[137]

5) Diegetic levels and their visualisation

After the pleasantly practical interlude of Chapter 4, I will now plunge back into the realms of theory in order to consider the forms of narrative embedding which are an endemic but habitually overlooked feature of the narrative arts, both verbal and visual. We saw several examples of this in the Charlemagne window, in the discussion about the milky way and in the having and telling of dreams, while another set of dreams will be considered in detail as part of my ‘paradigmatic excursus’ in chapter 8.b. First however I must explore some of the narratological theory behind what is known as ‘diegetic embedding’.

In a discipline overly fond of neologisms, the term ‘diegesis’ has an unusually long pedigree. Plato, in Book 10 of his Republic, distinguished between two modes of story-telling; diegesis or ‘indirect presentation’ (telling or recounting something) and mimesis or ‘dramatic imitation’ (showing or re-enacting). That distinction has, in essence, been retained. Such terminological conservatism has not however restrained the natural tendency of narratologists to complicate matters and the word ‘diegesis’ has recently attracted a bewildering array of prefixes, notably ‘extra-’, ‘intra-’, ‘hetero-’, ‘homo-’, ‘iso-’, ‘meta-’, ‘hypo-’, ‘auto-’ and even ‘pseudo-’. 1 Fortunately I need concern myself here only with the first four of these compounds - the minimum set for considering diegetic levels within biblical and hagiographic texts and their implications for visual narratives.

At its simplest, diegetic level is a measure of the extent to which a narrative is embedded or nested within other narratives. For example, a deeper level of nesting occurs when a character within a story relates a different story to the other characters. In such a case, the default story-world (in which the character tells his story) is described as the extradiegetic level, while the world of the story-within-the-story is described as intradiegetic. Such intradiegetic segments within a narrative can be of almost any length. At their longest they conform to the so-called ‘frame narrative’ model, in which the bulk of the story (or more often, a series of independent stories) is told at the intradiegetic level, packaged within the thinnest of extradiegetic wrappers. This form was particularly popular in medieval literature as a way of bundling together compendia of shorter (and generally unrelated) stories - Chaucer’s Canterbury Tales and Boccaccio’s Decameron being the most obvious examples.

[138] Having dealt with the position of the narrator in terms of the depth of diegetic embedding, the other key axis of categorisation concerns the participation of that narrator in the story-world. In this regard Gerard Genette distinguished between the homodiegetic narrator, who is an active participant, and the heterodiegetic narrator, who is not. 2 Homodiegetic narration is characterised by the use of first-person deictic pronouns to denote the speech, thoughts and actions of the narrator, who is normally in a state of ignorance with regard to the thoughts and intentions of the other characters in the story. By contrast, heterodiegetic narration is characterised by an external, omniscient (and often only implied) narrator who is party to, and hence in a position to report on, all that happens in the story-world.

In terms of the biblical texts, there are a few homodiegetic sections in the Gospels, notably those sections in the Acts of the Apostles where Luke, present in the story as a companion of St Paul, uses first-person pronouns (the so-called ‘we-passages’). Although they form only a small part of the Opus Lucanum, it has been argued that these ‘eye-witness’ references help to lend verisimilitude to the Gospels as a whole. 3 Homodiegetic narration is also important in some hagiographic texts (and occasionally also in their visualisations) in relation to the witnessing and validation of events, particularly miracles. Again, the inclusion of the homodiegetic narrator or the author in the visualisations of such stories is an important modality marker, lending the somewhat spurious authority of the eye-witness to a scene. We have already seen one example of this in the Charlemagne window at Chartres; the depiction of Archbishop Turpin (the supposed author of the narrative therein depicted) in the scene of Charlemagne and his armies heading off to war in the opening scene of the first Spanish campaign (panel 10 - see Fig. 4.10). Perhaps the most common and the most important examples of visual homodiegesis in medieval art however appear in depictions of the Book of Revelation, a high proportion of which include John (often depicted outside the frame or else in some way separately framed within it) in the process of ‘witnessing’ his Apocalyptic visions.

Leaving aside Revelations and Luke’s ‘we-passage’, the Bible’s normal narrative mode employs an unnamed, implied narrator, who is heterodiegetic (i.e. not a named participant) and extradiegetic (i.e. viewing the action ‘from above’) addressing the implied reader to report the events of the story directly. Periodically this extradiegetic discourse is interrupted by momentary intradiegetic fragments, as the narrator temporarily hands the baton to one of his characters to describe events. This is normally signalled by a switch to reported speech, direct speech or even free indirect discourse (all of which occur within the Bible to varying degrees). It could be argued that any use of direct speech in a written text is mimetic, [139] rather than (intra-)diegetic (i.e. the speech is ‘enacted’ rather than described). However this view was opposed by Aristotle himself, who argued that nothing short of a dramatic re-enactment, with the words spoken by an ‘in-character’ actor, should be regarded as mimesis. This position was reiterated by Gerard Genette in his argument that ‘...narration is a fact of language, and language signifies without imitating’. Hence ‘narrative mimesis’ in a text or an image is always illusory - and locutionary acts within the text should always be seen as embedded narratives. 4

As well as locutionary episodes, signalled by the normal speech-reporting verbs (‘said’, ‘told’, ‘shouted’, etc), I also include in this category cases where the subordinate clause is governed by verbs of cognition (‘dreamed’, ‘saw’, ‘witnessed’, etc.), since in each case there is the same implicit shift in the narratorial voice. Although this may seem like unnecessary complexity, the matter of narrative position is of great importance in determining the reader or viewer’s assumptions about narratorial reliability. This is particularly relevant when the dominant narrative mode employs an omniscient, extradiegetic narrator, as is the case in the Bible and most hagiographic texts. In this mode, a simple statement of events by the narrator is considered to be absolutely true (at least within the ontological and epistemological systems of the story-world). By contrast, a statement of events made by one of the characters within the story may be either true, false, or uncertain. There is, in other words, a world of difference between the following phrases concerning some event ‘[z]’ appearing in an extradiegetic narrative (words attributed to intradiegetic narrator [x] are italicised):

a) ‘[z] happened’,

b) ‘[x] told [y] that [z] had happened’ or ‘[x] exclaimed “[z] has happened!”’

In case a), the reader has no doubt that [z] actually has happened. In case b) however the reader believes an utterance has been made - but their view on whether or not the utterance was truthful (and hence whether [z] really has happened) depends on a range of factors - not least their prior knowledge about the morals and motivations of the character [x]. Take for example the segment of Luke, Chapter 22, where Peter denies Christ:

⁵⁶But a certain maid beheld him as he sat by the fire, and earnestly looked upon him, and said, This man was also with him. ⁵⁷And he denied him, saying, Woman, I know him not. ⁵⁸And after a little while another saw him, and said, Thou art also of them. And Peter said, Man, I am not. ⁵⁹And about the space of one hour after another confidently affirmed, saying, Of a truth this fellow also was with him: for he is a Galilaean. ⁶⁰And Peter said, Man, I know not what thou sayest. And immediately, while he yet spake, the cock crew.

[140] In these verses, the words uttered by Peter and the other people sitting by the fire (which I have italicised) are all intradiegetic fragments, presented as reported speech within an otherwise extradiegetic narrative. Because of the conventions regarding the truth-value of the biblical narrative, those who take the Gospel as gospel will accept that these words were definitely spoken (i.e. the locution was truthfully reported), just as they will accept that the fire was burning and that the cock crew. However, anyone who had been following the story up to this point would know that the intradiegetic fragments belonging to Peter (italicised and underlined in the extract above) contained lies. It is precisely this recognition of intradiegetic embedding by the reader that allows different modalities to apply to different parts of a narrative text in this way. Intradiegetic elements, however subtle, are of great importance in the building of narrative discourses, whether textual or visual. They normally fulfil one of four distinct narrative functions, each of which will now be addressed in turn, with an emphasis on their implications for visual narratives.

5.a) embedded narratives

For the purposes of this discussion, an embedded narrative can be defined as any intradiegetic episode which itself displays the characteristics of narrativity as described previously in chapter 2 and which, unlike the fragments of reported speech discussed in the previous section, constitutes a valid segment in its own right (i.e. it contains at least one nucleus). Perhaps the most obvious examples of embedded narratives in a biblical context are Christ’s Parables - and these provide a useful example for illustrating the general concept of diegetic level and its associated terminology.

Luke’s Gospel abounds in embedded narratives but as an example, let us take the story of the Good Samaritan (Luke 10:30-35), which Christ relates to ‘a certain lawyer’ who has asked him what to do to ensure eternal life:

²⁹But he, willing to justify himself, said unto Jesus, And who is my neighbour? ³⁰And Jesus answering said, A certain man went down from Jerusalem to Jericho, and fell among thieves, which stripped him of his raiment, and wounded him, and departed, leaving him half dead. ³¹And by chance there came down a certain priest that way: and when he saw him, he passed by on the other side. ³²And likewise a Levite, when he was at the place, came and looked on him, and passed by on the other side. ³³But a certain Samaritan, as he journeyed, came where he was: and when he saw him, he had compassion on him, ³⁴And went to him, and bound up his wounds, pouring in oil and wine, and set him on his own beast, and brought him to an inn, and took care of him. ³⁵And on the morrow when he departed, he took out two pence, and gave them to the host, and said unto him, Take care of him; and whatsoever thou spendest more, when I come again, I will repay thee. ³⁶ Which now of these three, thinkest thou, was neighbour unto him that fell among the thieves? ³⁷And he said, He that shewed mercy on him. Then said Jesus unto him, Go, and do thou likewise.

[141] Here then we have two distinct narrators operating at two diegetic levels; Luke’s overall account of events and Christ’s telling of the parable (the story within the story, shown in italics in the extract above). In this example, Luke is an extradiegetic narrator (and heterodiegetic too of course, since this is not one of the ‘we-passages’). Christ is intradiegetic and homodiegetic in relation to the Gospel but He is heterodiegetic in relation to the parable (since He does not appear in it). It should be clear from that last caveat that these qualifying terms like intra-, homo-, etc are always relative to some narrative ground level or other. Note also that the reported-speech of the Samaritan within the parable (underlined in the quotation above) could be seen as occupying an intradiegetic level one lower than the parable itself (level -2). Thus an evaluation of its associated truth-value (how likely the inn-keeper is to be recompensed further) depends on what the narratee feels about the honesty of Samaritans. The positions of the various participants and levels in the extract given above can be summarised thus (Table 5):

Table 5 - Diegetic levels and participants in the Good Samaritan parable

The story of the Good Samaritan appears quite regularly in medieval art, particularly in stained glass, where it often provides a framework upon which to hang Old Testament typologies. Normally however, when this, or other parables are shown in visual form, they are treated as being fully self-contained narratives and the framing story is ignored. A unique visualisation of the parable of the Good Samaritan which does include its framing narrative (or at least the opening frame) can be found in the eponymous window at Chartres (bay 44), where panel 04, the first panel after the signature panels, shows Christ as the intradiegetic narrator, recounting his story to the lawyers (see Fig. 5.01) - a scene elegantly described by Wolfgang Kemp as a ‘rare “incipit” formula of Christ as narrator’. 5 The contents of the parable are then depicted in panels 05-12, before switching to its Old Testament antetypes in panels 13-23. The change of diegetic levels between panels 04 and 05 is certainly not telegraphed - though it is perhaps subtly indicated. Christ sits to the right of the first frame, his hand raised, making a gesture that can be used either for blessing or as a discourse marker. Two Pharisees (clearly indicated by a titulus beneath them reading “PARISEVS”) sit on the left, a little way apart from Christ on the same long bench. One of them looks away towards his colleague and fingers the strap of his cloak. As we have seen (chapter 4 regarding bay 07, panel 04), this is a multivalent sign in the [142] Chartres windows but often it connotes nervousness or uncertainty. Unfortunately the head of the other man, whose body seems to be turning to his right, is a rather clumsy stop-gap - so we will never know whether he looked towards Christ or towards the next panel. The only clue to the real nature of this scene is that Christ’s eyes are slightly lifted - he looks not at his interlocutors but over their heads and towards the first scene of the parable in panel 05, where the ‘PEREGRINVS’ is about to depart from Jerusalem. This upwards glance may be scarcely noticeable when viewed in isolation but it stands out clearly when compared to those other panels in this window depicting conversations, where gazes are always aimed directly at interlocutors.

5.b) reporting the perceived

A more common need for the visualisation of intradiegetic elements is in indicating the contents of perception, whether of direct external experiences or of dreams or visions (oneiric content). Frequently in the Bible, it is necessary for one character to inform others of something they have either seen or dreamt - such as when Mary Magdelen informs the disciples that she has been to the tomb and found it empty (John 20:2) or when Pharaoh describes his dreams to Joseph (Gen 41:17-24). As Sixten Ringbom pointed out in one of the few dedicated studies of this topic, this kind of illocutionary act was traditionally indicated by the simple contiguity of two scenes within a picture cycle.6 One scene shows a character experiencing something, the next shows the same character (now effectively in the role of intradiegetic narrator), addressing one or more other characters; the implication being that he/she is telling them about the event depicted in the preceding scene. In biblical illustrations we find innumerable examples of this practice. Ringbom (ibid, p.42) quotes Otto Pächt’s example of the Apostola Apostolorum in the St Albans Psalter (see Fig. 5.02A). Here, on one opening, the left page shows the three Maries at the empty tomb, while the right page shows Mary Magdalen, index finger raised, addressing the Apostles - the implication of this juxtaposition being that she is telling them what she has witnessed in the preceding scene (the meaning in this particular pair of images is made even clearer by the fact that on both pages Mary is shown in a similar pose and in the same location on the far left of the frame).

The other model for depicting both the act and the contents of perception is by showing two distinct episodes juxtaposed within a single image - effectively a reversion to the complementary mode of illustration but within a polyscenic framework. Ringbom illustrates this with the example of the ‘Flight to Egypt’ scene in the ‘Oxford-Paris-London’ [143] Bible moralisée (British Library MS. Harley 1527, f.12v) which combines the angel that appeared to Joseph in a dream, and the consequences of that dream; the holy family travelling to Egypt. The artist has shown the family en-route in the customary manner but with a large angel reaching out of a cloud-frill and apparently shoving Joseph along on his way (Fig. 5.02B). As Ringbom points out, this ‘juxtaposition’ method has its weaknesses:

The content of the angel’s message has been made clear enough - but at a price. For without knowing what the whole story is about, or reading the caption there is no way of telling that the miniature represents St. Joseph’s dream, or, indeed, that it represents a dream at all.7

Perhaps in this case the ambiguity was indeed seen as problematic, since the designers of the Toledo Bible moralisée opted for a simpler approach, replacing the angel with a greatly reduced and more conventional vision of God emerging from the cloud, addressing his warning to the voyagers generally and without physical contact (Fig. 5.02C).

The juxtaposition of separate episodes within a single frame in an image cycle effectively became an ‘industry standard’ for depicting dreams and dreamers in medieval art. As Ringbom put it, ‘...a standardized figure, the reclining dreamer, becomes a kind of pictorial quotation mark, an index telling the beholder that the picture deals with a dream’.8

This is why the viewer of the ‘Constantine’s dream’ image discussed in the preceding chapter (Chartres, Charlemagne window, panel 03) would not be unduly worried that the sleeping ruler was about to be trampled by the crusader’s horse. Nevertheless, distinguishing between the act of experiencing a dream or vision, the contents of the dream and the recounting of those contents remained problematic - an issue that will be addressed in detail in chapter 8.b in relation to the ‘having’ and the ‘telling’ of dreams in Genesis 40.

5.c) showing topics of conversation

The preceding examples showed how contiguity could be used to denote that the perceptual experience of one character - something seen or dreamt in one {chronotopic} frame - could be communicated to other characters in the subsequent {chronotopic} frame. A related issue is the problem of indicating the current topic of a conversation; not strictly speaking an intradiegetic element in verbal narratives but one which is similarly dependent on deixis and which therefore behaves like one in narrative art.9

[144] When both interlocutors are shown in the presence of the thing being they are discussing, indicating the topic of conversation is a simple matter of the speaker looking at the other person and gesturing toward the object being discussed. For example in the previous chapter, in the scene where Charlemagne announces that he just wants some relics, he looks at Constantine and gestures towards the reliquaries (Fig. 4.06). This method works reasonably well, even when the topic of conversation is an unfamiliar object or complicated idea. For example in the ‘St Thomas in India’ window at Chartres (bay 23), an angel takes the soul of the King’s dead brother Gad to see the ‘many palaces of Heaven’ which had been built by Thomas’ charitable actions. The ‘many palaces’ mentioned in the Acts of Thomas do not of course refer to physical buildings but are a metaphor based on John, 14:2 ‘In my Father's house there are many mansions’. Nevertheless, the conversation is an easy one to illustrate, as can be seen in Fig. 5.03. The angel is shown facing towards Gad but pointing towards a reification of the ‘many palaces’.10 It is particularly in this reification that these examples begin to resemble intradiegetic embedding - the reified object is an ontologically distinct element embedded within the main image. Unlike a semantic enclave it uses the same semiotic system as the main image (a blend of iconic and conventional visual signs) but it clearly occupies a different level of reality.

A similar type of illustration can also be used to indicate the shared topic of a conversation when that object is a real thing (an existent within the same story-world) - but one that is not physically present in the narrative space occupied by the interlocutors, or else one that is not capable of being represented directly. Take for example the third temptation of Christ;

And the devil led him into a high mountain, and showed him all the kingdoms of the world in a moment of time. (Luke 4:5)

Luke’s allusion to an instantaneous global vision (in the Vulgate; ostendit illi omnia regna orbis terrae in momento temporis) is clearly not something that could be visualised directly. Many artists chose instead to use the alternative reading of ‘riches’ for ‘kingdoms’, though there were also ways of showing this scene without losing the original sense of the text. The third temptation is depicted in the prefatory cycles of two related English Psalters which, between them, demonstrate both methods of illustrating the scene. In the first, a leaf associated with the mid twelfth century Eadwine Psalter (London, BL.Add.37472v), the third temptation appears in a 3x4 grid of scenes - second row down in the third column (Fig. 5.04A).11 Here the artist has shown Christ atop a hill, while the devil, standing down below, points at a pile of golden plates [145] and chalices on the hill between them. A later twelfth century manuscript derived from the same tradition whose prefatory cycle employs the same 3x4 grid of squares is the ‘Canterbury Great Psalter’ - Paris Bibliotheque nationale, Latin 8846.12 Here, on f.2r, the artist has split the episode over two squares but has followed the text more literally (Fig. 5.04B). In the first frame, Christ stands on a hilltop between two devils, who gesture towards the next frame. This adjoining frame is quartered and in each quarter is a stylised representation of city, with a few golden vessels arranged outside. Thus ‘all the kingdoms of the world’ have been synecdochically reduced to four cities, each shown by the conventional sign for a city, such as one might find on a map - but with some treasures thrown in for good measure. Note that here the relationship between physically adjacent frames in the square grid is different from the relationship between frames elsewhere on the page and throughout the rest of that prefatory cycle - the second {picture} frame is not the {chronotopic} frame that follows the first one but instead serves it as a subordinate or hypotactic clause representing a distinct ontological level within the same {chronotopic} frame, reifying the indirect object in the phrase ‘ostendit illi omnia regna orbis terrae.

Another feature of the ‘illustrated topic of conversation’ image type discussed by Ringbom is the use of cloud-frills to differentiate between levels of being. He gave the example of a Bible moralisée image, where God instructs Moses regarding the choice of soft-furnishings for the Tabernacle (see Fig. 5.06). 13 In the illumination, God and Moses stand on the mountain, where God points into a cloud-frilled semicircle wherein are shown the curtains of which He speaks. Moses looks not at his God but into the visionary cloud, thereby showing his understanding of the mutual topic of conversation. Interestingly, when adding the moralisation, the artist has retained both the form and the semantic relationships of the Old Testament image, substituting Christ telling a group of clerics about the many mansions of His Father’s house (John, 14:2). In both cases we find the same division between the earthly vision in the lower part of the image and the ‘spiritual vision’ in the upper, the ontological division marked by the cloud-frill.

5.d) depicting figures of speech

As well as indicating things experienced, this same visual syntax can also be used for what are essentially just ‘word-pictures’. Unlike the Tabernacle curtains or the cities of the world discussed above, these are embedded elements whose contents have nothing to do with the [146] narrative proper. They do not depict anything existing within the story-world, or relevant to it, but instead serve like pictographic captions illustrating figures of speech.

The two Psalter prefatory cycles mentioned in the previous section both feature an example of one such illustrated metaphor - Christ’s enigmatic response to the scribe who wishes to become his follower in Luke 9:58:

And Jesus saith to him: The foxes have holes, and the birds of the air nests; but the Son of man hath nowhere to lay his head.

In the earlier example, the Eadwine leaf (London, BL.Add.37472v), the square frame is split horizontally (see Fig. 5.05A). In the lower half, Christ faces the scribe and jabs his index finger right through the horizontal division and into the upper half of the frame, to point at the foxes and birds in the scene above. 14 In the later version (Paris 8846), the regular 3x4 grid of undivided frames is maintained without interruption and as with the image of the third temptation the artist has used two regular frames. In the first the scribe turns towards Christ, who gestures past him towards the next frame where, again, the saying is visualised in a charmingly literal form. In both these cases, the contents of Christ’s speech-act are depicted in the same visual style as the rest of the narrative - there is no formal or pictorial differentiation to indicate that the well-housed fauna are simply figures of speech and not active participants in the story. Indeed, from the way the artists in both manuscripts have shown the foxes hungrily eyeing the nesting birds, the viewer could easily misconstrue the whole frame and confuse it with something out of Aesop’s fables. Even more than usually, this is an instance where prior knowledge of the text is a prerequisite for understanding the images that ‘illustrate’ it.15

Common to most of these examples, both for figures of speech and topics of conversation, is the same deictic gesture; Christ’s jabbing forefinger in the ‘foxes & birds’, God’s in the conversation with Moses, and Satan’s in the temptation. The extended index finger in medieval art is a multivalent gesture which can denote, amongst other things, seeing, telling and debating. Indeed, the index finger in narrative art can function in any of the three modes of Peircean semiotics; symbolic (as a gesture), iconic (as a finger, e.g. in the incredulity of Thomas), or indexical (as deixis). It is in the last of these - as a deictic marker - that the index finger reverts to its most natural function as a pointer. In the case of these two ‘birds and foxes’ images, the indexical finger serves to clarify that it is this subsidiary image that contains the [147] topic of conversation - and perhaps also suggests that this image is to be construed as hypotactic and not simply as the next {chronotopic} frame.

5.e) conclusions

My earlier discussion of frames (in chapter 3) introduced the idea that distinct ontological levels can be represented within and between narrative images. This chapter was intended to be a brief introduction to one of the practical functions of such ontological differentiation, namely the representation of diegetic embedding. Some of the ideas discussed here will be explored in more detail in section 8.b (an extended case study of several depictions of the various dream sequences and prophecies that form the core of Genesis 40) which should help to clarify the visual aspects of intradiegesis. Diegetic embedding is ubiquitous in textual narratives and has understandably attracted widespread attention from literary theorists but given the importance of embedded narratives and oneiric content in biblical and hagiographic stories it is also a practice that deserves closer attention from those concerned with visual narratives.

Given the space and time limitations which constrain any PhD thesis, a number of interesting avenues for further research have necessarily been ignored in this chapter. Literary theorists and linguists have long been fascinated by the nature of ‘deixis’, the words or tags which relate the content of an utterance to its context – i.e. the relative time, place and person of the senders and receivers of that utterance in relation to its subject. It might therefore be worth considering the equivalent deictic functions within medieval narrative images, both in terms of receivers within the image (for example Christ’s interlocutor in the ‘Foxes have holes’ parable) and the external viewer. The deictic relationship between internal and external receivers is particularly interesting in relation to those issues of embodiment and eye-witnesses (or ‘viewer-surrogates’) that have been explored by historians working in the early-modern period. Other fruitful topics that deserve further attention might include the development and use of the kind of ‘figures of speech’ images discussed above, the general characteristics of oneiric content in medieval narrative art, and perhaps most usefully, an expansion and updating of Sixten Ringbom’s fascinating but rather superficial study of ‘pictorial conventions for the recounting of thoughts and experiences'.16 Given the ubiquity of the device in medieval art, perhaps what we really need is a thorough and comprehensive investigation into the ‘archaeology of the cloud-frill’.