Scriptural Text Divisions in Their Medieval Context: an Overview

As historians such as Beryl Smalley and Gilbert Dahan gleaned from their analyses of medieval scriptural commentaries, the community of theologians that coalesced around the establishment of the universities and that flourished within them at the end of the twelfth and beginning of the thirteenth centuries developed a particular style of commentary dependent upon distinctive methods.Footnote ² Hierarchical division was a general method according to which medieval scholastics organized all kinds of materials for scientific analysis. Working scientifically meant defining and dividing one's objects by their genus and difference, that is, by the group to which they belonged and by the property differentiating them from others within that group. All students trained in dialectics acquired, therefore, the mental habit of dividing and subdividing. Furthermore, they all became acquainted with a vertical tree figure representing such a hierarchy: a centuries-old diagram that since the thirteenth century was called “the tree of Porphyry.”Footnote ³ A highly intense interest in hierarchy was evident in both natural philosophy and theology. It was therefore as natural as it was unique to subject philosophical and religious texts alike to this same all-encompassing preoccupation with hierarchies that swept the thirteenth century.

While earlier commentators occasionally provided a principal division into the parts of the texts they were explicating, the scholastic hierarchical, progressive division into smaller units was a radically new and different practice.Footnote ⁴ John F. Boyle's description of the divisio offers us a useful point of departure:

Text divisions thus became almost obligatory in commentaries on Aristotelian texts, on Peter Lombard's Sentences (which functioned as the newly introduced textbook of the faculty of theology), and in commentaries on the Bible itself.

This sort of progressive analysis is remarkably similar to modern approaches to narrative like that suggested by Roland Barthes that were inspired by linguistic analyses of the single sentence.Footnote ⁶ Did medieval grammar similarly inspire textual analysis? From what we know of Latin curricula and the speculative grammar of the twelfth and thirteenth centuries, the answer is likely negative. Latin syntactical theory seems to have other emphases.Footnote ⁷ A more likely source, yet only to a limited degree, can be located in classical rhetoric. Rhetoric was widely applied by thirteenth-century theologians.Footnote ⁸ Examples include Thomas Chobham, who enriched his Summa on the Art of Preaching with insights and terminology drawn from rhetorical theory, and William of Auvergne, who did the same in his Rhetorica divina on prayer.Footnote ⁹ The divisio textus itself also assumes an ability, developed in the context of rhetorical studies, for recognizing and naming the functions of different parts of the text. Terms such as narratio, redarguitio, and perhaps commendatio as well, are present, as we shall see below, in our diagrams, though perhaps less frequently than one would expect. Classical and medieval rhetorical theory emphasizes the general importance of organizing the orator's materials. The Rhetorica ad Herrenium even recommends the orator include a “division” as part of his speech (the third out of six) by means of which “we make clear what matters are agreed upon and what are contested, and announce what points we intend to take up.”Footnote ¹⁰ It also recommends the use of enumeration, “when we tell by number how many points we are going to discuss.”Footnote ¹¹ The medieval commentator who makes a divisio textus may be viewed, therefore, as taking the position of the author, enumerating the parts of the primary text to be discussed. However, these divisions and enumerations are still only acts of parsing and naming. Ancient rhetorical theory did not encourage subdivision or structural analysis of the relations between these parts, something that lies at the heart of the divisio. Its method of grouping and regrouping resulting in an overall complex structure is uniquely medieval.

The question of textual structure is equally relevant to composition and analysis. And, indeed, the systematic theological works authored since the early years of the thirteenth century present in explicit fashion a similar structure to that produced through divisio textus.Footnote ¹² Theologians and/or their disciples compiled voluminous summae with similar multileveled structures out of small units of debates over specific issues. They grouped and edited these issues into larger quaestiones, then into treatises and books. As they were engaged with these constructions, and as they wore hats of both commentators and authors, it seems likely that high-scholastic theologians believed ancient authors shared their own conceptions regarding the essential relation between doctrine and structure and that reconstructing the underlying structure of ancient texts would lead them to a deeper understanding of their meaning.Footnote ¹³

Scholars of medieval biblical exegesis who have addressed the particularities of the university mode of commentary have devoted scant consideration to divisio textus, merely noting that it expresses subtle exegetical and doctrinal views. Many of its most promising aspects remain, however, understudied. This may be attributed to the responses of modern readers, who may feel, as Sten Ebbesen expresses it, that “the thirteenth century divisions of the text carry the art of boredom to its perfection.”Footnote ¹⁴ Part of what makes these divisions difficult to swallow lies in their being purely verbal. Most scholars, especially those using modern editions, encounter text divisions in their verbal form, which often follows a pattern such as “the book has three parts … the first part divides into two … the first of these sections divides into three … , etc.” written in the same successive lines in which the rest of the commentary is written. Verbal descriptions of complex structures are destined almost by their nature to be tedious and difficult to process cognitively. Yet university masters also represented such divisiones visually, by way of tree diagrams, a fact that remained largely unknown until now. These visual representations can facilitate deeper study and appreciation of the analytic potential of divisiones and shed new light on how medieval Schoolmen understood textuality.

Various examples of divisio diagrams can be found in the margins of manuscripts of Aristotle's works and of biblical commentaries, as well as of commentaries on Peter Lombard's Sentences.Footnote ¹⁵ To the latter belongs one of the earliest and richest repositories of divisio diagrams, found in many manuscripts of Richard Fishacre's commentary, where they are called “ramified trees” (arbores ramificatae). Richard's arbores — indeed the entire genre of divisiones sententiarum — deserves a study of its own following the art of theological authorship and argumentation.Footnote ¹⁶ My current interest, however, is to investigate what happens when theologians applied this structural sensitivity together with its diagrammatic expression to stories. In doing so, I argue, Schoolmen engaged in an implicit narratology, the art of analyzing narrative structure, plotlines, and dialogues.

A general description will not suffice to substantiate this claim comprehensively or demonstrate this technique and the effectiveness of its visual representations, but diving into its rich mode of analysis will provide introductory familiarity. The following discussion thus addresses the biblical divisio and the narrative perception it conveys through a close inspection of an unedited series of such diagrams that I discovered in the lower margins of a late-thirteenth-century manuscript from the Bibliotheca Communale in Assisi. The series in question depicts the structure of the first three chapters of the Book of Job. I will closely examine the way its author analyzes narrative features such as events and dialogues compared with choices of his contemporaries. As noted above, purely verbal analysis of spatial structures is liable to obscure more than to elucidate. Accordingly, full transcriptions and translations of the diagrams — to the best of my knowledge the first of their kind to be edited — are appended.Footnote ¹⁷ Readers are heartily encouraged to consult them to this article frequently so as to experience the effect of visual representation firsthand. I have also included the biblical text, parsed and numbered to facilitate reading.

Delineating the Structure of Job: Anonymous Diagrams in Assisi Bib. Com. 51

The Schoolmen of the thirteenth century composed a significant number of commentaries on the Book of Job. Among them were the Dominican masters Roland of Cremona, Hugh of St. Cher, Guerric of St. Quentin, Thomas Aquinas, and Albert the Great, as well as the Franciscans William of Middleton, Peter of John Olivi, Matthew of Aquasparta, and Richard of Mediavilla.Footnote ¹⁸ The most famous of these commentaries is Thomas Aquinas's Literal Exposition.Footnote ¹⁹ “One of Aquinas’ more mature and polished commentaries,”Footnote ²⁰ it presents a thorough discussion of providence and related matters and exemplifies the scholastic exegetes’ new emphasis on the literal sense of scripture.Footnote ²¹ Although Thomas employed divisio textus in other commentaries, he did not propose an explicit one in this exposition. Boyle has suggested that since Thomas perceived the book of Job as “a give and take narrative argument” the literal interpretation of the text was not served by such a division.Footnote ²² Jaffe, in the introductory essay to his English translation, insisted that Thomas delineated a very clear structure, although “Thomas’ reader is not made aware of this structure in advance.” He proceeded therefore to fill this gap and provided one of his own.Footnote ²³

At least two late-thirteenth-century manuscripts of Aquinas's exposition also compensated for this lacuna. On the last page of the first, MS Florence, Biblioteca Medicea Laurenziana Plut. 20.18, dated to before 1280, someone added an unfinished draft of a verbal divisio textus.Footnote ²⁴ The second manuscript is MS 51 of the ancient Franciscan Bibliotheca communale in Assisi, dated to the late thirteenth century.Footnote ²⁵ This codex contains one of the earliest copies of Thomas's commentary, biblical commentaries and theological questions by Matthew of Aquasparta (fols. 121–202 are said to be his autograph of the commentary on the Apocalypse), commentaries by Jean of La Rochelle, as well as an anonymous, decapitated commentary on Job.Footnote ²⁶ In the bottom margins of the first five folios of Thomas's exposition an anonymous hand has carefully drawn ten diagrams, depicting an unknown divisio of the first three chapters of the Book of Job, to which we shall now turn.

The diagrams in question branch horizontally to the right. Each unit or cluster of units of the biblical surface text is assigned a name or “title” and, usually, indicated by its opening words. In diagrams of this sort, the latter are usually underlined, but in these specific diagrams they may not appear at all, or they may appear above the title with no underlining. The units’ titles can be a noun in the nominative case; a partial sentence to be completed with each of the alternative titles of the units of the next level; or the completion of a partial sentence begun in the preceding level, always declined in the appropriate case. One can thus decode a syntactically correct, or approximately correct, sentence moving through the tree. The movement rightward, upward, and then downward in Figure 1, for instance, introduces the speeches of Job's friends in something approaching a sentence as follows: “Hic determinatur de temptationibus ipsius Iob a domino factis per / medium triplex, ab mundo, id est amicis” (This part is about the temptations of Job brought about by the Lord through / three intermediaries, [the third being] by the world, that is, his friends).

Figure 1: Partial transcription of diagram 2r.

All in all (excluding the two diagrams in fol. 5, which analyze Job's speech and therefore do not belong to the narrative), eight diagrams describe the complete structure of the first two chapters, presenting the relations among a total of eighty leaves, a leaf being the smallest undivided textual portion. Combined in a large folio, these diagrams together would have made up one tree. Diagram 1r shows the primary division of the entire book. In its first two divisions, the bottom branches remain undeveloped, while the upper branch is explicated to its end. Diagram 2r departs from that second promised branch, and the diagrams in 3r, 4r, 4v^a and 4v^b develop the branches that remained open in 2r. The two last bottom branches of diagram 3r continue in 3v, while 4v^c completes 4v^b, whose open branch itself continues in 5r^a and 5r^b. Out of forty-five divisions and subdivisions of sections of all levels, the vast majority divide into two (53%) or three (29%) branches. Divisions into more than three parts (18%) are usually at the leaf level. This tendency to generalize and regroup a minimal number of units results in a high number of levels, enhancing the impression of the text as a highly complex entity.

The principles of parsing vary according to the matter. Some have a more formal nature, some are closer to the specific content. A pure formal division, for instance, appears in diagram 5r^a, the first division being between the “title of the coming narratio” and “the narratio.” Other principles of division follow common dichotomies like good vs. bad, a general group such as “goods” and its species (temporal vs. spiritual), or emotional phenomena such as mourning or sorrow and their various expressions. The sentences describing the death of Job's children are distributed into the elements composing an event: time, place, agent, etc. Significantly, parsing pays no regard to the actual surface length of the units or to formal division into chapters or verses. The first division treats almost the entire book in its first part, leaving a handful of verses to the second. The number of levels differs as well: the shortest unit is at five-level depth (diagram 1r, units 1–3), and the longest at fifteen-level depth (diagram 3v, units 52–55). Subdivisions may result in equal-length branches, mainly towards the end of the division, or in highly asymmetrical ones, such as a first, short branch then two much longer ones, or other combinations. While thematical symmetry is strengthened time and again by the frequent division into two, quantity and length were clearly not an issue for the art of narrative construction in medieval eyes.Footnote ²⁷

This set embeds numerous theological and structural choices. I cannot dwell here on each, but will focus on several points revealed in the first two diagrams. The first, showing the principal division of the book and the full explication of the first verses, will serve to demonstrate the subtlety and interpretative force of this technique and its diverse expressions with regard to coherence and the narrative role of a character's description; the second will examine the way dialogues and argumentative exposition are tackled in the frame of narrative analysis.

The narrative role of character description

Looking more closely at diagram 1r and Figure 2, we see that, unlike the ideal type described above, the surface — the opening words of the specific textual unit — is frequently missing. In three of its five occurrences, it is written above the title without the typical underlining. The diagram begins with a twofold division of the story: the multitude of Job's sufferings, which occupy almost the entire book, and the symmetrical plurality of his comforts as described in its final chapter. This division is identical to the one made by Matthew of Aquasparta, whose commentary reveals its source in Psalms: “secundum multitudinem dolorum meorum in corde meo consolationes tuae laetificaverunt animam meam” (Ps. 94:19) (in the multitude of my anxieties within me, Your comforts delight my soul). While the author of this diagram did not specify the precise words opening the second part, Matthew sets its beginning at 42:1.Footnote ²⁸

Figure 2: Assisi, Bib. Com. 51, fol. 1r. Foto di Giorgia Menghinella, Assisi – 2017 © Archivio fotografico del Sacro Convento di S. Francesco in Assisi, Italia.

The first verses present the protagonist through his name, origin, virtues, and possessions. What is the function of this characterization in the overall structure? And why should prosperity be under the section of the story allegedly focusing on Job's sufferings or anxieties? Olivi, for example, chose to accord treatment of Job's prosperity an independent place, dividing the story into three parts — initial prosperity followed by adversity and finally good fortune — thus suggesting a ternary form and emphasizing the chiastic relation of its opening and ending (ABA form).Footnote ²⁹ Another contemporary Franciscan master, William of Middleton, understood all that precedes the devil's attack to be a description of Job's person as the person fit for the battle he will be subjected to presently, thus undermining the prosperity-adversity opposition altogether.Footnote ³⁰ Matthew of Aquasparta, who similarly to our anonymous author splits the “sufferings” section into “prosperity” and “adversity,” explains his division by slightly changing the section's general theme. The section, he argues, aims to demonstrate Job's personal perfection, which is revealed in times of both prosperity and misfortune. A similar principle guided Richard of Mediavilla. He also divided the book into two parts: persecution or perfection, and retribution. The first part presents not only Job's persecution, but also his perfection, which is demonstrated in times of prosperity and anxiety alike.Footnote ³¹ Our author, however, conceived the narrative role of the description of Job's prosperity differently: it forms a background or setting for his suffering, for, in order to understand Job's afflictions, one must first know his preceding prosperity.

This raises the question of how Job's name and place of origin participate in prosperity or perfection. Thomas Aquinas, whose exposition appears above our diagram, could be understood in two ways. At first, he argues that the description of prosperity begins only after the person of Job and his virtue are presented. The author of the division attached to Thomas's exposition in MS Florence does precisely that (see appendices). Following Thomas, prosperity begins only after “person” and “virtue,” with the words “natique sunt ei” (and [seven sons and three daughters] were born to him). But Thomas also writes that the author of Job intended to depict him in the most perfect way, so as not to make us suspect that his adversities were due to anything he had done. His being a vir, a man rather than a woman, a mature individual but not yet an old man, accords with this perfection.Footnote ³² In the same manner, he interprets Job's virtues as spiritual goods, and the following sentences as describing the temporal goods.Footnote ³³

The author of the Assisi diagram, however, found a middle way. He chose to see Job's name and origin as part of the “prosperity” section, but not a sheer expression of it, or as a kind of good. The name of the second branch of prosperity is “prosperity” as well, suggesting that he conceived the details in the first branch as a preliminary to the issue proper.Footnote ³⁴ Virtues, however, were interpreted as a genuine part of the spiritual goods bestowed upon Job, as he divides this second category of “prosperity” into spiritual goods, temporal goods, and a third category, the script of which was unfortunately damaged.Footnote ³⁵ The spiritual goods — that is, Job's virtues — are then further divided according to their relations to the good and the bad and then into even more specific subcategories emphasizing symmetrical relations to the smallest detail. Divisio textus is this wrestling with how to understand the true place of an author's description of his protagonist, the ways it serves the author's message, and the unfolding of the story, while considering different options suggested by other masters regarding the exact place of the first words within the fabric of the entire tale. This task required remarkable effort and subtlety and was facilitated by using diagrams, as I hope readers experience themselves as they struggle through my description. Readers are welcome to look further for the diverse ways by which even the section describing Job's possessions was structured differently by different commentators in the appendices.

Arguments, imbalance, dialogues, and repeated patterns

Diagram 2r develops the theme of adversity but then relegates it to the background, taking God's temptation of or attack on Job as this section's organizing principle. This choice in itself embodies interpretative force by refuting any possible designation of Satan as the initiator of the entire episode. All that follows is the story of God's temptation, which, the diagram tells us, is executed either through intermediaries — Satan, Job's wife, his friends — or directly, when God speaks to Job from out of the storm. These diverse events thus acquire a kind of equivalence, sharing the same role, although they differ in terms of narrative length. Job and his friends’ disputation occupies more than thirty chapters, while Job's wife's temptation occupies less than two lines.

This quantitative asymmetry, as with that of the very first division in diagram 1r, stems partly from the book's peculiar combination of enveloping frame-story and lengthy theological discussion and might easily be solved if the discussion were entirely separated from the frame story. This is precisely what Thomas did, distinguishing as the very first step of his commentary the disputatio from the ystoria. The anonymous author of the divisio in the Florence codex followed him (see appendix E). Albert the Great set discussion apart as well. Taking “temptation” as the organizing principle for the entire book, he divided the book into three sections: 1) the attack or temptation itself, including the preceding state of prosperity; 2) the disputatio between Job and his friends regarding the cause of temptation; and 3) Job's state after his temptation.Footnote ³⁶ Our author, however, takes the lengthy disputation to be a genuine part of God's indirect temptation and attack, one way among others to challenge Job's perfection.Footnote ³⁷

Here, as in other cases, medieval commentators distinguish between the occasion, condition, or setting for a principal event, on one hand, and the event itself, on the other, strengthening coherence within units where it is not always evident but also exacerbating imbalance. The function of some actions is to launch a response or event, which forms the true center of the episode. The analysis of Satan's attack imposes a series of such asymmetrical structures on the text, each of which has a clear center of gravity. The author first distinguishes the setting (Satan's request for permission to attack Job) from the actual attack (2r, 4th division); then, on a minor scale, he distinguishes between two preparative stages and the “attack itself” (diagram 3r, 1st division), then distinguishes another “attack itself” section, presenting the killing of Job's children. Narration of events, according to this anonymous commentator, is partly constructed of pairs in different levels of a preceding setting and the main, “actual” event it launches.Footnote ³⁸

Narrators often weave dialogues into their fabric. But the parsing and analysis of dialogical scenes can be challenging. One might focus on the actual stops interlocutors make while exchanging words, or one might address more general features of the conversation's dynamic. For instance, should “yes” and “I totally agree,” when split by a remark of the second interlocutor, be taken as two units or as one expression of consent? The author of diagram 2r observes that the scene revolves around one principle act, obtaining permission.Footnote ³⁹ Therefore, for the sake of analysis, Satan's submission of his request precedes God's granting it. This deep structure defies the chronology of both the surface level of the text and its story, where God appears earlier. Each character is treated with four units, distinguished by their function in the dialogical act rather than by the natural division created by the verbal exchange. The titles of each section draw the reader's attention to a hidden symmetry in the setting of the scene:

Having shuffled the chronological sequence in this manner, our author also provides the readers with a key to reconstructing its original order. The number of dots at the end of each line (from one to eight) signifies their order of appearance in the biblical surface text: one dot for the first section, two dots for the following, etc. (see Figure 3).

Figure 3: Assisi, Bib. Com. 51, fol. 2r. Foto di Giorgia Menghinella, Assisi – 2017 © Archivio fotografico del Sacro Convento di S. Francesco in Assisi, Italia.

The uniqueness of this choice becomes apparent when compared with Matthew of Aquasparta's very different approach to the same dialogical challenge. Matthew keeps the original order and coherence by focusing on one interlocutor and his personal drama: Satan. The heavenly scene precedes the diabolic attack by presenting Satan's hostility, which will be later executed. This presentation of hostility is divided into its “occasio” (occasion or trigger) and its “manifestation.” God asks Satan from where he has come and after receiving Satan's answer teases him by asking whether he saw his servant Job. God thus creates in several steps the context for Satan to manifest his hostility. This manifestation of diabolic hostility, according to Matthew, is the essence of the second part of the scene, which consists of Satan's reply, his request for permission to afflict Job, his reception of divine authorization to proceed, and his departure to his task.Footnote ⁴⁰

This last section reveals another structural challenge. Satan's withdrawal from the scene, can be conceived in two ways: as the end of this narrative unit or as a short opening for the next narrative unit. Matthew chose the first option, our author the second. In diagram 3r, Satan's departure and the designation of the period of time for his actions form two short units preceding the description of his attack, which occupies the largest part of the diagram. As in its other parts, structural analysis of the series of misfortunes endured by Job enables the reader to recognize both similarities and differences between accounts, just as it did in the graphical parsing of the dialogue between Satan and God. Although the trimming of the outer margin of the page has cut off the script, one can see that “vicinity” and “multiplicity,” for instance, appear in the description of the loss of Job's oxen and she-asses, as well as his camels, but in a different order. Job's response and its praise close this final unit: a sensitive choice, for Job's mourning could well be interpreted as a response to the entirety of his losses, including his property, rather than for his children alone. Diagram 3v carefully analyzes his response in detail: three signs of his mourning; his approach to God in sign, deed, and word, etc.

Many classic stories repeat almost identically patterned scenes with slight variation: the youngest pig meets the wolf, then the older brother, then the oldest; Goldilocks finds a big/mid-sized/small chair/cup/bed, etc. In our story, Satan attacks twice, and the biblical author includes two almost identical celestial scenes before each attack. In both, a similarly phrased dialogue between God and Satan precedes the actions themselves: the first against Job's possessions and family, the second against his own body. The third branch of diagram 2r presents the division between the attacks and the dialogue of the first scene, but it is not duplicated for the second, identical, scene. The author just notes in 4r that the division is exactly the same. On the other hand, in 4v, there are diagrams that develop the two other indirect divine temptations — through Job's wife and through his friends — and painfully illustrate their structural difference. The brief wife's scene is not introduced with any remark, such as “his wife was there,” while the temptation presented by his friends is preceded by an extensive preparatory description of their arrival, their expressions of sorrow, etc.

Conceiving Texts as Trees: The Meaning of the Diagrammatic Representation of Stories’ Divisio

The divisio diagrams belong to a large family of horizontal tree diagrams with which theologians, philosophers, and jurists represented different kinds of relationships and that occupied the margins and sometimes entire folios of medieval manuscripts from the late twelfth century at the earliest through the fourteenth and fifteenth centuries. Many of these were distinctiones, presenting conveniently organized lists of several spiritual meanings pertaining to an object or of the reasons why a certain metaphor is used. Others visualized the generation of Aristotelian syllogisms, while still others presented diverse kinds of lists.Footnote ⁴¹ All these diagrams, as I explain in detail elsewhere, are built as sentences with alternative parts.Footnote ⁴² This principle dictates their horizontal form and sharply distinguishes them from other trees, mostly vertical ones, such as divisions of sciences or the Porphyrian tree and appeared in the medieval diagrammatic tradition prior to the thirteenth century.Footnote ⁴³

In Richard Fishacre's diagrammatic divisiones of Peter Lombard's Sentences, we find a unique and explicit comment on the advantage of using a tree diagram to represent the divisio: it allows one to distinguish each section (singula) clearly.Footnote ⁴⁴ But the tree does much more than that, clarifying the mutual relations of these single parts and enabling a flexibility of perspective. It conveniently shows an overall structure and enables diverse “zoom-in” views, which are not dictated as in a verbal divisio where the author leads the reader by a single linear route.Footnote ⁴⁵ Clearly exhibiting the parts and their relations, it facilitates reorganization. One can easily move a line, draw sketches, decide to group or to separate items, compare divisions, and in doing so modify interpretations and ideas.

Richard and others visualized their analyses of the Sentences; others visualized analyses of logical treatises. But the application of diagrams to narrative enhances the implication of the spatialization of time. Diagrams essentially illustrate the implicit spatial nature of narrative. Indeed, already the fundamental visual representation of a story in its written form — the page and the book — turns time into space.Footnote ⁴⁶ Movement in time along this space is partially free. One's eye may browse back several pages to recall a forgotten piece of information, meditate on a particular sentence for a while, or take a glance at the end. At the same time it is limited. The graphic forest of script signs in successive lines makes it difficult to see quickly where to browse, thus favoring the dictated linear progression. It was in the twelfth and increasingly throughout the thirteenth centuries that different aids familiar to us today were invented to facilitate location of specific information in texts without reading all the way through, helping readers to navigate the textual space of page and codex. More and more texts and manuscripts were manufactured “for use” rather than for reading, and their layout was designed accordingly.Footnote ⁴⁷ Even manuscripts of chronicles, which constituted the genre of linear time, were produced with indices and original visual subject signs, enabling one to create different paths of reading through a single text, depending on whether one seeks information on bishops, for instance, or the history of secular emperors, or any other subject the narrative records.Footnote ⁴⁸

The diagrams analyzed here both build on and complicate this spatialization of time and the readerly freedom it entails. Observers who lay an eye on a diagram first see one object, a story that exists in a certain manner at a single moment. It is noteworthy that terms such as opening, beginning, or ending are entirely absent. Yet, as in a sculptural plan of a giant church portal or a complex picture, observers freely move in time from one detail to another. Indeed, two cases excluded, were we in possession of the full text of each unit and not just its opening words we could read the entire narrative top down. But the diagram's heart is the underlying structure and not the surface, which is frequently missing or improperly underlined. In fact, the diagram presupposes the mental or physical existence of the surface at hand so that one can now play with its spatial settings.

The divisio — both verbal and visual — prompts movement back and forth across the original narrative sequence in a series of leaps. The reader is pulled forward and backward in ever shorter intervals, from the opening chapter to the last, then back to the beginning again, then to the beginning of the second part of the first part, etc. While the diagram is based on the sentence and therefore is read mainly from left to right, as the unfolding of an idea, it enables readers to work their way back in the opposite direction, placing each detail in widening circles of context and meaning. They may also read vertically, creating thereby abstracts of the text that differ in their increasing resolution. Some diagrams afford more freedom, while others dictate navigation more rigidly. For instance, the reader who follows 2r is limited to a specific route, though she passes through junctures that post the names of other narrative avenues.

The assumption that the underlying structure is a tree seems, however, to support linearity from a different aspect. As adjacent units are grouped into larger ones, there is seemingly little room for links between non-sequential parts, which skip over their neighboring units. In the sequence that runs 1, 2, 3, 4, 5, 6, for instance, strong ties and common features are usually found between successive units such as 2–3 or 4–5–6, while a group holding just 2 and 5 is rare. There therefore are few opportunities to interpret something that appears in an earlier scene as a clue to be solved further on in the story or to a narrative that interlaces plotlines. Although such sequentiality is the typical case, scholastic commentators conceived less common divisiones whose branches interlace as well. One example is found in the anonymous divisio in the above-mentioned MS Florence Medic. Laur. Plut. 20.18, which closely reflects Aquinas's implicit divisions and is edited in both verbal and reconstructed diagrammatic form in Appendix E. Following Aquinas, it demonstrates Job's virtue, as well as his care for his and his family's moral purity in two non-sequential sections of the text: his description (4–7) and his later pious conduct (15–20).Footnote ⁴⁹ Material prosperity (8–14), the third branch, lies between these two sections. In fact, Thomas's first division should result in a significantly non-sequential structure as well. The above-mentioned distinction between the historical-descriptive part of the book and the disputation over divine providence implies that the first part consists of the first three chapters and the last one, thus enveloping or framing the disputation.

I will round out this survey with consideration of the possible pedagogical functions and uses of these diagrams. There is no direct evidence that such diagrams were used in the classroom, but it is plausible that medieval teachers employed this new device, since it reflected the general understanding of their milieu regarding the best way to comprehend texts. Medieval diagrams have also often been interpreted as mnemonic devices, especially since Mary Carruthers's highly influential studies.Footnote ⁵⁰ Classification and subordination, as well as graphical representation, were recommended by medieval authors as aids for memorization. Divisio diagrams could have been used, therefore, in order to memorize either the commentary or the biblical text itself. If so, this memorization involved a great deal of additional interpretation, very different from that required by images such as the figures contained in the anonymous Ars memorandi for the Gospels.Footnote ⁵¹ There is yet, however, no direct evidence for this function.

Considering the remarkable increase in search and reference tools during the twelfth and thirteenth centuries, which have been well documented by Malcolm Parkes and Mary and Richard Rouse, one might argue that both the divisio and its tree diagram should be explained in this context.Footnote ⁵² Indeed, the thirteenth century saw the division of the Bible into chapters, sections, and, later, verses. Although vertical lists of contents and indices were already in use earlier, they certainly won unprecedented popularity at this time. Nevertheless, it is the very duplication of means that clarifies the great difference between these divisions and lists, on one hand, and the divisio and its diagrams, on the other hand. Indexing and numbering chapters and verses facilitates uniform naming, meaningless in itself, yet which allows one to locate and refer to a textual portion precisely without relating to its context. Similarly, a list of contents allows one to jump to the right page and ignore the rest. A vertical list of items does not represent internal relations or structure, or, if it does, it is only the simplest of possible structures. The divisio, however, as shown above, can ignore the formal division into chapters and verses altogether, even if it occasionally refers to them for its purposes.

Conclusion: Towards a Medieval Narratology

The divisio textus was an intriguing, uniquely medieval practice and micro-genre, which finds parallels only in strikingly similar modern analyses in literature, cognitive science, and computational linguistics. Embedded in what Minnis treated as the implicit medieval theory of authorship, discourse analysis did not form a specific branch of knowledge in the medieval world, whether as part of grammar or of rhetoric or any other discipline. Nor did medieval scholars produce a general theory. Although implicit, the study of the divisio textus and its diagrammatic expression demonstrates how subtle, sophisticated, and varied their perceptions of textual coherence and narrative structure were. The series of diagrams depicting the structure of the narrative opening the book of Job represents an illuminating example of this genre and of the narratological insights it conveys. It reveals to us the way the Schoolmen understood texts as carefully constructed and layered arrays of semantic elements, as well as the mechanisms they discovered that constitute the art of narration.

Comparison with other divisions of the same text has demonstrated the range of perceptions regarding the function of a character's description in relation to the main message, as well as the challenges that argumentative sections and dialogues pose to structural thought. Further studies into divisiones may be significantly enhanced by graphical reconstructions and may throw additional light on the practice. Such work may trace its historical development from its rise to its decline, identify tendencies of both individuals and of schools, inquire into resemblances to contemporary literary techniques, and contribute to the vast field of sermon studies and studies of medieval prose. All these, I hope, will further enhance our understanding of how medieval scholars perceived authors and texts and how visualization served them in discovering and expressing these understandings.