The Great Stemma: a Late Antique diagram of Christ's descent from Adam
Jean-Baptiste Piggin, Hamburg
Introduction to Data Visualization
To anyone interestedin visualizations of abstract information, a diagram in Florenceat the Laurentian Library must surely be the richest and most spectacular artefact of the entire first millennium. Marking a milestone in the history of writing systems, it maps biblical history onto a great chart. Its rediscovery offers evidence of a previously unsuspected graphic-arts technique in patristic narrative writing.
The Latin document we are concerned with spans eight folios of an 11th-century codex, Plutei 20.54 (ff 38r-45v). At first sight it presents a genealogy, starting from Adam and ending with Jesus Christ, in a form that seems to resemble a modern family tree.[1]
The extended diagram, which until recently had been thought to date from no earlier than the 8th century, and perhaps to have been first drafted as late as the 10th century, contains about 540 biblical and historical names, mostly encased in roundels that are connected to their neighbours by lines. The persons named include:
- patriarchs in Genesis,
- kings of Judah and Israel,
- minor prophets,
- Roman emperors.
This genre, a large-scale data visualization, is extremely rare and perhaps unique in anyliterature of the first millennium, yet it would evince little surprise among modern readers precisely because information graphics are nowadays such a common feature in newspapers and textbooks.
Figure 1: snippets and lines
Data visualization is a feature of advanced writing systems. Snippets of text are made into closed entities, which can then be grouped on the page into sets or strung together with connecting lines.Different kinds of connectors represent relationships among the entities such as sequenceor hierarchy. The chick-bird-egg example in Fig. 1 is a simple but powerful way of describing the idea of a cycle.
Visualization is a practice that we acquire as we learn literacy, but are only dimly conscious of possessing. Once the notational conventions are understood, the boxes and lines can be elided: white space can be used to create virtual containers, and position alone can convey the relationships between the nodes.
Such drawings are often described as "back of a napkin" or "back of an envelope" in style, suggesting informality, but are also a foundation of PowerPoint presentations and are of interest in the field of information science, where the term mind mapping is sometimes applied to the visual conventions that the drawings are based on.
The road map, on which highways connect towns, is the implicit model, and perhaps was even the early inspiration for such drawings. Indeed, in information architecture, the connections between such data are called paths.
The most frequently used pattern in the paths between such texts is that of hierarchy, because visualization is more effective at communicating the intricacies of hierarchy than linear speech is.The principle of hierarchy brings visual order to things as various as corporate organization charts, cladograms in biology, flow diagrams for military planning and family trees. Formally, genealogies are no different from other hierarchies.
Any evidence that the pre-medieval world was familiar with such visualization techniques would not only shake a still widespread perception thatGraeco-Roman intellectual life was discourse-centric, but also tend to contradict the view that Antique technical draftsmanship and mapping skills were poor and impracticable.
Figure 2: document overview
The Florence Diagram
The Florence diagram makes sophisticated use of this mind mapping technique. This article will investigate the diagram's:
- text,
- visual logic,
- date of origin, and
- role in Christian controversy.
My re-assembly of the sections in the codex into a continuous chart appears to be the first presentation of the diagram as a more or less single unit for 1,000 years. I have also made some organizational alterations vis-a-vis the codex, based on evidence about how the document is most likely to have been spatially arranged in its archetypal form.
Fig.2 shows a reconstructionof the diagram, here arranged in two parts for a modicum of legibility. A register has been added to mark the stages of bible history that are represented, from the Creation to the Incarnation. The entire diagram was perhaps two metres wide when originally drawn.
Figure 3: Family of Esau detail
Fig.3 shows a detail from the manuscript:the family of Esau. A literal reading of Gen. 26, 28 and 36 indicates that Esau had fivewives and five sons. The connections that radiateaway from Esauhere are of three sorts. One marks his descent from Isaac. One is a trunkconnector to Esau's wives, indicating that all five share a common relationship with him.Five are ramifications, from the Latin term, ramus, for branch, to his offspring. Together, the scheme constitutes a stemma, roughly equivalent to what by the Renaissance period had come to be called a family tree.[2]
Persius and Isidore indicate to us that the termsramusand stemma were an established pair in Latin usage:
Persius: An deceat pulmonem rumpere uentis stemmate quod Tusco ramum millesime ducis censoremue tuum uel quod trabeate salutas? (Saturae 3.28)
Isidore: Stemmata dicuntur ramusculi, quos advocati faciunt in genere, cum gradus cognationum partiuntur ... (Orig. IX.vi.28)
A readable arrangement of the diagram, in a single strip where zooming in reveals the fine detail, can be found on the website. The text and roundels were retraced using vector-drawing software, which offers the added benefit of separating the content into searchable layers, which can made visible or invisible at will so as to facilitate study and discussion. Fig. 4shows a schematic redrawing of the family of Esau.[3]
Because the Florence diagram is an assembly of many such stemmata into a kind of super-stemma, I call it the Great Stemma for short.
Figure 4: Esau family redrawn
No author's name can be discovered for the Great Stemma. Moreover, it carries no title. A heading, Genealogia ab Adam usque ad Christum per ordines linearum, appears on at least two medieval Spanish copies, but this appears to be merely a librarians' descriptive label. The Great Stemma bears no date of authorship, although the Florence copy does have a dated text attached to its final folio. This is the Ordo Annorum Mundi, a document to be discussed in more detail below, in a recension of 672 A.D. That date can be adopted as aterminus ante quembeforewe seek evidence of an earlier date.[4]
It is astonishing that the Great Stemma has received such limited scholarly attention. Two Vetus Latina scholars, Bonifatius Fischer and Teófilo Ayuso Marazuela,quoted extensively from the document in the course of their research on the early western bible.[5]
Several art historians, including Wilhelm Neuss, John Williams and Yolanta Zaluska, have studied how medieval Spain decorated it.[6]
Most recently, the cultural historian Christiane Klapisch-Zuber took a closer look at a late copy, from Gascony, but rapidly dismissed it as an affront to the principles of ‘graphic semiology’, and argued that no coherent biblical genealogical diagram existed before a medieval work, the Compendium, devised by Peter of Poitiers.[7]
Specialists from other fields have yet to begin picking over the rich fruit that the Great Stemma contains. Until I published an online collation in 2010, there had not, to my knowledge, been any complete edition of the Great Stemma text, let alone a graphic rendering.
Zaluska laid the foundations for this work more than 25 years ago by defining recensions and giving them five principal sigla: Alpha, Beta, Gamma, Delta and Sigma. She based her analysis on copies that are found in a cluster of Spanish bibles and as frontispieces to the Apocalypse commentary by the Spanish monk Beatus of Liébana.
Unfortunately she initially left out of account the Florence manuscript, which contains the earliest text, without illumination and largely free of medieval contaminations.
Zaluska did notice her omission, and very late in the day drew her readers' attention to the Florence manuscript, but she has not published on the topic since.[8]
I have denoted the Florence diagram with a new siglum, Epsilon, and it has become the basis for my electronic edition of the Great Stemma and my effort to reconstitute the lost archetype of the work.
The 18 manuscripts are listed below with their common long names, Zaluska’s adaptations of the Neusssigla, archival locations and finally the recension to which they belong. They are tabulated by their codex type (the Commentary of Beatus, a bible or a history) and date, with the Morgan codex, penned in about 940, believed to be the oldest extant manuscript.
Table: The Manuscripts
Beatus
1 / Morgan / M / New York, Pierpont Morgan Library / M. 644, 4v-9v / β2 / Tábara / T / Madrid, Archivo Histórico Nacional / cód. 1097B, 0-1v / α
3 / Gerona / G / Girona, Museu de Catedral / Num. Inv. 7(11), 8v-15r / α
4 / Urgell / U / Seu de Urgel, Museu Diocesán, Archivo de la Catedral / Inv. 501, I-V / γ
5 / Facundus / J / Madrid, Biblioteca Nacional de España / MS Vitrina 14-2, 10v-17r / β
6 / Saint-Sever / S / Paris, Bibliothèque National / ms. lat. 8878 / σ
7 / Turin / Tu / Turin, Biblioteca Nazionale Universitaria di Torino / lat. 93 / Sgn. I.II.1, 8v-15r / α
8 / Rylands / R / Manchester, JohnRylandsUniversity Library / ms. lat. 8, 6v-13r / α
9 / Cardeña / Pc / Madrid, Museo Arqueológico Nacional, and New York, MetropolitanMuseum / ms. 2, 3r-6v, and 1991.232.6, 2-3r / α
10 / Las Huelgas / H / New York, Pierpont Morgan Library / M. 429, 6v-12r / α
Bible or Beatus?
11 / Fragment Vitr. 14-2 / Fi / Madrid, Biblioteca Nacional de España / MS Vitrina 14-2, 2r-5v / βBibles
12 / León Bible / Le / León, Colegiata de San Isidoro / cód. 2 / β13 / San Juan de la Peña Bible / Ma / Madrid, Biblioteca Nacional de España / lat. 2 (A. 2) / γ
14 / Second León Bible / León, Colegiata de San Isidoro / cód. I. 3. / β
15 / Calahorra Bible / Ca / Calahorra, Cathedral Treasury / ms. 2 / δ
16 / San Millán de la Cogolla Bible / Ac / Madrid, Real Academia de la Historia / cód. 2-3 / δ
Histories
17 / Roda Codex / Ro / Madrid, Real Academia de la Historia / cód. 78 / α18 / unnamed / Florence, Biblioteca Medicea Laurenziana / Plut. 20.54 / ε
Writing the first detailed study of the work, Zaluskabuilt on Fischer's survey of the Vetus Latina word-forms for hisGenesisto demonstrate that the text of the biblical names hadsubmitted to no influence from the new transliterations of the Hebrew names undertaken by Jerome of Stridon for his Vulgate translation at the end of the 4th century. Among the examples she isolates is Chor (Gen. 36:22), a personal name in the Septuagint and Vetus Latina, which Jerome had corrected to Hori. Zaluska also detectedthe Vetus Latina nameof a Horrite chief –Ucan– which Jerome had suppressed, as well as another name, Chat, which had arisen through scribal error. Both names continued to be reproduced in the Great Stemma into the high medieval period although they were no longer current in the Genesis narrative.
The original core text seems to have been written without any awareness of the writings of Jerome, or of his contemporary Augustine of Hippo, although all the recensions other than Epsilon have been augmented with many annotations from both of those writers, as well as from Isidore of Seville.
Perhaps it wasthese augmentationswhich caused Fischer to labour under the misconception that a late 8th-century Iberianauthor had drafted the diagram while consultingan already ancient Vetus Latina bible, thus creating an impression that this wasa clumsy and failed Dark Ages attempt to invent the family tree.
My hypothesis is that the recensionsare in fact the damaged remnants of one of the boldest achievements of Antique book design, dating from that period centuries earlierwhen the roll had yet to give way to the codex.
There can be little doubt that the diagram was originally conceived as a single piece, since a major copying error plainly occurred while the document was still in one piece: thirteen forefathers in the Matthew genealogy, from Uzziah (here Oziam) to Azor, have been attached to the wrong wives. It seems that instead of copying each parental couple in sequence, an earlyscribe copied the sequence of husbands en bloc, then the sequence of wives, only to mismatch them. One major block of roundelswas thus accidentally transposed by four positions to the right, disrupting other data throughout the area. This error crosses the page boundaries of the Epsilon and other recensions and infects all the known manuscripts.
The visual logic of the diagram, with lines that run the full sweep from beginning to end, indicates that itsreader needed to see itas a whole, unrolled on a table or hung on a wall, to grasp its layering of various data types into a harmonious whole.
Figure 5: Timeline
The timeline
The first of those layers of content is a chronicle of world history. This comprises the names of many biblical rulers, which are also enclosed in roundels like the genealogies. Thesedo not ramify, but are instead strung out like beads on a string. This sequence represents historical time, a long scale that probably began with the Flood (though there are now gaps and the start is no longer extant) and ended with the reign of the Roman emperor Tiberius, which is the secular time marker corresponding to the Incarnation.A possible arrangement of the material is shown in Fig. 5 in a schematic sketch that somewhat compresses the diagram’s horizontal sweep. The timeline clearly provides a backbone to the document and aims to persuade the reader that the genealogies to be arranged around it are not myth, but embedded in world history.
For three principal sections, from Kohath to Joshua, from the Samarian king Jeroboam to his distant successor Zechariah, and from Cyrus of Persia to Tiberius, the sequence proceeds, roundel by roundel, from left to right. This must have been the rule for the whole. One sequence, from King Cushan-rishathaim to Jephthah,judge of Israel, has been arranged in the main extant manuscripts as a string of roundels running down the page and then turning away to the left, doubtless to accommodate the exigencies of transferring the material from an expansive roll to the more limited room on a codex page.
There is no direct evidence that the roundelsmaking up the Great Stemma timeline were arranged to any fine scale of years, and indeed this would have been difficult to achieve where every roundel remained the same size, regardless of the length of reigns that varied between half a year and more than 40 years in length. However a coarser gradation in the order of quarter-centuries would have beenentirely feasible, and I have experimentally applied this in the reconstruction without wishing to insist that it was in fact the Great Stemma author's intention to do so.
In general, the chronological data does not surpass the available space when it is arranged in a band, roughly matching the genealogical datapoint for point. The data only had to be compacted in one stretch, the succession of Samarian kings. My reconstruction employs a zigzagging line (visible in Fig. 2 above the word ‘Israel’) as a provisional solution to accommodate this group. In every other section, the timeline fitted comfortably into the available space as a straightline.
The main consequence of drawing this timeline to scale was to make the area at the right of the diagram less dense, leaving a large amount of vacant space. It is conceivable that in the archetype, this space may have been occupied by a survey of the main 52 persons making up the Davidian royal court. I return to this issue below, although the evidence for such an inclusion is not conclusive.
Figure 6: The Fila
The Fila
Layered onto the timeline comes the document's dynamic factor, its four "ancestral streams". Fig. 6 shows these in the schematic view.The streams are the four arrows which sweep around the outer edge of the document, framing it, with Adam, David and Christ as nodal points.
In their purpose, these streams are analogous to the fila that measure the onward march of dynasties in the Chronological Canons of Eusebius of Caesarea. The regnal years of the various kingdoms there form filum sequences which run in parallel down each page of the Canons, providing a comparison between the different national time-reckoning systems.
But unlike the Eusebian fila, which are columns of year-numbers rigidly arranged on a vertical axis and in perfect alignment, the Great Stemma's fila sweep and turn to represent four main lines of descent in the bible story. Each of these lines has a significance of its own in salvation theology.
The Afilum is the main line comprising God's chosen in the Hebrew scriptures, the succession of patriarchs from Adam's son Seth down to David.
The Bfilumcomprises the antediluvian,accursedchildren of Cain. This graphic divergence of the A and B fila alludes to an early Christian topos, the contrary paths adopted by the children of Seth and of Cain, which Augustine was to elaborate in the 15th book of his City of God.
The C filum is principally made up of the house of Judah, the ‘royal’ descent from King Solomon down to Christ in the Gospel of Matthew.
The D filum comprises the succession of obscure post-Davidic names brought to us by thegenealogy in the Gospel of Luke, commencing with Nathan, who is described as a son of King David.