Show Summary Details

Page of

Printed from Oxford Research Encyclopedias, Literature. Under the terms of the licence agreement, an individual user may print out a single article for personal use (for details see Privacy Policy and Legal Notice).

date: 06 July 2022

The Chapterfree

The Chapterfree

  • Nicholas DamesNicholas DamesDepartment of English and Comparative Literature, Columbia University


First known as a kephalaion in Greek, capitulum or caput in Latin, the chapter arose in antiquity as a finding device within long, often heterogenous prose texts, prior even to the advent of the codex. By the 4th century ce, it was no longer unusual for texts to be composed in capitula; but it is with the advent of the fictional prose narratives we call the novel that the chapter, both ubiquitous and innocuous, developed into a compositional practice with a distinct way of thinking about biographical time. A technique of discontinuous reading or “consultative access” which finds a home in a form for continuous, immersive reading, the chapter is a case study in adaptive reuse and slow change. One of the primary ways the chapter became a narrative form rather than just an editorial practice is through the long history of the chaptering of the Bible, particularly the various systems for chaptering the New Testament, which culminated in the early 13th century formation of the biblical chaptering system still in use across the West. Biblical chapters formed a template for how to segment ongoing plots or actions which was taken up by writers, printers, and editors from the late medieval period onward; pivotal examples include William Caxton’s chaptering of Thomas Malory’s Morte d’Arthur in his 1485 printing of the text, or the several mises en proses of Chrétien de Troyes’s poems carried out in the Burgundian court circle of the 15th century. By the 18th century, a vibrant set of discussions, controversies, and experiments with chapters were characteristic of the novel form, which increasingly used chapter titles and chapter breaks to meditate upon how different temporal units understand human agency in different ways. With the eventual dominance of the novel in 19th-century literary culture, the chapter had been honed into a way of thinking about the segmented nature of biographical memory, as well as the temporal frames—the day, the year, the episode or epoch—in which that segmenting occurs; chapters in this period were of an increasingly standard size, although still lacking any formal rules or definition. Modernist prose narratives often played with the chapter form, expanding it or drastically shortening it, but these experiments usually tended to reaffirm the unit of the chapter as a significant measure by which we make sense of human experience.


  • British and Irish Literatures
  • Middle Ages and Renaissance (500-1600)
  • 19th Century (1800-1900)
  • 20th and 21st Century (1900-present)
  • Literary Theory
  • Western European Literatures
  • Print Culture and Digital Humanities

Few bibliographical terms, and few textual devices, have had the psychological and cultural impact of the word “chapter.” It is a common way in which the technology of the book has entered everyday speech and everyday thinking about our selves, our lives, and our pasts. To “close that chapter of my life” with regret; to finish a repetitive story by saying “and so on to the end of the chapter”; to “start a new chapter,” either with relief or enthusiasm: these are at once experiences of reading and experiences of living. Perhaps only the page—to “turn the page” or “be on the same page”—has had the same ability to move from the vocabulary of the book to biographical metaphor. These chapter-metaphors signal a way in which our lives, in fact, take on the shape of a book. More specifically, a certain kind of book: the novel, the genre in which the technology of the “chapter” has taken root most firmly, and in which the chapter develops from a textual device to a way of constructing the world. The chapter seems in fact to be an inescapable part of the machinery of the novel, a necessary joint within its larger architecture without which its usual bulk could scarcely be supported.

Yet although these demotic ways of using the term “chapter” refer to breaks in an overarching narrative—a division in a story—the origin of the chapter as a term and as a textual unit lie not in narrative at all, but in fact in non-narrative, informational, quasi-encyclopedic texts in classical antiquity, particularly in the compendia of Roman legal, grammatical, and scientific knowledge. To understand the chapter’s specifically literary meaning, it is necessary first to excavate its history prior to the novel, a history that extends back almost two millennia before the popularization of the term “novel” in European languages, and one that centers on a phenomenon of editorial and bibliographical practice in an environment prior to the codex itself. To do so raises the question: How does such a transition happen—a transition between a material practice of arranging texts, a habit of mise en page, and a formal practice of literary composition, something capable of meaning and as such readily available to metaphorization?

Two considerations immediately reveal themselves. The first is that such a material practice would have to be ubiquitous, spanning divisions of language or genre. The second is that it would also have to be innocuous and conventional—something so ordinary, so uncontroversial, so stubbornly persistent, that its initial justifications and contexts could easily be forgotten, reimagined, reinvented; it would be a tool, capable of instrumentalization but, at least not initially, of significance, nor worthy of particular reverence. It would, in short, accumulate instances but not rules. Certainly both are the case with the chapter, that inescapable mode of segmenting long prose narratives. What follows is a capsule history of the ubiquity and conventionality of the chapter. That ubiquity has been and remains capable of innumerable uses and styles; no two authorial manners of chapter structure are ever exactly similar. But the history of the chapter’s gradual transition from an editorial to a compositional unit nonetheless reveals certain patterns and recurrences that speak to deep tendencies in what might seem like a bewildering, potentially endless set of examples.

The Chapter in Antiquity

What is, and was, a chapter? The word itself is only a partial help. “Chapter”—caput, capita or capitulum in Latin, kephalaia in Greek—derives in each case from “head”; thus “heading” is a fair enough translation of these early terms. But “heading” is an ambiguous term in classical antiquity’s discourses on textuality. A caput or capitulum can at once identify a brief summary of a unit of text (what we ordinarily think of by “heading”) and the unit of text itself (what we now call a “chapter”); frustratingly, it can also refer to any passage of text at all, not necessarily even a passage textually delineated as such. A cluster of subsidiary terms, such as titulos (or “title”) and breves (brief headings) display a similar tendency to migrate between standing for the label of a textual unit and the unit itself. But this confusion conceals an important truth. The “chapter” is, from its origins, never a freestanding textual whole; it is always a division within some larger entity: a hiatus, a caesura, a segment, or what classical rhetoricians denoted a pericope. If it is a unit, it is a unit always conscious of its place within a sequence of similar units, bound within a larger unity: its essence is its segmented, incomplete status. This is not so with the other major textual unit-term of classical antiquity, the book (biblos, tomos, or liber). Our earliest extant descriptions of the “book” as a unit imply a wholeness the chapter never suggests. The 1st-century bce historian Diodorus Siculus, in his Bibliotheca historica, praised the model of the earlier historian Ephorus, to whom Diodorus credited the invention of writing in carefully shaped biblon: as Diodorus described it, “each one of his Books is so constructed as to embrace events which fall under a single topic. Consequently we have given our preference to this method of handling our material and, in so far as it is possible, are adhering to this general principle.”1

The unity of a “book” is premised, therefore, on the unity of its material form, the papyrus scroll or volumen. Diodorus here comments on the appropriateness of a composition that arranges a conceptual unity (a “single topic,” a genos) to match a physical unity. This was a difficult art, since the length of the papyrus scroll was more or less fixed; as historians of ancient books have calculated it, scrolls were usually thirty to thirty-five feet long, which with average column sizes and bookhands—themselves relatively stable across centuries—would yield something like 20,000 words at most. This rather arbitrary measure, which combined factors of cost with factors of size—the scroll was a carefully calibrated compromise between a certain amplitude and a certain ease of use, particularly in the matters of holding and unrolling—became itself a compositional measure; it is no accident, for instance, that the Gospel of Matthew is 18,000 words long. Matching conceptual unity to a fixed material length became an acknowledged art of classical authorship; as Henri Marrou has described it, even as late as Augustine the “book” was considered, at its best, a “literary unity.”2

The chapter, on the other hand, is not isomorphic with its physical referent. In fact, it has no physical referent; unlike the “book,” it does not mimic other documentary or material forms (the letter, the testimony or affidavit, the scroll); it is always a subset rather than a set. Tellingly, chapters have never even tended toward any numerological tradition; whereas numbers like the twelve or twenty-four “books” of epic poetry became standard, and Greek and Latin historical works were organized in “decades” or units of ten that could involve their own higher orders of unity, chapters remained stochastic, even random. How many, for instance, know offhand the total number of chapters in Tom Jones, War and Peace, Les Misérables, Middlemarch, or Buddenbrooks?3 This is not the case in Chinese fiction, where the 100- or 120-chapter novel defined distinct forms, but in the West chapter numeration rarely if ever reflected significant schemes.4 Put simply, the chapter, from its origins, expresses fragmentation rather than unity.

That is because it was first understood as a navigational aid for what we now call consultative reading: the kinds of scanning, information-seeking perusal that seemed increasingly called for by the information-dense texts of mid to late classical antiquity. The first texts we possess that seem to have been chaptered from the outset, rather than chaptered by subsequent editors, are telling: Cato the Elder’s De agri cultura (mid-2nd century bce), Pliny the Elder’s Naturalis historia (1st century ce), and Aulus Gellius’s Noctes Atticae (2nd century ce). These each fall into the category of florilegia: collections of potentially disconnected information (on, respectively, agriculture, the natural world, and grammatical and legal thought) which required “headings” to organize the consultative, discontinuous reading they were likely to solicit. These proto-encyclopedic works were not written with an eye toward any conceptual unity, and they do not imagine immersed, continuous reading. Chapters are, in this period, the accompaniment of information culture rather than literary culture, a scanning reading rather than an immersive one.5 As a result, the arrival of chapters in literary or more specifically fictional texts is quite late; Greek and Roman fictional romances, from Achilles Tatius’s Leucippe and Clitophon to Heliodorus’s Ethiopica, were still composed in “books.”

Take, for instance, the famous moment of Augustine’s conversion from his Confessions. As Augustine narrates it, he understood the song of nearby children—Tolle, lege, “take it and read”—to be a sign to “open the book, and read that chapter which I should first light on (legerem quod primum caput invenissem).” And so he does: “I snatched it up, I opened it, and in silence I read that chapter which I had first cast my eyes upon (Arripui, aperui et legi in silentio capitulum, quo primum coniecti sunt oculi me).”6 While most scholars agree that capitulum here means pericope, or short passage, it is nonetheless important that the vocabulary of chapters comes into play in situations—like this famous moment of bibliomancy—where reading in snatches, rather than continuously, is apparent.

At this early point in chapters’ history, they are usually signaled by simple numbers in the margin of the text, with headings, should there be any, listed in a separate prologue frequently referred to as capitula lectionum or capita rerum. Chapter headings do not begin to intervene in the body of texts until much later, the 12th and 13th centuries in particular, although there are a few significant exceptions. The example of Gellius’s Attic Nights is particularly useful here: its twenty books were divided into 398 chapters, whose headings or titles directly follow the collection’s preface; not until certain manuscripts of the Carolingian era were the titles redistributed to the individual books in which their chapters were found, and only later still were they each inserted into the manuscript to directly precede each individual chapter.7 The history of how Gellius’s capita were understood by editors demonstrates the gradual transition of “headings” into distinct “units,” but his list of headings, one of the few complete, presumably authorial collections from Latin antiquity remaining to us, is equally revealing, as the following samples from Book Two demonstrate:


Quo genere solitus sit philosophus Socrates exercere patientiam corporis; deque eiusdem viri temperantia (“How Socrates used to train himself in physical endurance; and of the temperate habits of that philosopher”)


Quae ratio observatioque officiorum esse debeat inter patres filiosque in discumbendo sedendoque atque id genus rebus domi forisque, si filii magistratus sunt et patres privati; superque ea re Tauri hilosophi dissertatio et exemplum ex historia Romana petitum (“What rules of courtesy should be observed by fathers and sons in taking their places at table, keeping their seats, and similar matters at home and elsewhere, when the sons are magistrates and the fathers private citizens; and a discourse of the philosopher Taurus on this subject, with an illustration taken from Roman history”)


Qua ratione verbis quibusdam vocabulisque veteres immiserint h litterae spiritum (“For what reason our forefathers inserted the aspirate h in certain verbs and nouns”)8

These are “headings” that modulate between two functions, the referential and the indexical. They attempt a relatively full account of the contents of each unit, while advertising in a brief form the subject at hand—be it biographical, cultural-historical, or linguistic in nature—in order to guide either a rapidly scanning, or merely curious, reader.

Those chapter headings are, put another way, enticements, solicitations of interest. As such, their task is to modulate between the readerly need for a finding aid—to locate matters of interest—and a desire, whether that of the reader or of the cunning writer, to stray from that initial interest. Gellius’s own description of his chapter headings is characteristic but telling:

Capita rerum quae cuique commentario insunt, exposuimus hic universa, ut iam statim declaretur quid quo in libro quaeri invenirique possit. (“Summaries of the material to be found in each book of my Commentaries I have here placed all together, in order that it may at once be clear what is to be sought and found in every book.”)9

Discontinuous, consultative reading is, in other words, the rationale behind chaptering, which is here identified with the summary-work of the “heading” rather than the division of the units themselves, although the latter is clearly implied by the former. The capita here imply a seeking, a curiosity, and advertise a possible finding, while leaving open the possibility that what is to be found need not necessarily overlap with what was sought.10

This rationale for the chapter as a textual unit had a remarkable durability. It was already to be found in Pliny, whose dedicatory preface to the Natural History explained to Vespasian that “I have subjoined to this epistle the contents of each of the following books, and have used my best endeavors to prevent your being obliged to read them all through. And this, which was done for your benefit, will also serve the same purpose for others, so that any one may search for what he wishes, and may know where to find it.”11 We can find it again in the 4th-century Peri Kataskeues Anthropou (On the Making of Man) of Gregory of Nyssa: “For clearness’ sake I think it well to set forth to you the discourse by chapters (kephalaion), that you may be able briefly to know the force of the several arguments of the whole work.”12

It is still the stated rationale four hundred years later, when the Ostrogothic textual scholar Cassiodorus explains, in his Institutions of Divine and Secular Learning, why the Bibles produced by his scriptorium include chapter headings at the start of each biblical book:

To make the text of the Octateuch available to us in a summarized version, I thought that the chapter-headings (titulos) should be set down at the beginning of each book, chapter-headings that had been written by our ancestors in the course of the text. The reader might thus be usefully guided and made profitably attentive, for he will easily find everything he is looking for, seeing it briefly marked out for him.13

Jump ahead five hundred years to the middle of the 12th century, and one finds this, from the prologue to Peter Lombard’s Four Books of Sentences, one of the first medieval compilations of patristic wisdom:

Moreover, so that what is sought may more easily come to meet the reader, we have prefaced this work with titles [titulos], by which the chapters [capitula] of each of the books is distinguished.14

One might note that here, by the 12th century, the terminology for title versus unit has been settled, but the language of “seeking” remains the same. It would linger well into the 18th century, albeit now with an openly commercial connotation, as when in Joseph Andrews (1742) Henry Fielding refers to chapter titles as “so many Inscriptions over the Gates of Inns … informing the Reader what Entertainment he is to expect, which if he likes not, he may travel on to the next.”15

If early chapterings are associated with prose compilations like Pliny’s or Gellius’s, by the 4th and 5th centuries of our era they are much more prevalent across a greater variety of textual types, for two reasons: the triumph of prose, and the triumph of the codex. The inhibiting unity of the scroll had largely yielded, by the 5th century, to the codex, which, by virtue of its more flexible size, was commonly understood as a collection or anthology of disparate texts. As Isidore of Seville put it in his 7th-century work Etymologies, Codex multorum librorum est; liber unius voluminis (“A codex is composed of many books; a book is of one scroll”).16 That is, textual culture had migrated to a point where the unity of material container and textual contents was no longer conceptually so firm. Texts could be of any length, particularly in prose; and as such, they began to borrow segmenting techniques from the most elastic genre of antiquity, the florilegium or compendium. The 4th century in particular is a key moment in the early history of the chapter: continuous prose texts like Eusebius’s Ecclesiasical History were composed in chapters, no longer in books, while older prose narratives—such as, crucially, the Gospels—were chaptered by editors producing the new codex bibles called for by the mass Christianization of the Roman empire. Discontinuous or consultative reading, and its textual accompaniment the chapter, had migrated from the periphery to the core of Western literary culture, while the previous avatars of literary unity—the scroll, the “book”—began to fade.

The Biblical Paradigm

While instances of authorial control over chaptering, such as those of Pliny or Gellius, can be revealing, the norm in antiquity and early medieval Europe was quite otherwise: the division of texts into capita or kephalaia was more often the work of editors and scholars. In fact, the careful dividing and labeling of textual units was a serious, if currently obscure, aspect of intellectual labor in the centuries before print. Bede is a key example; among the list of works attributed to him is a series of capitula lectionem, or “summaries of readings” of scripture. The work here is analytic—to determine the correct structure of the text, to divide it into monads that represent the curve of an argument, or the wholeness of a scenic unit—and compositional, as Bede was careful to produce headings that were elegantly varied in lexical style as well as appropriate and just in their encapsulation.17 To provide divisions and headings was understood less as a dispersal or deconstruction of the text than, in the words of Hugh of St. Victor, to “gather” it (colligere) into significance: “Now ‘gathering’ is reducing to a brief and compendious outline things which have been written or discussed at some length. The ancients called such an outline an ‘epilogue,’ that is, a short restatement, by headings, of things already said (brevis recapitulatio supradictorum appellata est).”18 Chaptering came to reflect the medieval efflorescence of the concept of ordinatio, where segmentation and labeling express a text’s inner logic.

It was an ancient practice, but that practice, certainly by the 12th century, now had a justification, one that in turn intensified and institutionalized the practice.19 In many ways its roots went back to the most significant scriptoria of late antiquity, such as Cassiodorus’s Vivarium, and before that to the first libraries and scriptoria to place Old and New Testaments in pandect codices, such as the library of Caesarea, which by the early decades of the 4th century could be called, with only minimal anachronism, a research institution.20 Chaptering as intellectual labor had a long lineage, but its style—and as a result its meaning and eventual fate as a technique within fictional prose—changed greatly in the period between Caesarea and the medieval university. Nowhere is that change more visible, and culturally influential, than in the evolution of the chaptered Gospel from the 4th to the early 13th century.

Roughly speaking, at some point in the first thirty years of the 13th century, either in the young university at Paris or at monasteries in England, the entirety of the Bible was divided into the chapters still in use. Tradition has assigned this work to Stephen Langton, who taught at Paris in the early years of the century before returning to his native England to become archbishop of Canterbury at a uniquely difficult moment in relations between the papacy and the British crown (years which, it bears noting, include Magna Carta). As the traditional story has it, in Paris Langton faced a fairly international set of theological students who brought with them Bibles divided according to a bewildering set of different systems, and scattered remarks by university officials—particularly Peter the Chanter, who was head of the university council in the 1190s—lamented this state of affairs.21 Thus, in the interest of classroom harmony—the need for a common scriptural text, the need for a stable system of citation—Langton took on the work of evolving a new system of chapters that, disseminated by the stationers of Paris in their popular student Bibles, grew into the standard system of the Christian West.

We should be careful not to discard the traditional account too quickly, but there are some problems with it. For one, most of the explicit evidence comes from significantly later, such as the attestation of the English Dominican annalist Nicholas Trivet, which cites Langton as the author of the new system over a century after the work would have been performed. For another, Langton’s own extensive commentaries on the Bible cite older chapter divisions, throwing into doubt the chronology of the matter. Last, recent work by the textual scholar Paul Saenger has centered on the Abbey of St. Albans in Hertfordshire, where an intriguing set of Bibles with modern chapter divisions was produced before the first Parisian examples with new chapters—and, even more intriguingly, with evidence of a Hebraic influence, suggesting that scholars at St. Albans, using the divisions (sederim) of the Palestinian lectionary cycle as a model, were the first to confect the new divisions, which at some point migrated to Paris.22 And yet, Langton was, in his commentaries, clearly interested in the matter of chapter divisions; in several instances he explicates tricky texts by adducing the evidence of the context of the chapters in which they fall, and he defends chapters by citing Jerome, who had justified them as a readerly aide-mémoire in the same fashion as my earlier examples.23 The matter is at present undecided. But it still can be said that, around 1200, the modern divisions came into being.

Langton and his unnamed predecessors in biblical chaptering faced a similar dilemma: balancing the need for producing stable citational units of more or less similar size (a question of standardization) with finding appropriate places in the text at which to anchor their units (a question of naturalization). Each system solved this problem differently; how differently, at least on the question of size, can be determined from Table 1. From Eusebius to Langton, biblical chapters get progressively larger, and their numbers fewer. The scriptorium at Caesarea evolved the so-called Eusebian canons, which, working off of an even older series of marginal distinctions called the Ammonian numbers, divided Gospel texts into short, several-sentence-length sections for comparison across the Gospels. But Eusebius and the Caesarean scriptorium are also linked, persistently if not conclusively, with the three oldest extant pandect Bibles—among them Vaticanus (or Codex B) and Alexandrinus (or Codex A). Unlike their perhaps slightly older cousin Sinaiticus, both Vaticanus and Alexandrinus have marginal chapter divisions—kephalaia in the Greek of these pandects—each in significantly greater numbers than in the modern version (with the odd exception of John in Alexandrinus). The size of the biblical chapter was smaller than we might expect, a trend that persisted into the Latin pandects and Gospel books of important scriptoria like Wearmouth-Jarrow in Northumbria, which produced Codex Amiatinus, perhaps the most significant exemplar of the Vulgate text.24

Table 1: Major Systems of Division of the Gospels, 3rd–13th centuries.

Ammonian/Eusebian (3rd–4th c.) [Caesarea]

Vaticanus “Codex B” (4th c.) [unknown]

Alexandrinus “Codex A” (4th–5th c.) [unknown]

Amiatinus Vulgate “A” (8th c.) [Northumbria]

Modern “Langton” (13th c.) [Paris]

























The gradual lengthening of the chapter from late antiquity to medieval Paris is one lesson. But Alexandrinus can take us even further. Unlike Vaticanus, Alexandrinus comes not just with marginal chapter notations but with titloi or chapter headings as well, which are placed in a list at the start of each Gospel, and even, on selected pages, as what we would now call running heads. The list itself furnishes us with a clue to the method of chapter division in late antiquity, in the nascent state of Christian biblical segmentation. The vast majority of these titloi begin with peri (“about,” “concerning”) followed by a genitive construction: thus, the fourth chapter of Mark is Peri tou leprou (“Concerning the leper”), while chapter 5 is Peri tou paralutikou (“Concerning the paralytic”). The form of these titloi would, in fact, be influential for centuries, dictating the traditional grammatical constructions of chapter titles in history, treatise, and novel, well into the 18th century.

More important, the items following the incantatory “peri” are illustrative. Starting with the first marked chapter (not the beginning of the Gospel itself, which is not a chapter, but considered an incipit)—Peri tou diamonidzomenou (“Concerning the demon-possessed man”)—the list is a list of others: it is structured as a list of Jesus’ encounters, with the leper, the paralytic, the man with the withered hand, the blind man, the Phoenician woman, and so on. Only once is Jesus mentioned in a titlos (“Concerning the transfiguration of Jesus,” when Jesus becomes an other); otherwise he is the still center around which others revolve. Occasionally distinct parables (of the sower, say), or distinct miracles (walking on water, the five loaves) make their own chapters as well. We might say, in fact, that the chapter divisions delineate discrete actions, something like Barthes’s famous proairetic code: encounters, moments of advice or rebuke or healing, particular speech acts. Biblical scholars have explained this tendency of Alexandrinus’s titloi as an emphasis on miracle narration over ethical or theological instruction: the miracles, usually performed on unnamed individuals like the demon-possessed man, “pop” out in the kephalaia.25 Others think of the kephalaia as essentially lectionary units, their brevity explained by their liturgical function.26 To a literary scholar, however, the tendency seems more generally one of action, atomized into small monads. A kephalaion, one might say, is a discrete act, with a beginning (Jesus meets someone), a middle (healing or rebuking, conversing), and an end (the “other” vanishes from narration).

It is also a hook for readerly memory. The function of those “others” is to be remembered: What happened to the leper? To the paralytic? Our readerly recollection snags on these particulars of people and actions, rather than abstractions like “charity” or “salvation.” And, given the texture of Gospel narration, these discrete actions are short. Take kephalaia 12 and 13 of Alexandrinus. Here is number 12 in its entirety, titled Peri tēs thugatros tou archisunagogon (“Concerning the daughter of the synagogue-ruler”); it is little more than an introduction and a difficulty—a dying daughter—proposed:

Then came one of the rulers of the synagogue, Jairus by name; and seeing him, he fell at his feet, and besought him, saying, ‘My little daughter is at the point of death. Come and lay your hands on her, so that she may be made well, and live.’ And he went with him. And a great crowd followed him and thronged about him.

Kephalaia 13 (Peri tēs haimorrouseis, “Concerning the woman with the flow of blood”) begins when, on his way to the Jairus’s house, Jesus is interrupted by a woman “who had had a flow of blood for twelve years,” whose touch of his garment briefly robs Jesus of his power. Yet 13 continues when, having healed this woman, Jesus and his by now large entourage enter Jairus’s house, where the girl arises healed after a brief command (Talitha cumi, “Little girl, I say to you, arise”). We might think of this as clumsy chaptering, which ignores the way in which the episode of the haimorrouseis is an odd interruption of an ongoing action, climaxing with Jairus’s daughter rising from her deathbed. But we might also see this as the way in which the kephalaia of Alexandrinus are not quite wholes, but instead incitationes—editorial markings for the moment when a new element (a new person, a new question or feat) enters the text for the first time. Put another way, their function is purely one of signals to a reader’s memory; as wholes they are either oddly quick, or sloppy, in either case disjunct with their surroundings. They are notabilia, orientation points, mnemonic hooks.

Not so with later chapterings, which segment this moment differently, largely by understanding it as a kind of narrative whole. One of the major chapter systems of medieval Europe—commonly known as the “Alcuinic” system, after the Northumbrian monk who became a major figure at the Carolingian court in the 8th century, and witnessed by the Northumbrian codex known as Amiatinus—begins its capitulum 14 of Mark with Jairus’s appearance and request, and ends with the reviving of Jairus’s daughter, thus understanding the episode of the haimorrouseis as an inset digression in the frame story of the daughter’s “rising.” Even the Alcuinic chapter title admits as much: Vadens mortuam suscitare filiam Iairi, mulierem a profluvio sanguinis, suscitat protinus et puellam (“He hurries to raise from the dead the daughter of Jairus, he heals the woman with the flow of blood, and then goes onward to raise the girl”).27 This would be the more commonsensically proairetic version, one that reads it as a single, if interrupted, action. That is, by the 8th century we are moving away from kephalaia’s notion of editorial and readerly hooks or incitations.

Much more so is the Langton or 13th-century chaptering. Both kephalaia 12 and 13 of Alexandrinus—as well as 11 and the brief sentence of 10—fit within chapter 5 of what is still our chaptering of Mark. Langton’s chapter 5 of Mark includes not only the episodes of Jairus’s daughter and the haimorrouseis, but also the possession of a man by Legion (cast into swine by Jesus, who then hurl themselves off a cliff into the sea). What binds these episodes into a single chapter? They are all miracle healings, it is true, but there are many of those in the first half of Mark. What organizes them into a single chapter here is much simpler: their location. Mark 5:1 (the last sentence of Alexandrinus 10) states: “They came to the other side of the sea, to the country of the Gerasenes”; Mark 6:1 (the middle of Alexandrinus 13), “He went away from there and came to his own country; and his disciples followed him.” Langton’s fifth chapter is, quite simply, a record of what happens in a particular location, the “country of the Gerasenes.” This is characteristic of the Langton chapters, which are largely organized by either place or time: either a single site, as here, or a single day. (Langton’s Mark 6, for instance, is a record of Jesus’ brief return to Galilee; his Mark 15 is a record of Jesus’ last day, starting with the chief priests bringing Jesus to Pilate. In each case the Alexandrian and Alcuinic systems break these up into smaller units.)

What should we make of this? As commonplace as the Langton rationales seem, they are nonetheless syntheses: attempts to bring several discrete actions under the umbrella of another, impersonal category. “Impersonal” because, for instance, none of the miracles recorded in Mark 5 is in any way characterized by being set in the country of the Gerasenes—that is simply a background element, a part of the Gospel’s realist notation. Similarly, the temporal structure of some of the Langton chapters—a day, a week—is beyond personal agency, not directly informative of the actions performed in that day. Geographical thinking, ephemeral or hebdomadal thinking are impersonal syntheses which, by placing the results of agency (discrete actions) into a category beyond agency, invent a new perspective: the distant, or retrospective, whole (as in, the things that happened that day; the things that happened in that place). If we extend this synthetic thinking to even broader categories—the things that happened at that visit, the things that happened when I was young—we approach something like the later novelistic chapter: a whole, even a meaning, that is only apparent in retrospect, whose governing aspect is not palpable at the time of the actions performed within it. “Alien syntheses” might as well be a motto for a kind of literary form that subsumes actions under the rubric of something vaster, something beyond us, something from the realm of the Author, the Narrator, not the agent. Readerly memory is no longer a sufficient explanation of their function; instead, these chapters are something like conceptual puzzles.

The foregoing is not intended to argue that Langton, or the unknown scribes of St. Albans, or whoever forged these early 13th-century biblical segmentations, invented the chapter as a literary form. Instead these should be viewed as an instance—although, given the cultural centrality of the Gospels, an important instance—in which an editorial tool, an aspect of the mise en page of late antiquity, slips the moorings of its initial justification in order to float into wider and deeper waters, into what we might call literary meaning, without ceasing to function as an editorial tool. Given the fact that the Paris Bible contained fewer chapters than its predecessors, it would seem that their function as citational units would be less efficient, in fact, than the numerous kephalaia of Alexandrinus or the capitula of Amiantinus. In one sense this is right; eventually these modern chapters would, in the 16th century, need to be joined by verse numeration to permit more exact citation. Nonetheless, they persist as citational units, while serving as much more than that. The 13th century is a major site, if not the only site, in which the textual functionality of the chapter—a paratext which aimed toward a particular reading mode—develops into something aslant of its original function, something that invents a new literary function that can parallel, in fact seamlessly accompany, its older rationale.

Early Canons of Narrative Chaptering

If the chapter moves, thanks to the late classical and medieval editors of Christian scripture, from an editorial unit to something like a literary form—a narrative unit structured by non-agential, abstract, or “alien” rubrics—we might expect early fictional narrative immediately to inherit this form. In fact, the novel oddly recapitulates the history of scriptural chaptering, moving again from short editorial units to longer abstract units, but this time often with an inherent self-consciousness of that movement, given that now chapters are just as often shaped by the decisions of writers as of editors, translators, or scholars. Even the chapterings of editors reveal themselves increasingly as aesthetic choices, poised toward a wealth of new understandings about how to manage the collision between a discontinuous, informational technique of segmentation and an immersive, continuous reading practice, one more propulsive than the lectio of the monasteries, which was emerging into cultural dominance.28 Prose, the instrument of that submergence in an unrolling, forward-oriented universe, was increasingly—almost inevitably—accompanied by segmentation into chapters, now its necessary counterpart.

Nowhere perhaps is this clearer than in the practice of mises en prose associated with the tastes of the 15th-century Burgundian court.29 Often anonymous prosateurs, in many cases commissioned by Philippe le Bon, turned to the Old French verse chansons de geste and rendered them in modernized Middle French prose, in the process compressing, abridging, and rearranging the work of verse epic sources for a courtly, urban audience. No small part of the modernization process of dérimage, or prosification, was division into chapters, into labeled units that did not exist in the original verse. Chaptering is essential to this prose, as is apparent in the prologue to an early 16th century prosification of the verse cycle of Lorraine by one of the few named prosateurs, Philippe de Vigneulles of Metz:

(“And therefore I, Philippe de Vigneulles the merchant, in honor of God and the city, decided to change the story from ancient verse or chanson de geste into chaptered prose, and as briefly as I could or knew how to do … because many people did not understand well the language that was used in the past, and they did not take pleasure in reading it because of its antiquity, and people now want to have short and pleasing things, as their minds become every day more acute and subtle. Therefore I inform all readers and listeners of this story that I the writer have abridged it … and that in the former story there are long unnecessary passages that I have omitted in order to avoid prolixity.” 30)

The modern virtues of prose—subtlety and precision—are produced by both abridgement and morcellation; the chapter is here part of a technique of modernization aiming at a more rapid form of comprehension. It is clearly seen here as a response to a cognitive demand with several facets: segmentation achieves an effect of speed (as frequent divisions mark progress more rapidly than non-segmented text), performs what cognitive science now calls “chunking” (subdividing in order to provide footholds for long-term memory), and in its brevity and labeling can express clarity or precision.

This morcellation is both a visual and conceptual technique, and it owes little to its verse sources. Neither Philippe de Vigneulles nor the earlier, anonymous prosateurs of Chrétien de Troyes’s Cligés and Erec et Enide adhere in their chapterings to the laisses or verse divisions of their source texts. Nor do they necessarily adhere to a notion of episodes defined by proairesis; in this they resemble Langton’s biblical chapters.31 Instead, the Middle French prosifiers developed an astonishingly flexible, if also somewhat messy and chaotic, roster of aesthetic effects for which chapters could be put to use. Their primary desideratum seems to have been to produce a page design that would display frequent divisions—roughly one chapter division per folio page, so that the eye would continually be greeted by aerations.32 Within that vague parameter, a great freedom was possible.

Even a partial catalogue of uses toward which chapter divisions could be put in the mises en prose reveals the elasticity and adaptability of the form:

Shift of protagonists or plot strand. First lines of chapters often announce a shift between two distinct narrative points of view or plots, and are used to manage the sometimes multiple, simultaneous diegetic strands within source texts.

Points of suspension. Heightened moments are often split directly by chapter division to achieve paradoxical effects of momentary suspense and dramatic tableaux; the prosateur of Erec et Enide, for instance, interrupts moments of combat, and at one point inserts a chapter break between Erec falling wounded at the point of death and Enide running to him.33 The break here both prolongs, and yet relieves, the effects of suspense.

Emphasis on pregnant, lingering final words or key speech acts. Important character pronouncements can be given added heft and significance by ending chapters with them; prosateurs increasingly understood chapter division as a technique of rhetorical amplification.

Encapsulating an action that spans multiple days. Diurnal transitions are occasionally buried within chapters, and at these occasions the framing of the chapter insists upon the priority of human agency by naming, often in the chapter title, an ongoing purpose.

Interrupting an ongoing action with reference to non-agential frames (such as days or location shifts). Similarly, actions can conclude, and even plot strands change, in the midst of chapters that are framed instead by temporal or locative principles.

Reminder of the narratorial or textual frame to events. The formula dist l’histoire, referring to the source text, often begins chapters, as a way of pulling the prose out of its immediacy and establishing the diegesis as framed within a documentary process. The chapter, that is, periodically reannounces the telling of the tale.

These are multiple ways in which the chapter division knits its segments into a definite shape. They can combine at moments; at other moments one or more can be flagrantly disobeyed—shifts in plot strands can happen in the middle of chapters as well as at chapter breaks; reminders of the narrative’s textual frame need not happen at the start of chapters. The procedure is ad hoc, supple, adhering above all to a general rhythm of length that itself was relatively flexible.34

All of these effects are exhibited by the 15th century’s most famous example of prose chaptering, the editing and publication by William Caxton in 1485 of Thomas Malory’s Morte d’Arthur. Thanks to the 1934 discovery of the Winchester Manuscript (or “W”), which shows signs of having been in Caxton’s print shop during the preparation of the 1485 Morte, we have some evidence of how Caxton made his divisions of the text into the twenty-one books and 507 titled chapters in which he printed it. Caxton’s preface is explicit about his divisions: “And for to vnderstonde bryefly the contente of thys volume / I haue deuyded it in to xxj bookes / and euery book chapytred as here after shal by goddes grace folowe.”35 The preface specifies the number of chapters in each book, a number which is variable enough to seem random. Again, the primary criterion seems to have been mise en page: 61 percent of Caxton’s pages include a chapter break, and only 24 percent of his pages fail to include any kind of white space or prose segmentation.36 The effort to provide relief to the eye at almost every moment (only eleven pairs of facing pages lack a segmentation of some kind) dictated the rough length of Caxton’s chapters, which average 668 words, with a range from 157 words at the shortest to more than 1,400 words at the longest. Like the prosateurs working shortly before him in Burgundy, Caxton understood the chapter as first a material practice—here, one involving the casting off of W or a similar copytext into units ensuring frequent visual breaks—but second as a highly variable aesthetic tool.

Figure 1: The Visuality of Caxton’s Chaptering [from Book 4, chapter 25].

© University of Manchester.

Careful analysis of W indicates that Malory used indications within it, such as its occasional large capitals, to guide many of his chaptering choices.37 Chaptering in Caxton is far from a free or materially undetermined practice. But within these constraints, important and influential effects can still be generated. Moments of suspension are even more common in Caxton than they are in the work of the prosateurs, often quite artfully so; in Book 17, the fourteenth chapter ends as Launcelot stops right before the door that will take him to the Sangrail, while the fifteenth chapter starts with the opening of that door. Itself a portal or threshold, the chapter division here insists on itself as a pause, a caesura within events that is generated from outside them. Often it is pure suspense that is generated: in Book 6 Launcelot comes to a castle and is greeted with ominous words “Soo he loked aboute / and sawe moche peple in dores and wyndowes that sayd fayr knyghte thou arte vnhappy.”38 The arrival of two giants is withheld until immediately after the chapter break. Some of the narrative’s most pivotal speech acts, such as Sir Palomides’s defiant proclamation to Tristram of his love for Isoud, that “loue is free for alle men,” conclude chapters without any response, giving them a weight simply by way of placement.39 Particularly beloved by Caxton are chapter endings that provide only a temporary, provisional closure, with an uneasy affect of some kind lingering and promising future results; when in Book 18 Sir Bors proclaims that “there was treason amonge vs,” the chapter ends with an ominously litotic promise of consequences to come: “Thenne some sayd to sire Bors we may wel bileue your wordes / and soo some of them were wel pleasyd/ and somme were not so.”40

Certainly not all of Caxton’s chapterings have these elegant, deft effects; most are merely mechanical, opening with a rote “Then” or “Now turn we unto”; but one can see how the brevity of the chapters, enforced by page design among other considerations, is emerging as a temporal structure within the narrative of some importance. The causes, in other words, were accidental or contingent: the legacy of the chapter as an editorial marking, as a visual element within the page design of prose, as a citational unit. The consequences were lasting, even if obscure at first. Among them was a critique with a long history, becoming slowly more prevalent by the 17th and 18th centuries, asserting that chaptered prose made complicated arguments, and protracted acts of comprehension and cognition, difficult; Locke made this complaint most famously.41 Subtending this critique, however, were the possibilities for thinking of narrative time as a practice of rhythmic interruption that medieval chaptered prose created. Not until the novel of later centuries, however, were those possibilities fully exploited.

The Novelistic Chapter

In The History of Charlotte Summers (1750), commonly attributed to Sarah Fielding, a Miss Arabella Dimple, lying half-naked in bed, calls her maid Polly to fetch “the first Volume of the Parish Girl I was reading in the Afternoon.” The maid returns and sits down with the book: “Pray, Ma'am, where shall I begin, did your Ladyship fold down where you left off?—No, Fool, I did not; the Book is divided into Chapters on Purpose to prevent that ugly Custom of thumbing and spoiling the Leaves; and now I think on't, the Author bid me remember, that I left off at the End of—I think it was the 6th chapter.”42 Such witty self-consciousness about the conventions of chapter division was common in the middle decades of the 18th century, as the period’s novels, fast becoming a cultural phenomenon, often reflected upon their implied contract with their readers. No small part of that contract was the chapter, which was felt to have both an indexical and a physiological function: both to mark or label, and, as Laurence Sterne put it in Tristram Shandy’s “chapter upon chapters,” to “relieve the mind,” or to “assist—or impose upon the imagination.”43 Both the indexical and the physiological effects of chapters are adaptations to the consumption of prolonged narrative, particularly to the problems of information management and cognitive endurance, presented by forms that require considerable time of the reader.

Time—the temporality of both reading and narrative, and the ways they intertwine—becomes the key aspect of the novel’s use of the chapter division. One salient aspect of the history of the chapter from the mid-18th century through the Victorian period was its increasing naturalization; metacommentary like that in Charlotte Summers and Tristram Shandy dwindles as the chapter develops into a way not simply to mark readerly or extradiegetic time, but a kind of phenomenology of time within realist fiction. As is continually the case in the accumulative history of the chapter, all the techniques developed during ancient and medieval editorial practice remain available for use by the novelist, but are put to new use. In particular, the medieval understandings of the chapter—as a means to indicate shifts of plot strand, to highlight moments of suspense, to lend prominence to particular speech acts, and to remind the reader of the narrative frame, whether enunciatory or documentary—are retained, but subordinated to an obsessive notation of diurnal time. The chapter thus is set against the day, which is in turn compared to units of biographical self-understandings of characters (“periods,” “epochs,” “moments”).

Middlemarch is an excellent example of this tripartite balance. It contains 146 days of narrated action, set within eighty-six chapters.44 Diurnal boundaries in Eliot are usually clear, noted by innocuous adverbial phrases. The most common such phrases involve “day” itself: “the next day” and “one day,” as in the following sentences:

But Fred did not go to Stone Court the next day, for reasons that were quite peremptory. (Ch. 26)

In this way it happened that one day near four o’clock, when Mr. Brooke and Ladislaw were seated in the library, the door opened and Mrs. Casaubon was announced. (Ch. 39)45

Following these in frequency are “the next morning,” “one morning,” “one evening,” and “the next evening.” Sequence clearly matters here—the linking of one unit to the next—and there is a decisive preference in favor of daylight, of solar time. At times the transitions are as minor as the deictic “now,” dropping us into a scene. There are also occasional unmarked transitions, where only the subtle swoop from a narratorial present tense or modal form to a preterite summons us to a new day and scene. The diurnal or solar boundary, that is, tends to be clear in the Victorian serial novel, clearer than any more calendrical index, and is roughly aligned with the chapter. Middlemarch’s present action begins on September 30, 1829, when Sir James Chettham comes to dine with the Brookes; it ends more vaguely in May of 1832. But we can say that the novel’s main time scheme covers almost a thousand days of elapsed time, of which 146 have been selected for narrative representation—something like 15 percent, scarcely a continuous temporality.

More importantly, these indications of diurnal shifts tend to occur at or near the openings of chapters. Eliot tends to mark the textual divisions of book, serial number, or chapter by starting them with a new day of diegetic time, although at key moments that marking gives way. In Middlemarch, eighteen of its 146 days take more than one chapter to narrate, spanning two to four chapters; among those eighteen, three times one multiple-chapter day directly follows another. The last of these pairs, as an example, is a six-chapter dilation spanning Dorothea’s attempts to reconcile Rosamond and Lydgate and Rosamond’s admission to Dorothea of Ladislaw’s love for her. All of these three pairings involve middle-of-the-night activity: a strained marital conversation, a death, a breakdown. They are days unnaturally elongated and joined, sleepless and blurred. These multiple-chapter days, however, cluster in the novel’s middle sections, as Figure 2 demonstrates.

Figure 2: Chapters and Days in Middlemarch.

The first and last books of Middlemarch present us with a rhythm in which the number of narrated days roughly equals the number of chapters; that is, at the novel’s beginning and end there is a balance obtained between diegetic/diurnal time and textual segmentations. But in the novel’s middle—when plot complexity has reached its maximum, and when the number of interrelationships being tracked is at its presumed height—this rhythm is thrown off. Chapters become temporally capacious, their boundaries porous. In Books 5 and 6, the number of single-day chapters declines to a minimum, while the number of narrated days increases. Days begin to hurry; less sealed off from each other, they encapsulate briefer events that are less apt to conclude with the close of the day’s human activity. Their frames, that is—waking and rest—become weaker indices of significance; the path of narrative energy no longer so neatly adheres to the path of the sun. Then, with the resolutions of Books 7 and8, the boundaries of the day and the chapter begin to cohere once more, as if Middlemarch’s dominant chord has been resolved back into the tonic.

In the beginning and end of the novel, then, a concord between extrinsic time—the conventional, public time of the day—and textual time, the time of the chapter, is produced. Neither time unit, however, is necessarily psychological or agential time: their frames tend to ironize or situate even fairly strong proairetic moments. Take Middlemarch’s chapter 18, which moves relentlessly toward a decisive action—Lydgate’s simultaneously reluctant and defiant vote for Tyke as chaplain of the new fever hospital—and then veers away in its final four paragraphs to Lydgate’s low estimate of Farebrother’s willpower. Chapter 4, which, while almost concluding on another decisive action—Mr. Brooke’s delivery of Casaubon’s letter of proposal to Dorothea—finishes instead on Brooke’s perplexity before the question of female motivation. Eliot’s chapters end often on aphoristic reflections, but those aphorisms are just as often inconsequent, aslant of or curiously unresponsive to the chapter’s récit.

This is the realist novel’s cunning adaptation of the “alien syntheses” of medieval chapterings: the temporal frames it uses (the day, the chapter) tie action to boundaries which do not respect the temporality of individual agency—the temporal projection of individual plans, decisions, strategies. A single example from Anthony Trollope will suffice to illustrate this. Trollope is very much of his moment in his chaptering technique: while in earlier chaptered novels, such as those of the Fieldings, chapter division called for much explicit, ironic, and self-referential comment, by the time of Trollope these have largely vanished. Chapter titles in Trollope no longer predominantly reflect the genitive or locative syntax of classical forebears (beginning with “In which” or “Of”), but are instead short, elliptical mottos, names, or brief tags, which bear at first no obvious relation to the events of the chapter. Nor are those events straightforwardly proairetic, the narration of an action and its consequences. Instead, in a virtuoso manner implying a thorough (if implicit) consciousness of the chapter form of scripture as he knew it, Trollope continually writes chapters whose boundaries are set by multiple forms of abstract, alien, non-agential rhythms: circadian rhythms, such as the biochemical or physiological rhythms of daily time (waking, full attention, exhaustion, sleep); the rhythms of social time (gatherings, meals, visits); the rhythms of the attention of others (the light cast on our actions by the presence of an onlooker). If the chapter is an “event,” in Trollope it is only an event at the level of the textual unit, or for its consumer—not for any of its participants.

Chapter 23 of The Small House at Allington, entitled “Mr. Plantagenet Palliser,” is concerned with almost anything but that. In general it narrates the afternoon and evening when Adolphus Crosbie, the ambitious and opportunistic government clerk with aristocratic friends and tastes, decides to propose marriage to Lady Alexandrina de Courcy, despite the fact that he has already become affianced to the much poorer, if purer, Lily Dale. But such a pivotal moment in the novel’s plot—in fact, the central traumatic act of treachery whose consequences the rest of the novel will elaborate—is set deliberately within a structure irrelevant to Crosbie’s key decision. Just like the episode of Jairus’s daughter, interrupted oddly by the haimorrouseis, the chapter narrates two separate conversations between Crosbie and Lady Alexandrina—the first flirtatious, the second decisive—interrupted by a description (the first in the Barchester series of a main protagonist of the later Palliser series) of Plantagenet, who is on a brief visit to Courcy Castle en route to a meeting with his Parliamentary constituents. Palliser is engaged in a comically tepid flirtation of his own with the married, timid Lady Dumbello, and this engages the hyperbolic attention of the castle’s gossipy guests, who are, of course, equally interested in the much more serious dalliance between Crosbie and Lady Alexandrina. In this sense, the chapter title reflects the passive, distantly interested perspective of the castle’s guests, who believe the house party to have been deliberately contrived to throw Palliser and Lady Dumbello together (and who, to the extent they notice Crosbie’s dilemma, accord it second place in their attention). Put another way, the chapter as a unit is structured by the interests of somewhat ignorant, quite marginal minor characters, who, in a kind of grace note common to Trollopian narration, are continually described in the same room as characters engaged in important actions, half-listening; when Crosbie finally makes his disastrous proposal, he is in the same room as Lady Alexandrina’s mother, “reading a work on the Millenium, with a light to herself in one corner,” and “a young gentleman and lady … playing chess.”.46 For Crosbie, this is the pivotal act of his life. For the novel, it is—with evident irony—set within a world oriented elsewhere.

And not just conceptually. The chapter begins with Crosbie’s ruminations in midafternoon, but ends that night, with a deliberately irrelevant bedtime conversation between one of Lady Alexandrina’s brothers and his wife about the exact nature of the succession of the dukedom to which Palliser is heir. Exasperated by his wife’s questions, George de Courcy mutters, in the chapter’s last words, “I’m going to sleep.”47 The course of the day, or a portion thereof—from “the half-hour which always occurs before the necessity for dinner preparation has come” to actual sleep—is as structurally important to the chapter as anything about Crosbie’s decisions or actions: in other words, the things that happened between the onset of evening and its close.48 Crosbie may have just ruined his life and Lily’s, but around him is a houseful of people who know or care little, and their perspective governs the chapter as a unit, which is, at least officially, “about” Palliser. What we do, in Trollope, is always set in the context of those frames—the opinions of others, the passing of daily time—that proceed inexorably and without reference to us. This is a kind of narrative puzzle, akin to those of the Langton chapter system (something like, why is this chapter a chapter?), but with an unsettlingly ironizing solution: what motivates a chapter’s size and shape is a rhythm and a consciousness alien to our actions, our reasons, our tragedies. Trollope’s chapters, that is, adopt a centuries-old structural style associated with narrative “chapters,” and do so to an entirely different, recognizably novelistic goal, to ironize human agency.

Trollope’s chapter is exemplary not only for the manner of its operation but for its sheer innocuousness. It is finally one among countless chapters, in his work and in the fiction of the 19th century—a novelistic technique so ordinary and unavoidable that any attempt to instantiate it or summarize its purpose is likely to seem arbitrary and partial. That innocuousness and conventionality, the subtlety of the way in which its effects are achieved and managed, is, however, the key to its survival. Although the chapter was sometimes discarded within modernism (as in the work of Samuel Beckett) or occasionally reified into something far more hard-edged (as in the distinct stylistic worlds of the chapters of James Joyce’s Ulysses), it remains a familiar aspect of narrative fiction; chapter titles may come in and out of fashion, chapter lengths may be entirely flexible, but the idea of a textual unit that is also a temporal unit—a unit that is oblique to the agency of characters within a fiction—has persisted; it is perhaps the most notable, if usually invisible, debt the novel has to premodern textuality.

Links to Digital Materials

Pericope: A collective of biblical scholars and paleographers who study how the Bible has been differently segmented over time.

The Malory Project: A comparison (with images) of Caxton’s chaptered version of Malory and the unchaptered “Winchester manuscript.”

Further Reading

  • Blair, Ann. Too Much to Know: Managing Scholarly Information before the Modern Age. New Haven, CT: Yale University Press, 2010.
  • Brown, Marshall. “Plan vs. Plot: Chapter Symmetries and the Mission of Form.” Stanford Literature Review 4 (1987): 103–136.
  • Doutrepont, Georges. Les Mises en prose des Épopées et des Romans chevaleresques du XIVe au XVIe siècle. Brussels: Palais des Académies, 1939.
  • Drucker, Joanna. “Graphic Devices: Narration and Navigation.” Narrative 16.2 (2008): 121–139.
  • Fredouille, Jean-Claude, Marie-Odile Goulet-Cazé, Philippe Hoffmann, and Pierre Petitmengin, eds. Titres et articulations du texte dans les oeuvres antiques: Acts du Colloque International de Chantilly, 13–15 Décembre 1994. Paris: Institut d’Études Augustiniennes, 1997.
  • Genette, Gérard. Paratexts: Thresholds of Interpretation. Translated by Jane Lewin. Cambridge, U.K.: Cambridge University Press, 1997.
  • Grafton, Anthony, and Megan Williams. Christianity and the Transformation of the Book: Origen, Eusebius, and the Library of Caesarea. Cambridge, MA: Harvard University Press, 2006.
  • Hunter, J. Paul. “From Typology to Type: Agents of Change in Eighteenth-Century English Texts.” In Cultural Artifacts and the Production of Meaning: The Page, the Image, and the Body. Edited by Margaret Ezell and Katherine O’Brien O’Keeffe, 41–69. Ann Arbor: University of Michigan Press, 1995.
  • Light, Laura. “The Bible and the Individual: The Thirteeenth-Century Paris Bible.” In The Practice of the Bible in the Middle Ages. Edited by Susan Boynton and Diane Reilly, 228–246. New York: Columbia University Press, 2011.
  • Martin, Henri-Jean. La Naissance du livre moderne: Mise en page et mise en texte du livre français (XIVe-XVIIe siècles). Paris: Éditions du Cercle de la Librairie, 2000.
  • Rouse, Richard H., and Mary A. Rouse. “Statim Invenire: Scholars, Preachers and New Attitudes to the Page.” In Renaissance and Renewal in the Twelfth Century. Edited by Robert Benson, Giles Constable, and Carol Lanham, 201–225. Oxford: Oxford University Press, 1982.
  • Saenger, Paul. “The Anglo-Hebraic Origins of the Modern Chapter Division of the Latin Bible.” In La fractura historiográfica: Las investigaciones de Edad Media y Renacimiento desde el tercer milenio. Edited by Javier San José Lera, 177–202. Salamanca: Seminario de Estudios Renacentistas and Sociedad de Estudios Medievales y Renacentistas, 2008.
  • Stallybrass, Peter. “Books and Scrolls: Navigating the Bible.” In Books and Readers in Early Modern England. Edited by Jennifer Andersen and Elizabeth Sauer, 42–79. Philadelphia: University of Pennsylvania Press, 2002.
  • Stevick, Philip. The Chapter in Fiction: Theories of Narrative Division. Syracuse, NY: Syracuse University Press, 1970.


  • 1. Diodorus Siculus, Bibliotheca Historica, vol. 3, translated by C. H. Oldfather (Loeb Classical Library; Cambridge, MA: Harvard University Press, 1970), 98–99.

  • 2. Henri Marrou, “La division en chaptires des livres de La Cité de Dieu,” in Mélanges Joseph de Ghellinck, S. J., vol. 1 (Gembloux: Duculot, 1951), 249. For more on the literary unity of the ancient “book,” see Marguerite Harl, “Recherches sure le Peri Archon d’Origène en vue d’une nouvelle edition: La division en chapitres,” Studia Patristica 3 (1961): 57–67; Louis Holtz, “Les mots latin désignant le livre au temps d’Augustin,” in Les débuts du codex, edited by Alain Blanchard (Turnhout: Brepols, 1989), 105–113.

  • 3. Exceptions, of course, exist: the twelve chapters of Malcolm Lowry’s Under the Volcano (1947), meant to mirror the twelve hours of the day it narrates, offer a kind of numerological significance, matched to clock time. In general, however, this is a rarity, particularly prior to the 20th century.

  • 4. See Andrew Plaks, “Shui-hu Chuan and the Sixteenth-Century Novel Form: An Interpretive Reappraisal,” Chinese Literature: Essays, Articles, Reviews 2.1 (1980): 3–53.

  • 5. One potential source for the use of capita in Roman encyclopedic texts derives from the sectioning of Roman legal statutes; see Shane Butler, “Cicero’s capita,” in The Roman Paratext: Frame, Texts, Readers, edited by Laura Jansen (Cambridge, U.K.: Cambridge University Press, 2014), 73–111.

  • 6. Augustine, Confessions, vol. 1, translated by William Watts (Loeb Classical Library; London: Heinemann, 1912), 464–465.

  • 7. See Leofranc Holford-Strevens, Aulus Gellius: An Antonine Scholar and His Achievement (Oxford: Oxford University Press, 2003), 30.

  • 8. Aulus Gellius, Noctes Atticae (The Attic Nights), vol. 1, translated by John C. Rolfe (Loeb Classical Library; Cambridge, MA: Harvard University Press, 1967), xliv–xlv.

  • 9. Ibid., xxxvi.

  • 10. See Joseph Howley’s As the Romans Read: “Aulus Gellius and Imperial Intellectual Culture” (Cambridge, U.K.: Cambridge University Press, 2017), which argues for the definite disjunction between Gallium’s headings as a device for discontinuous ‘seeking,’ and the literary sophistication by which chapters present unexpected elements to be ‘found’.”

  • 11. Pliny the Elder, Natural History, vol. 1, translated by H. Rackham (Loeb Classical Library; Cambridge, MA: Harvard University Press, 1967), 20–21.

  • 12. Gregory of Nyssa, “De hominis opificio,” Patrologiae Cursus completus (Series Graeca), vol. 44, edited by J. P. Migne (Paris: Migne, 1865), 128.

  • 13. Cassiodorus, Institutions of Divine and Secular Learning, translated by James Halporn and Mark Vessey (Liverpool: Liverpool University Press, 2004), 115.

  • 14. Peter Lombard, Magistri Petri Lombardi Parisiensis episcopi Sententiae in IV libris distinctae, edited by Ignatius Brady (Spicilegium Bonaventurianum 4; (Grottaferrata: Editiones Collegii S. Bonaventurae ad Claras Aquas, 1971), 4.

  • 15. Henry Fielding, The History of the Adventures of Joseph Andrews, vol. 1 (London: J. Murray, 1742), 134.

  • 16. Isidore of Seville, The Etymologies of Isidore of Seville, translated by Stephen Barney, W. J. Lewis, J. A. Beach, and Oliver Berghof (Cambridge, U.K.: Cambridge University Press, 2006), 142.

  • 17. For the list of Bede’s capitula, see the Bede, Ecclesiastical History of the English People, vol. 2, translated by J. E. King (Loeb Classical Library; Cambridge, MA: Harvard University Press, 1954), 384–386; see also Paul Meyvaert, “Bede’s Capitula Lectionum for the Old and New Testaments,” Revue Bénédictine 105 (1995): 348–380.

  • 18. Hugh of St. Victor, Didascalicon of Hugh of St. Victor: A Medieval Guide to the Arts (New York: Columbia University Press, 1961), 93. For an explanation of this scholarly technique as a Gothic tendency to impose precise boundaries onto spatial and temporal artifacts, one that spanned architecture and chaptering alike, see Brian Stock, The Implications of Literacy: Written Language and Models of Interpretation in the Eleventh and Twelfth Centuries (Princeton, NJ: Princeton University Press, 1983), 82.

  • 19. See Malcolm Parkes, “The Influence of the Concepts of Ordinatio and Compilatio on the Development of the Book,” in Medieval Learning and Literature: Essays Presented to Richard William Hunt, edited by J. J. G. Alexander and M. T. Gibson (Oxford: Oxford University Press, 1976), 115–141. For an argument that leans more toward ordinatio as a bibliographical concept than as an analytic one, see Richard H. Rouse and Mary A. Rouse, “Statim Invenire: Scholars, Preachers and New Attitudes to the Page,” in Renaissance and Renewal in the Twelfth Century, edited by Robert Benson, Giles Constable, and Carol Lanham (Oxford: Oxford University Press, 1982), 201–225.

  • 20. See Anthony Grafton and Megan Williams, Christianity and the Transformation of the Book: Origen, Eusebius, and the Library of Caesarea (Cambridge, MA: Harvard University Press, 2006), 215.

  • 21. See Amaury d’Esneval, “La Division de la Vulgate latine en chapitres dans l’édition parisienne du XIIIe siècle,” Revue des Sciences Philosophiques et Théologiques 62 (1978): 559–568; Beryl Smalley, The Study of the Bible in the Middle Ages (South Bend, IN: University of Notre Dame Press, 1964), 222–224; Mary Carruthers, The Book of Memory: A Study of Memory in Medieval Culture (Cambridge, U.K.: Cambridge University Press, 1990), 93–99; Laura Light, “The Bible and the Individual: The Thirteenth-Century Paris Bible,” in The Practice of the Bible in the Middle Ages: Production, Reception, and Performance in Western Christianity, edited by Susan Boynton and Diane Reilly (New York: Columbia University Press, 2011), 228–246; Paulin Martin, Introduction à la critique générale de l’ancien testament: de l’origine du Pentateuque, vol. 2 (Paris: Maisonneuve, 1888), 461–474. For an account of Langton’s career as a whole, see F. M. Powicke, Stephen Langton (Oxford: Oxford University Press, 1928).

  • 22. Paul Saenger, “The Anglo-Hebraic Origins of the Modern Chapter Divisions of the Latin Bible, in La fractura historiográfica: Las investigaciones de Edad Media y Renacimiento desde el tercer milenio, edited by Javier San José Lera (Salamanca: Seminario de Estudios Renacentistas and Sociedad de Estudios Medievales y Renacentistas, 2008), 177–202. See also Joop H. A. van Banning, “Reflections upon the Chapter Divisions of Stephan Langton,” in Method in Unit Delimitation, edited by Marjo Korpel, Josef Oesch, and Stanley Porter (Leiden, The Netherlands: Brill, 2007), 141–161.

  • 23. Smalley, Study of the Bible, 224.

  • 24. For a fuller list of chaptering systems and their variations, see Donatien de Bruyne’s magisterial work of research and collation, Sommaires, rubriques et divisions de la Bible latine (Namur: Auguste Godenne, 1914). De Bruyne’s work should be supplemented with some revised statistics offered by Patrick McGurk, Latin Gospel Books, from A.D. 400 to A.D. 800 (Paris: Éditions Érasme, 1961).

  • 25. See Greg Goswell, “Early Readers of the Gospels: The Kephalaia and Titloi of Codex Alexandrinus,” Journal of Greco-Roman Christianity and Judaism 6 (2009): 134–174.

  • 26. This was the contention of Philip Carrington in According to Mark: A Running Commentary on the Oldest Gospel (Cambridge, U.K.: Cambridge University Press, 1960), worked through with mathematical exactness. See also Kurt Aland and Barbara Aland, The Text of the New Testament: An Introduction to the Critical Editions and to the Theory and Practice of Modern Textual Criticism, translated by Erroll Rhodes (Leiden, The Netherlands: Brill, 1987), 252. Such a theory, if true, nonetheless fails to explain why the necessary number of divisions (to cover the liturgical year) were placed exactly where they were.

  • 27. See Codex Amiatinus: Novum Testamentum Latine, Interprete Hieronymo, edited by Constantine Tischendorf (Leipzig: Avenarius and Mendelssohn, 1854). The translation of the chapter title is mine.

  • 28. Its slow emergence remains a key element in cultural histories of the late Middle Ages. Ivan Illich’s description of a “new way of reading” in the 12th century, one that “no longer creates an auditory and, therefore, social space,” is paradigmatic: “The reader flips through the pages. His eyes mirror the two-dimensional page.” Ivan Illich, In the Vineyard of the Text (Chicago: University of Chicago Press, 1993), 81–82. See also Armando Petrucci, Writers and Readers in Medieval Italy: Studies in the History of Written Culture, translated by Charles Radding (New Haven, CT: Yale University Press, 1995).

  • 29. The foundational study here is Georges Doutrepont, Les Mises en prose des épopées et des romans chevaleresques du XIVe au XVIe siècle (Brussels: Palais des Académies, 1939). See also Mettre en prose aux XIVe–XVIe siècles, edited by Maria Colombo Timelli, Barbara Ferrari, Anne Schoysman (Turnhout: Brepols, 2010).

  • 30. Philippe de Vigneulles, Le chanson de geste de Garin le Loherain, mise en prose par Philippe de Vigneulles, de Metz (Paris: Henri Leclerc, 1901), 2–3. The translation is mine. For a detailed study of his practice, see Catherine Jones, Philippe de Vigneulles and the Art of Prose Fiction (Cambridge, U.K.: D. S. Brewer, 2008).

  • 31. See Maria Colombo Timelli’s note in her edition of the prose version of Cligés: “la coincidence episode/chapitre représente l’exception plutôt que la règle,” in Le Livre de Alixandre empereur de Constantinople et de Cligés son filz: roman en prose du XVe siècle, edited by Maria Colombo Timelli (Geneva, Switzerland: Droz, 2004), 21. See also Jones, Philippe de Vigneulles, 89, who agrees.

  • 32. See Doutrepont, Mises en prose, 468–475; Jones, Philippe de Vigneulles, 89.

  • 33. See the break between chapters 32 and 33: L’Histoire d’Erec en prose, roman du XVe siècle, edited by Maria Colombo Timelli (Geneva, Switzerland: Droz, 2000), 196.

  • 34. Colombo Timelli, L’Histoire d’Erec, 31, tells us that the median chapter length of the prose Erec is seventy-nine lines of text, with its longest and shortest chapters being 150 and thirty lines, respectively.

  • 35. Malory, Le morte darthur (London, 1485), 2v. Facsimiles of both the Caxton first printing and the Winchester Manuscript are available for comparison at The Malory Project, directed by Takako Kato and designed by Nick Hayward.

  • 36. See Takako Kato, Caxton’s Morte Darthur: The Printing Process and the Authenticity of the Text (Oxford: Society for the Study of Medieval Languages and Literature, 2002), 29–31.

  • 37. In W, 104 of the 106 large capitals coincide with chapter beginnings in Caxton’s text: see William Matthews, “The Beseiged Printer,” in The Malory Debate: Essays on the Texts of Le Morte Darthur, edited by Bonnie Wheeler, Robert Kindrick, and Michael Salda (Cambridge, U.K.: D. S. Brewer, 2000), 54–55.

  • 38. Malory, Le morte darthur, 99v.

  • 39. Ibid., 284r.

  • 40. Ibid., 367r.

  • 41. In 1707 Locke wrote that the text of Paul’s Epistles were “so chop’d and minc’d, and as they are now Printed, stand so broken and divided, that not only the Common People take the Verses usually for distinct Aphorisms, but even Men of more advanc’d Knowledge in reading them, lose very much of the strength and force of the Coherence, and the Light that depends upon it. Our Minds are so weak and narrow, that they have need of all the helps and assistances can be procur’d, to lay before them undisturbedly, the Thread and Coherence of any Discourse; by which alone they are truly improv’d, and lead into the Genuine Sense of the Author.” See John Locke, Essay for the Understanding of St. Paul’s Epistles, by Consulting St. Paul Himself (London, 1707), vii. Locke’s argument against chaptered prose, particularly biblical prose, had a long lineage; in the early 14th century Nicholas of Lyra lamented the chaptered Bibles of his day as follows: “They have chopped the text into so many small parts … that to some degree they confuse both the mind and memory of the reader and distract it from understanding the literal meaning of the text.” See Nicholas of Lyra, “Literal Postill on the Bible: Extracts from the General and Special Prologues, and from the Commentary on the Psalter,” in Medieval Literary Theory and Criticism, c. 1100–c. 1375: The Commentary Tradition, edited by Alastair Minnis, A. B. Scott, and David Wallace (Oxford: Clarendon, 1992), 269–270.

  • 42. Sarah Fielding, The History of Charlotte Summers, the Fortunate Parish Girl, vol. 1 (London, 1750), 68. “Arabella Dimple,” as it happens, is a metacharacter in the novel, a reader of Charlotte Summers itself, and in fact has misremembered—the author had in fact bid the reader remember the temporary halt at the end of the novel’s fourth chapter; in thinking that she left off reading at the end of the sixth chapter, from within the novel’s fifth, Arabella Dimple jumps ahead in the story she occupies, part of the mid-18th century British novel’s love of addressing the artificiality of the chapter form.

  • 43. Laurence Sterne, The Life and Opinions of Tristram Shandy, Gentleman, vol. 4 (London, 1761), 97.

  • 44. Arriving at the figure of 146 days required surprisingly little in the way of hermeneutic deftness. My counting method employed one framing rule: to include only narrated action—not narratorial digressions, and not any specified amounts of elapsed unnarrated time (such as, for instance, the specification in chapter 31 of Middlemarch of Rosamond’s uneasiness after ten days of Lydgate’s absence).

  • 45. George Eliot, Middlemarch, edited by David Carroll (Oxford: Oxford University Press, 1988), 213, 317.

  • 46. Anthony Trollope, The Small House at Allington, edited by James Kincaid (Oxford: Oxford University Press, 1980), 257.

  • 47. Ibid., 260.

  • 48. Ibid., 241.