1Intratextuality
Intratextuality is the phenomenon and the study of the relationship between elements within texts: it is concerned with structures such as ring composition, continuities, discontinuities, juxtapositions, story arcs and other repetitions of language, imagery, or idea, including gaps both in the hermeneutic circle and in the form of absent presences and roads not taken. It is interested in the problem of how texts are put together, by authors and readers, as unified wholes, or occasionally in creative disunities, and divided up into sections for ease of consumption or for other purposes. It interacts closely with intertextuality (relationship between texts), paratextuality (the edges of texts which create one form of interface between the text and the world),1 and with what might be called âextratextualityâ, that is, the way texts point outside of themselves, whether to real or fictional worlds. Like intertextuality, intratextuality is concerned, crucially, with the interpretability of links, contrasts, and other textual elements once they have been identified. Just as not all source criticism is intertextual, so not all juxtaposition is intratextual: rather, it becomes so when it functions as the basis for interpretation. Intratextuality is not inherently limited to one side or other of the hermeneutic battle (if so it is) between author and reader, since it is as validly concerned with echoes and structural patterns identified as signs of (and even evidence for) the controlling mind of an author2 as it is with both widespread and idiosyncratic tendencies towards divisional (âbite-sized chunksâ) and unificatory strategies in the phenomenology of reading. Since intratextuality begs questions about how the literary mind works, it might seem profitable to consider whether any light can be shown by other forms of investigation into the human mind.
2The âcognitive turnâ
In recent years, particularly since the new millennium brought about an artificial but powerful barrier between the 20th and 21st centuries, scholars interested in Theory, myself included, have been wondering where we go next, now that post-structuralism has undergone a mixture of debunking and naturalisation, and totalising theories have become somewhat tarred, at least in their most explicit forms, with the damaging brush of 20th century totalitarianism. I remain convinced that one of the most interesting questions to be asked in our field is how meaning happens, what counts as a valid reading, how we communicate it, and what we think it is that we are doing with our 2000-year-old texts. Some scholars in recent years have looked to the cognitive sciences for help in exploring the phenomenon of literature.
The terminology of the âcognitive turnâ, however, is (tellingly) unfortunate, both in that the language and its underlying actuality pertain to a kind of insecure aping of our scientific masters in the contemporary STEM world, and also by an implication of radical newness, as if reading, writing, and telling stories had not always been cognitive processes. The history of literary scholarshipâs interaction with a range of relevant disciplines (primarily cognitive linguistics, cognitive psychology, and neuroscience) can be and has been told in a number of different ways. For me, the best work in cognitive literary criticism is that which is most firmly grounded in its literary history and which tempers its claims for the newness and significance of what cognitive sciences can contribute to the understanding and appreciation of literature. My story of cognitive literary criticism would be the one which traces its history in reader-response criticism, including especially the phenomenology of Iser 1974, whose work is regarded by some as cognitive avant la lettre.3 I would single out in particular a book by Terence Cave, a literary critic with extensive knowledge of âcognitive approachesâ, but free from the fundamentalist evangelising that mars some studies in this area. He begins: âit [cognitive literary criticism] should be a literary criticism worthy of the name, resisting integration into extraneous agendas for which literature would provide mere illustrative examples, but at relevant points inflected by frames of reference, terminology, conjectures, and the like drawn from across the spectrum of cognitive disciplines.â4
âCognitive approachesâ can be divided into two broad categories: scientific/empirical and reflective. The first includes both neural imaging (science) and participant-observation/questionnaires (social science), while the second (reflective) is closer to traditional humanities methodologies but draws in a secondary manner on insights from other disciplines. The reflective methodology is, I believe, both the most practically useful for Classics and yet also holds the greatest risks of overstatement. As a number of scholars have warned, transference of (perhaps improperly understood) concepts from one field to another may be at best metaphorical and at worst misleading.5 The question for this paper is what, if anything, âcognitive approachesâ can offer towards greater understanding of how texts fit together and of the work of readers in constructing interpretations of texts. While this paper cannot hope to do justice to the full range of literary readings that could justifiably be called âcognitiveâ, the following issues have seemed to me to be of relevance to intratextual enquiry. What do we mean when we talk about âreadingsâ, and about the engagement of âreadersâ with texts? Can insights about the workings of memory offer anything to intratextual study? How can recent formulations of the embodied and extended mind reflect on our work as critics of ancient texts?
3What is a reader?
Allow me for a moment to draw you into a fantasy, which I hope will illustrate not only what neuroscience cannot do but also what counts as the reading of an ancient text. Imagine, if you will, a research project, in collaboration between a classical literary critic and a real live scientist, in which a large number of people are wired up to an fMRI machine, while they read some Latin poetry, to see whether their brains light up in the same way. To take an intratextual example, we might like to try to find out what proportion of people have a brain-reaction, without being told to look out for it, to the resonance between dum conderet urbem at Virg. Aen. 1.5 and ferrumâŚcondit at 12.950. If so, would this actually prove the rightness of a reading which interprets the death of Turnus as the inevitable but despicable consequence of Roman imperialism?
Of course not. Not only would there be all sorts of practical difficulties in such an experiment, but also the recognition of a connection is only a preliminary to interpretation, which in such a case could go in any of several different, even opposing, directions. Neural imaging is very non-specific (shared by a vast range of activities, in animals as well as humans) and is currently unable to tell us anything about the subjective experience of literature.6 My example is a parody, but it raises questions that I think are important.
The greatest benefit for me in considering opportunities and threats raised by the empirical study of literature is being forced to think in detail about what we mean by a âreaderâ. On the empirical side of cognitive literary studies, scholars attempt and discuss experiments with âciviliansâ as readers, rather than literary specialists. For us readers of ancient texts, in a language which has no native speakers, we would have an additional difficulty in any such undertaking, arising from the fact that we have no âlayâ or âcivilianâ readers.7 Even if such an experiment as my disingenuous bit of Virgilian ring composition were possible, for any outcome to be meaningful â in the statistical terms on which such experiments are predicated â one would need to have a large evidence-base. Our first problem would be that although we could in theory wire up all the competent Latin readers in the world and probably thereby acquire a meaningful sample, it is impossible to imagine someone with sufficient linguistic competence who would never have heard of the possibility of ring composition, or of the debate surrounding the end of the Aeneid. Although one might be able to design questionnaire-based experiments which would have the potential to create useful data, we would not be able to test them on lay readers. But useful for what? I can imagine that it might be interesting to see how widely a particular reaction is shared, whether the information is sought by neural imaging or by self-reported response, but I suspect that the interest would be greater for the student of contemporary psychology and sociology than it would be for the literary critic. If it were possible (currently, I believe, it is not) to get meaningful data on all competent readers over a period of time, the extent to which those data change over time might have a contribution to make to literary theory about the embeddedness of any particular reading in its own historical moment. Most likely it would only tell us what we already believe to be the case.
But the crucial point, and the fundamental difference between empirical studies of readers and literary criticism, is this: even if we could devise an experiment using readers who are neither specialists nor students (or former students), how far would we regard the outcome of their brain-responses, whether neural or self-reported, as âproofâ of the rightness of the interpretation? I suggest that the answer is â not very far at all. Although some people might feel that this is an elitist and even arrogant attitude, my intention is not to undermine the validity of any individualâs reaction to an ancient poem. Rather, it is to describe what we actually do. Burke and Troscianko claim that the benefit of cognitive literary science (their term) is âthat instead of basing conclusions about textual effects on the singular experience of the critic-as-reader disguised as the generic reader, or accumulating new interpretations without acknowledgement of the cognitive factors on which they depend, we can understand interpretations as cognitive effects, and investigate their natural variations in others as well as ourselvesâ.8 I am not convinced that cognitive literary science is able to deliver in this way, but, be that as it may, my argument in this paper is intended to suggest that the experience of the critic-as-reader is neither âsingularâ nor âdisguised as the generic readerâ. Anyoneâs reading might be valid for him or herself, but readings which we come to accept as powerful and original (and therefore publishable) tend to be precisely those which stand outside the cloud.9
The ancient reader, obviously, cannot be either wired up to an fMRI machine or asked to fill in a questionnaire. Once we begin to think about real readers, we must consider the question of how far our âreaderâ is to be identified with an âancient readerâ. There are some 21st-century scholars of classical literature who would say that what they are interested in is the âoriginal readerâ, some who regard their main interest as the âmodern readerâ, and some who would want to think of a complicated, perhaps fuzzy, but often creative, combination of the two. I regard all these approaches as valid in their own terms, but as requiring careful scrutiny as to their underlying assumptions. For example, when we start thinking about readers in experimental terms, we have to become much more explicit about the evidence-base. By âancient readerâ of the Aeneid, do we include every competent reader of/listener to the Latin language from the 20s BC up to the fifth century AD? Or do we mean the elite group of people who may have been present at preliminary readings of parts of the poem before its completion, if indeed those took place? Virgilâs fellow poets and friends? All first century BC readers? My contention, to cut it short, is that all our so-called readers are hypothetical. It may indeed be critically acceptable for us to construct details of and constraints around our hypothetical reader, and (this is crucial) to do so differently for different purposes, but we should not deceive ourselves into imagining that we could ever meaningfully describe a real ancient reader.10
In conversation with my colleague Andrew Morrison, I have come to the following conclusion about the âreaderâ, as regards interpretation, one which would entail a stronger rejection of the role of empirical initial responses by âreal readersâ: interpretation is more than reading, and âreadersâ in criticism are always heuristic tools, which we use to help us conceptualise the effects of a text, and in turn to help create lar...