Previous Section Contents Page  



10) Conclusions

Pictures don’t tell stories; viewers tell stories. Art historians also tell stories, or at least we try. As with any story-telling genre, a PhD thesis is expected to share certain structural characteristics with all those examples that preceded it. It should have a beginning that promises much, followed by an interminable middle, meandering off on frequent digressions, and a distinct ending – some kind of conclusion that summarizes all of what went before. Peregrinations that once appeared promising and which swallowed up valuable research time before leading nowhere are normally tacked on after that conclusion as copious (often unread) appendices, alongside collections of primary data to show that one really did do the legwork in dusty archives and icy cathedrals.

This particular thesis does not deviate substantially from the genre norms, though it does have a few quirks. The early realisation that ‘marginal’ things were all too often central to my work, together with an absence of primary material that needed citing, led to an early appendectomy, while the need to sweeten the pill of too much theory with plenty of practical examples (along with an uncontrollable fascination with some of the artworks involved) resulted in the inclusion of more case studies than might have been strictly necessary, often under the questionable guise of the ‘excursus’.  All that remains, to fulfil the needs of the {thesis} genre, is a conclusion of sorts. To some extent the real ‘conclusion’ of any academic work is not in the words of the final chapter but in the reception of what precedes them and the extent to which it nuances subsequent discourse. Nevertheless, for the sake of genre compliance if nothing else, I will attempt to provide closure to this meta-narrative by framing what has gone before with a conclusion of sorts. If not a paragon among conclusions it will at least be a parergon.

This thesis had three main aims; to review the historical and current state of narrative theory within the literary-studies tradition; to adapt those theories, where appropriate, from verbal to visual narratives (creating, if not a coherent methodology, then at least a vocabulary or meta-language with which to discuss narrative images) and to present a series of case studies; the latter intended partly to raise some interesting issues about framing and narrative structure but also, more generally, to demonstrate how visual narratology can add an additional layer of analysis to more traditional studies. The key challenge in this (as in any attempt at enforced interdiscipliniarity) has been in resisting the temptation to introduce ‘new’ theories for their own sake but rather to ask of each ‘what does this imported concept offer us?’ The various examples and case studies scattered through the preceding chapters were all, in part, attempts to answer that question en passant, while chapter 9 was a more concerted effort to show what visual narratology can reveal about the role of the artist in early thirteenth century stained glass design.

[242] The Holy Grail in any study of meta-communication is the genuinely trans-medial and trans-generic phenomenon - that is to say a meta-communicative device which develops independently  in diverse media and genres rather than simply being copied verbatim from one object to the next. It is for this reason that I have focused so much on the practices of metalepsis and other frame violations in various media, since these really do fit that requirement. In the case studies these practices were examined in relation to stained glass, sculpture and manuscript illumination, sometimes in examples which are clearly unrelated to each other - but also in groups of objects (the ‘royal Psalters’ and Bibles Moralisées for example) where subsequent artists working within a recognisable tradition have copied its use - not through the simple repetition of a visual motif but independently, each deploying a common meta-communicative device in a different visual form. Thus the examples considered from amongst the ‘royal Psalter’ prefatory cycles (see chapter 6.c) each feature one example of metalepsis - but each artist included a different metaleptic element and chose a different context in which to deploy it. By contrast the artists of the Toledo Bible Moralisée copied a single metaleptic motif from the earlier manuscript (draught animals wandering out of the frame) but included it in a range of scenes where it had not been used by their predecessors. Each of these examples is evidence of artists consciously using a device which is not part of the narrative content but which serves to modify how the viewer ‘reads’ the narrative.

Naturally there wasn’t room in this thesis for every topic within the ever-broadening field of narratology. Text-oriented narratologists like to distinguish between the ‘disnarrated’ - passages describing those events which have not taken place - and the ‘unnarrated’ -  events which are deliberately not mentioned in a story even though one might have expected the narrator to include them (their exclusion often foregrounding a significant reluctance to mention something). 1 The disnarrated is just one of a number of narratological concepts which remained unnarrated in this thesis, not because I didn’t regard them as important or interesting (Charlemagne’s peccatum innominandum is an example where ‘the unnarrated’ would be highly relevant) but because I felt that rather than aiming for a truly encyclopedic Summa narratologica it would be better to explore a limited range of narratological concepts in more detail. Such has been the proliferation of topics in narratology in the last thirty years that a great many things had to remain ‘unnarrated’ - or else were skimmed over in the most cursory fashion. Many of these areas may yet reward further consideration. One could generally do a lot worse than simply opening  the excellent Routledge Encyclopedia of Narrative Theory (Herman, Jahn & Ryan, eds.) on any random article  and considering whether the concepts therein described might usefully be applied to medieval narrative art - and if so, how. The [243] exercise of considering a concept’s suitability for one’s own field can often be every be as useful and revealing as the fully developed concept itself.

As well as skimming over or ignoring several important narratological terms in this thesis, I have left my definitions of others deliberately vague. Despite the considerable efforts that have been devoted to such activities over the last half century, it is ultimately pointless to worry too much about defining precise scales of ‘narrativity’ - or to become obsessed with finding the ultimate, all-inclusive definition of ‘narrative’. Such definitions or scales have value only insofar as the process of trying to refine them helps us to elucidate the features or characteristics that lead a receiver to interpret a communicative structure as being more or less narrative. If we are ever to progress beyond the taxonomic fetish that made ‘structuralism’ a dirty word in some academic circles, such definitions should always be treated as means rather than ends.

Inscribed over both the entrance and the exit of this thesis are the words “pictures don’t tell stories; viewers tell stories”. As art historians we routinely use phrases such as “these wall paintings tell the story of Christ’s Passion” but such statements are misleading. Perhaps what one should say instead is something like “these wall paintings help the viewer to tell/recall/contemplate/discuss/visualise the events of Christ’s Passion”. Even that would to some extent overestimate the determinism of the image. The ‘meanings’ of visual narratives are never fixed but are transient and labile, constantly being re-negotiated between image and viewer. They are the product of the viewer’s changing experiences and expectations as much as they are controlled by the intentions or wishes of the patrons and designers of the images. Bakhtin saw religious texts as representing precisely the kind of ‘authoritative discourse’ that would be challenged or undermined by the heteroglossia of the novel but from the moment their production moved beyond the cloister and into the hands of lay craftsmen, medieval religious images were always and everywhere heteroglossic. Moreover the shifting of indeterminacy between the textual and the visual (see chapter 1.c) means that when stories are expressed through images rather than words, they have to make concrete a range of details (clothing, facial expressions, gestures, etc.) that can safely remain unstated when the same story is told verbally - but which each carry meanings of their own. These visually concretized details naturally tend to be the semiotic elements that are least canonical or fixed in their meanings and most open to interpretation by the individual viewer. They are also the details most likely to evoke thoughts in the viewer’s mind of unrelated matters, their quotidian familiarity sparking of a chain of semiosis entirely beyond the imagination of their original authors.

Thus the role of the viewer in consuming visual narratives is generally more active and less controllable than is the role of the reader (or auditor) in consuming verbal narratives. Indeterminacy is both the weakness and the strength of narrative images. It shifts power from the maker to the viewer - but it also draws the latter into a more active mode of reception. The [244] indeterminacy of narrative images liberates and ensnares in equal measure. From the mid-thirteenth century reminiscences of Cardinal Eudes de Châteauroux, recalling his uncertainty before the glazed stories in the cathedral and the readiness of a young layman to interpret them for him, through the ‘counterfeting gentilmen’ in the fifteenth century Beryn’s Tale, who bicker over the stories in Canterbury Cathedral’s windows, to those recent historians who see every image of a chivalric knight as representing the particular epic hero they have chosen to study; each viewer weaves his or her own story from the narrative images they encounter - and then seeks to explain their interpretations to anyone who will listen.2 My aim in this thesis has not been to weave yet another version of a familiar story or two but rather to foreground the techniques underlying the weaving process - to look beyond the weft and expose the warp threads that supported medieval narrative art.


1 See G. Prince, ‘The Disnarrated’ in Style, 22:1 (1988), p.6. For the ‘unnarrated’ see R.R. Warhol, ‘Neonarrative; or, How to Render the Unnarratable in Realist Fiction and Contemporary Film’ in Phelan, J. and Rabinowitz, P. J., eds., A Companion to Narrative Theory, Oxford 2005, 226-31.

2   For Eudes de Chateauroux’s childhood experience, see Kemp, W., The Narratives of Gothic Stained Glass, Cambridge 1997 [1987], pp.71-2. For Beryon’s Tale see Bowers, J. M., ed. The Canterbury Tales: Fifteenth-Century Continuations and Additions  Kalamazoo, 1992, p. 64


Previous Section Contents Page