: examining verbal protocols gathered in response to a literary text, she found that expressions of emotion in response to a literary text tended to be followed by evaluations and arguments, and this sequence of events provided the main impetus to the development of an interpretation.
If the reader, experiencing uncertainty and feeling, is searching for an interpretive context, we might also expect to find readers considering a wider range of ideas than at other moments in the text when interpretation is more straightforward. This is the implication of another study we carried out. When readers are asked to talk aloud about a story, and we then classify the various statements that readers make in response to each segment, we find that a wider range of different statements corresponds to the more highly foregrounded segments (i.e., number of types of level 2 constituents in our protocol analysis correlates with foregrounding: partial correlation, r(82) = .25, p < .025, controlling for syllables). Foregrounded passages, then, are strong candidates for locating moments of indeterminacy in interpretation, and these in turn provide the germinating points for the subsequent development of interpretations of the text. Thus foregrounding appears to play an important role in the interpretive process (cf. Miall and Kuiken,1995).
3. Conclusion
In conclusion, we believe that the empirical evidence available on the issue tends to support the formalist hypothesis rather more strongly than the opposing positions. Moreover, we agree with Van Peer, who argues that the theory of literariness "describes and explains a number of fundamental issues of literature in a powerful and elegant way" (Van Peer, 1995, p. 315). As we remarked at the outset, however, to adopt this position is not to return to the old, dogmatic formalism of the affective fallacy (Wimsatt and Beardsely, 1954), but to move forward to more fruitful ground -- that of research with actual readers. Our brief review of the debate and some of the evidence, suggests that such research might profitably focus on several related areas. We state these in terms of the following propositions:
1. Literary texts possess distinctive properties, which in the present context we have taken to be foregrounding (but further research will perhaps define more widely). There is, of course, no definable cut-off point beyond which a text is non-literary; rather, foregrounded features occur in a continuum of texts from the clearly literary to the clearly non-literary.
2. Recognition of foregrounding, i.e., the treatment of a text as literary, depends on a reader's linguistic competence, not on literary experience or training.
3. The encounter with foregrounded features plays a formative role in the interpretive effort of a reader. This is, of course, unlikely to be the only influence: text genre, narrative features, etc., also play a major role, varying according to context.
4. The kind of text processing initiated by foregrounding is distinct from that modeled in discourse processing theory, involving schema creation rather than schema extension or modification (cf. Miall and Kuiken, 1994b, Miall, 1989).
Each of these positions has some support from empirical research, but in general little research has been carried out. Thus, we make the usual and predictable call for further research. Why, however, would such research matter to the broader community -- especially the community in literature departments that we discussed earlier, which has largely written off the formalist position as untenable?
In brief, we would argue that literary reading takes its place within a larger cultural ecology, with roots that go far back in history and prehistory, and which has adaptive consequences in the life of the individual and culture of which we are, perhaps, hardly aware as yet. At the same time, the value of literary reading is conferred by a kind of contract that the reader makes with a text: by treating it as a single, albeit complex, communicative experience, the reader comes to recontextualize or redefine some significant aspect of experience. A reader taking up a literary text thus makes several related commitments that guide the act of reading. Let us call these the Formalist Contract. At this stage in our understanding, there are four components in this contract that we would identify, as follows:
1. The bounded text
During reading the reader treats the text as a whole thing, bounded and complexly interconnected. We would set this against the more recent dispersion model, from Barthes (1977) to Landow (1992).
2. Communicative intent
The reader is prepared to assume that new understanding can be acquired from reading, that the act of reading can in itself be made creative by the encounter with the text. This opposes the "always already" of poststructuralism in which it is argued that a text can only exist by replicating existing textual fragments and strategies.
3. The openness of reading
Foregrounding in the text, among other features, arouses memories, feelings, a sense of self, of empathy, etc., which the reader places at the disposal of the reshaping functions of the text. A corollary of this position is that readings will vary, often in major ways, between individuals from the same community. This opposes Fish's (1980) notion of the interpretive community.
4. The adaptive function of literature
If the dynamics of literary reading lie in defamiliarization, anticipation, schema formation, etc., then the primary function of literary reading is to equip us to better understand and respond to our environment. Literature is able to do this by invoking and reshaping our feelings "offline," as it were, that is, in isolation from behaviours and actions in the every day world that have real consequences.
We cannot help concluding that this view of literature is not only more
constructive, but more probable, than the view that has been emerging from
the advocates of poststructuralist literary theory.
References
Andringa, E. (1990). Verbal data on literary understanding: A proposal for protocol analysis on two levels. Poetics, 19, 231-257.
Barthes, R. (1977). Image-music-text, trans. Stephen Heath. London: Collins/Fontana.
Birkerts, S. (1994). The Gutenberg elegies. New York: Fawcett Columbine.
Culler, J. (1981). The Pursuit of Signs. London: Routledge & Kegan Paul.
Durant, A., & Fabb, N. (1990). Literary studies in action. London: Routledge.
Fish, S. (1980). Is there a text in this class? The authority of interpretive communities. Cambridge, MA: Harvard University Press.
Hanauer, D. (1996). Integrations of phonetic and graphic features in poetic text categorization judgements. Poetics, 23, 363-380.
Harari, J. V. (1979). Textual strategies: Perspectives in post-structuralist criticism. London: Methuen.
Hoffstaedter, P. (1987). Poetic text processing and its empirical investigation. Poetics, 16, 75- 91.
Landow, G. (1992). Hypertext: The convergence of contemporary critical theory and technology. Baltimore: John Hopkins University Press.
László, J. (1990). Images of social categories vs. images of literary and non-literary objects. Poetics, 19, 277-291.
Miall, D. S. (1989). Beyond the schema given: Affective comprehension of literary narratives. Cognition and Emotion, 3, 55-78.
Miall, D. S., & Kuiken, D. (1994a). Foregrounding, defamiliarization, and affect: Response to literary stories. Poetics, 22, 389-407.
Miall, D. S., & Kuiken, D. (1994b). Beyond text theory: Understanding literary response. Discourse Processes, 17, 337-352.
Miall, D. S., & Kuiken, D. (1995). Feeling and the three phases of literary response. In G. Rusch (Ed.), Empirical approaches to literature: Proceedings of the Fourth Biannual Conference of the International Society for the Empirical Study of Literature -- IGEL (pp. 282-290). Siegen: LUMIS-Publications, Siegen University.
Pavio, A. (1985). Mental representations: A dual coding approach. New York: Oxford University Press.
Schmidt, S. (1982). Foundations for the empirical study of literature: The components of a basic theory, trans. R. de Beaugrande. Hamburg: Helmut Buske Verlag.
Shklovsky, V. (1965). Art as technique. In L. T. Lemon and M. J. Reis, trans., Russian formalist criticism: Four essays (pp. 3-24). Lincoln, NE: University of Nebraska Press. (Original work published 1917)
Van Dijk, T. (1979). Advice on theoretical poetics. Poetics, 8, 569-608.
Van Peer, W. (1995). The empirical study of (literary) texts. In G. Rusch (Ed.), Empirical approaches to literature: Proceedings of the Fourth Biannual Conference of the International Society for the Empirical Study of Literature -- IGEL (pp. 313-318). Siegen: LUMIS-Publications, Siegen University.
Wimsatt, W. K., & Beardsley, M. (1954). The affective fallacy. In W. K. Wimsatt, The Verbal Icon (pp. 21-39). London: Methuen.