The Liber Genealogus

Notes

The Great Stemma carries neither a date nor an author's name, but we are fortunate to possess a sibling work, the Liber Genealogus, which provides both dates and a context. The Liber is generally regarded as an anonymous Latin work of Late Antique Christian scholarship which summarizes much of the genealogical material in the Bible, principally that part which Christians believed to constitute the ancestry of Christ, adding a small amount of chronological material from the time of Christ onwards for another four centuries. An edition of the work in its earliest form appears on this website. [*]Its siglum in Fischer's Verzeichnis der Sigel: Kirchenschriftsteller (since republished as Gryson, Roger. R�pertoire g�n�ral des auteurs eccl�siastiques latins de l'antiquit� et du haut moyen �ge (Freiburg: Herder, 2007)) is AN gen. Its number in the Clavis Patrum Latinorum is 2254.

At first glance, it appears to be a kind of handbook for bible study. One form of its text appeared in print in the 19th century under a different title, the Origo Humani Generis. It is generally recognized that both titles constitute variant editions of a single work written in or before 427 CE, perhaps in North Africa or in Gaul.[*]See my bibliography for the critical editions by Lagarde, Frick and Mommsen and for the two recent discussions of it (short titles Romains and Interpretatio) by Inglebert. Maureen Tilley in her The Bible in Christian North Africa calls it the Liber Genealogicus, but I have not seen any other writer using that spelling.

The work has a long history of publication in print. The T Recension, from the Codex Taurinensis, was edited and was published at Paris by Christoph Pfaff in 1712 (Google Books). Giovanni Domenico Mansi's 1761 edition offered readers the L Recension from Lucca (Google Books). In 1892, a German scholar, Carl Frick republished the Pfaff text with a learned introduction in Latin, entitling it the Origo Humani Generis, while another German, Paul de Lagarde, published his own fully annotated transcription of the Lucca and Turin texts (he called them M and C) together in a single print article.

In the same year, both were trumped by Theodor Mommsen, who collated both of the previous texts along with two other recensions (G and F), merging them into a unified critical text and issued it as part of his Monumenta Germaniae Historica series. Whereas Frick had misguidedly perceived the Origo, which is shorter, as a prior work, Mommsen considered it, with good reason, to be a later abridgement of the Liber. From that perspective he concluded that it was not entitled to separate consideration.

Mommsen was the grandest scholar of his generation but his decision to merge the four texts in this way was in some respects unfortunate. The unified text made it difficult for the reader to detect the earliest stratum, interpolated and contaminated as it was by later generations of editors. To overcome this difficulty, I have extracted the G Recension, the primitive text of 427, matched it against the St Gall manuscript, and re-published it integrally on this website.

In 1922, the French scholar Paul Monceaux reverted to a pre-Mommsen view, proposing like Frick that the Origo, which leaves out all the events after the birth of Christ while adding ethnographic material, was the earlier work. With the weakest of arguments, Monceaux dated the Origo to about 397. He contended that the Liber then grew from it, appearing in a series of recensions, initially within the secessionist North African church of the Donatists and then among Catholics, over the next 60 years.[*]Monceaux proposed that the Origo began as a Latin adaptation of Hippolytus's world chronicle and that it may have been composed in Gaul or Italy, followed by a now lost Donatist version composed in about 406. In essence, he argues (a) that the 438 version is not a child of the 427 version, but a sibling, and (b) that they are silent about an event of 411. Hence they must be based on a lost parent composed before 411 (Monceaux, P. Histoire littéraire de l'Afrique chrétienne depuis l'origines jusqu'a l'invasion arabe. Paris: 1901-23, vi. 250-251).

A recent examination, by Richard Rouse and Charles McNelis, has raised cogent questions about Monceaux's assumptions, suggesting that the Liber of 427 probably is the earliest appearance of the document. Another recent account by Hervé Inglebert has tended to support Monceaux's view, while leaving the question of the document's evolution open. Rouse, Richard, and Charles McNelis. 'North African literary activity: A Cyprian fragment, the stichometric lists and a Donatist compendium.' Revue d'histoire des textes 30 (2000): 189-238. Inglebert, Herv�. Les romains chr�tiens face � l'histoire de Rome. Collection des �tudes augustiniennes . S�rie Antiquit�s, ISSN 1158-7032. Paris: Inst. d'�tudes Augustiniennes, 1996. Neither the shortness of a recension, nor the fact that it is transmitted to us in the oldest available codex can be seriously used to argue that it is the most ancient recension. Inglebert also questions Monceaux's shaky argumentation for the 397/406 dates, but leaves the issue open.

If we bring the Great Stemma into the overall picture, as I outline below, we can scarcely avoid the conclusion that the Liber of 427 is indeed the earliest of the texts and that the Origo was inspired by the Liber, reworking the content and deliberately omitting the material for the period after Christ's birth. I would therefore agree with the finding by Rouse and McNelis, but for different reasons which I am about to explain.

As to the unsolved riddle of the authorship of the Liber, all we can say with any certainty is that the 427 author must have belonged to the Donatist church of North Africa, a schismatic Christian group, since the text criticizes Catholic persecution of Donatists.

Bruno Krusch has offered a weak argument, summarized by Monceaux, that a Gallic author, Q. Julius Hilarianus, might have been the author:

Quant à l'auteur, c'est sans doute Q. Julius Hilarianus, dont nous possédons deux petits traités écrits en 397, le De duratione mundi et le De ratione Paschae. La tradition manuscrite semble favorable à cette attribution. Le Taurinensis contient à la fois le De ratione Paschae d'Hilarianus et le Liber genealogus. En outre, le De ratione Paschae a été pillé par l'auteur du Liber Paschalis, de 455, qui, dans le Codex Lucensis, précède le Liber genealogus.[*]Monceaux, 250

However this argumentation is not convincing, and the broad consensus remains that the name of the Liber author is irredeemably lost.

The Liber presents the same two parallel ancestries as the Great Stemma, but in text rather than graphic form, accompanied by short analytical comments.

Its three principal editions are now denominated G, F and L from the manuscript groups that contain them. In Mommsen's combined edition, the Origo carries the siglum T: [*]The sigla were chosen by Mommsen and refer to the locations of the type manuscripts: St Gall, Florence, Lucca and Turin.

il faut distinguer G, l'édition donatiste de 427, F, l'édition donatiste de 438, et L, l'édition catholique de 455-463. Le Liber genealogus se présente avant tout comme une liste des généalogies bibliques du Christ. Il s'agit donc d'un résumé d'histoire biblique, qui utilise parfois des apocryphes comme III Esdras, dans lequel viennent s'intercaler des remarques différentes selon les éditions G, F, L. [*]Inglebert, Romains, 599.

Ayuso and Fischer would have been aware that the Great Stemma in the San Juan Bible contains eight etymologies that can only have been added to the diagram after an editor had consulted the Liber Genealogus. [*]Seth = resurrectio; Enos = obliviscens; Cainan = lamentatio (SJ) / natura dei (LG); Malelehel = plantatio; Iareth = dissensio (SJ) / descensio (LG); Enos = renovatio; Matusalam = missus; Lamec = percutiens (SJ) / bonae mentis (LG). This is likely to be the source of their understandable but mistaken assumption that the entire Stemma is simply a graphic adaption of the linear text that had previously existed in the form of the Liber Genealogus: their view has prevailed unchallenged for several decades as the conventional wisdom.[*]Ayuso, Extrabiblicos. Fischer, Algunas.

The reverse hypothesis, that the Liber author may have been guided in his research, data arrangement and writing by a stemmatic drawing has been a constant focus of my research and I have concluded that, on the balance of probabilities, some kind of diagram, probably the Great Stemma, did indeed serve as the basis for the Liber. The reasons for this finding, which is surprising in the light of all the previous scholarship, are set out below.

Because the documents are clearly linked, my collation of the Great Stemma text is cross-referenced throughout to the matching names and passages from the Liber Genealogus, employing Mommsen's section numbers, subdivided with my own subsection numbers.

We can readily see from this comparison that the two works cover the same ground and cover it in roughly the same fashion. They proceed from the extended genealogies laid out in Genesis, taking the reader onwards to the ancestries set out in the books of Ruth, Samuel, Kings and Chronicles. Both connect these to the genealogies claimed for Christ in the gospels of Matthew and Luke. Five points about their broad similarity will suffice:

• The overall data order in the two documents matches: the Liber presents the names in an order that resembles the left-to-right, top-to-bottom order of the Great Stemma: first setting out all the descendants of Jacob, then doubling back and repeating the names Judah, Perez and Hezron (Liber 360, 361) to explore the Davidic line in detail, then tracing Christ's ancestry according to Matthew, then doubling back to complete the line according to Luke.
• The Levites appear twice: the sons of Levi appear in the main stream leading to Christ, but they also appear again in a recapitulation. In the Liber, this appears after the genealogy terminates with Christ. In the codex form of the Great Stemma, the Levite appendix uses free space on a codex page which is otherwise only occupied by the ancestry chain from Esrom to Boaz (or Jesse). This supplement continues the Levite genealogy down to the children of Aaron and Moses. I have argued elsewhere that this passage was originally attached to a timeline.
• Genealogies of the children of king Saul, of Eli and of Samuel also appear. Like the Levites, these rulers are significant for the chronological underpinning of the content, but their family relationships do not form a part of Christ's genealogy.
• Two further supplementary genealogies are included in both works, although they have neither a bearing on Christ's ancestry nor on the chronography. These are an ethnic "map" of the Horites (Gen 36: 20-30) in stemmatic form and an account of the ancestry and friends of Job the Patient. But other biblical genealogies, for example the post-exile genealogies in the books of Ezra and Nehemiah, are not included.
• An approximate contemporaneity between the Great Stemma and the Liber is attested by the Vetus Latina forms of the biblical proper names, which were current at the time of writing, before the diffusion of Jerome's Vulgate translation— these names are generally Latin transliterations from the Septuagint Greek Old Testament.

We find, on close study, a multitude of fine-detail commonalities between the documents which include:

• The Joachimite hypothesis to explain Christ's ancestry.
• An array of uncanonical names, some of which are not attested in any other ancient work.
• The division of a single name, Manahat (Gen. 36:23), into two names, Macha and Cath.[*]First noted by Frick, 144, note 23: sunt potius quinque, sed ex uno (Manahat) duos (Macha et Cath) fecit. Fischer's Vetus Latina Genesis 378 suggests this arose as a dittography in the Greek, where a syllable was repeated like this, μαναχα χαθ, so that a single name separated into two. The error is unique to the Great Stemma and Liber Genealogus and could perhaps originate from a defective manuscript, or recension, of Genesis which is no longer extant..

Striking as they are, these similarities could perhaps be explained away by supposing that there might have been a common fund of biblical names, both scripturally attested and unattested, kept alive by oral tradition within some now unidentifiable community of Late Antique exegetes and scholars.

So much of the ancient documentary record has been lost that a 21st-century scholar must, in all honesty, allow for the possibility that these features do not mark the views of a single author or school, but were perhaps a form of received wisdom in an entire region, or a period, which we can no longer pinpoint.

This could not however be said for several similarities that indicate careful, word-for-word copying from models which must have constituted a distinctive book evolution. [*]Zaluska makes a similar point, but she cites three variations involving the omission of six names which I believe indicate not just a shared tradition but also an evolutionary fork within it: Les deux textes terminent chacun cette généalogie de façon apocryphe en introduisant à la fin Joachim, père de la Vierge, ce qui leur permet d'affirmer que la lignée est bien celle de Marie. En outre dans le parcours des générations, ils partagent trois variantes importantes qui entraînent l'omission de six noms par rapport à la liste "canonique". Ces analogies sont trop significatives pour être simplement le fait du hasard et doivent correspondre au moins à une tradition commune. Zaluska, Feuillets. Zaluska does not give the six names from the Lucan list, but apparently means (1) Melea, (2) Maath, Nagga and Esli, (3) Amos and Matthathia.

1. The Great Stemma shares with the Liber Genealogus a commentary which interprets the genealogies in the light of verse 5:5 of the New Testament Book of Revelation and asserts that Luke the evangelist's genealogy, from Nathan, in fact conducts to Mary, not Joseph. Both comment that these parallel biological descents are explained in Revelation with a lion symbolizing Solomon and a root symbolizing the obscure royal son Nathan.[*]Gen 49:9 associates the tribe of Judah with a lion cub; Isaiah 11:1, 10, echoed in Rom 15:12, establishes the phrase "root of Jesse". In translation, this exegesis says:

Whereas the evangelist Luke traces the origin of Mary back to Nathan, the evangelist Mathew traces that of Joseph back to Solomon, demonstrating an ancestry from the tribe of Judah. Thus it is clear that these two are biologically descended from a single tribe, leading down to Christ, so that what was written might be fulfilled, "Behold, the lion of the tribe of Judah, the root of David, has prevailed," whereby the lion is Solomon, the root is Nathan.

The two passages bear quoting in full in Latin. First the Great Stemma:

Sicut Lucas evangelista per Natan ad Mariam originem ducit, ita et Matheus euvangelista per Salomonem ad Ioseph originem demonstrabit id est ex tribu Iuda, ut apparet eos de una tribu exire et sic ad Christum secundum carnem pervenire. Ut compleatur quod scriptum est, "Ecce vicit leo de tribu Iuda radix David," leo ex Salomone et radix ex Natan.

As Ayuso and Zaluska note, the text in the Liber Genealogus is almost identical. It is printed as follows in Mommsen's edition of the Liber Genealogus:[*]Ayuso, Extrabiblicos, ...; Zaluska, Feuillets, 241.

Cuius Lucas evangelista ad Mariam originem demonstravit, similiter etiam et Salomonis Matheus evangelista ad Ioseph originem demonstrat, ut appareat eos de una tribu exire et sic ad Christum venire (pervenire), ut compleatur quod scriptum est: "Ecce vicit leo de tribu Iuda radix David," leo ex Salomone et radix ex Nathan.[*]simplified from Mommsen, no. 543.

While this text might perhaps be a quotation from a now-lost theological author in the patristic period, it is more likely that it is the central statement of purpose composed by the Great Stemma's author: the inclusion of the passage is deliberate and explains what the documents are meant to signify. Its association of the lion with Christ's male ancestry and the root with his female ancestry may also associate with some system of Antique symbolism. This would doubtless reward further attention. More detail.[*]While Eusebius of Caesarea is the source of the Stemma's chronographic framework, he preferred a different account of the genealogy contradiction.

2. A curious distortion of the data from the biblical book of Judges suggests both works share a connection to a distinct and irregular family of biblical manuscripts. This concerns the judge Shamgar at Judges 3:31. The New English Translation of the Vaticanus or B text of the Septuagint at this point reads:

And after him arose Samegar son of Dinach and he struck down the allophyles, fully six hundred men, with a bull's ploughshare. And indeed he too delivered Israel.The Vetus Latina translation is known from Augustine, who quotes it (Quaestionum in heptateuchum libri septem, Iudicum, 25): Et post eum surrexit Samegar filius Aneath, et percussit alienigenas in sexcentos viros, praeter vitulos boum: et salvavit Israel. This was duly compiled by Sabatier into his Bibliorum Sacrorum Latinae Versiones Antiquae but where the passage was located in the VL text that Augustine used is unknown. A complete Vetus Latina text, in the Codex Lugdunensis, has the passage at its normal position in Judges 3:31: Et post illum surrexit Semigar, filius Anath, et percussit ex alienigenis ad sescentos viros extra vetulos, et salvos fecit et ipse filios Istralel. This codex was rediscovered in the 1890s and published by Ulysse Robert.

There is nothing else. There is no mention of Shamgar/Samegar actually ruling, let alone for how long, but by a Jewish tradition of learning which is mentioned by Moore, Shamgar was counted as a judge and was reckoned to have ruled for a matter of months only, and therefore not to have earned a predicate of years.[*]Moore, Judges, 104 note: The midrash explains that he died in the first year of his office.

This conception was to continue for centuries. Even a much later work, the Chronicon Alexandrinorum, dated to about 630, reproduces the feature: Shamgar is found there as a judge (Semega iudicum) (line 236).[*]Mommsen, MGH AA, 9, p 117. See also: Albrecht, Stefan. "Chronicon Paschale" in Encyclopaedia of the Medieval Chronicle, ed. Dunphy, Graeme. Leiden: Brill, 2010, 387-388.

By the second century, this peculiar tradition undergoes two further changes: firstly, Shamgar is allowed a full year of his own in the historical calculations, and secondly, the Shamgar passage moves to a new place in the sequence. The latter transformation is evident in some of the Septuagint manuscripts of Judges: the sentence normally found at Judges 3:31 shifts to a position after Judges 16:31, with the name Shamgar altered to Samera.[*]The whole Shamgar issue is summarized in a note by Nestle, Eberhard. "Samgar". Zeitschrift f�r die alttestamentliche Wissenschaft, 32 (1912) 152. Nestle quotes the Liber Genealogus (calling it the vandalische Chronik and the Origo Humani Generis), and states that the position shift is characteristic of the Lucian family of Septuagint manuscripts, referring specifically to the Codex Zuqninensis rescriptus.

Why this happened we cannot say, but in the early Christian chronology compiled by Theophilus of Antioch for his Apologia ad Autolycum we read:

Then Samson judged them 20 years. Then there was peace among them for 40 years. Then Samera judged them one year; Eli, 20 years; Samuel, 12 years ...[*]This is Schaf's translation, now on the CCEL website, found in Book III, chapter XXIV.

Within 50 years of Theophilus writing, both Hippolytus of Rome and the scholar Julius Africanus have both recognized this One Year of Samera as a confirmed period of Jewish history. Hippolytus returned Shamgar to the early position, but we do not know at which position in the sequence Julius Africanus placed the year.[*]The position remains unresolved in Wallraff, Chronographiae, T40, and Gelzer's attempt at a solution (cited by Nestle) is not accepted.

The Great Stemma and Liber Genealogus then add yet another layer of oddity to this tradition by counting both Shamgar, the early (canonical) warrior, and Samera, his later alter ego, as separate rulers, each with his own place in the timeline.

Though later chronographers tended to realize that the second Year of Samera was an error, like many errors, it developed a life of its own. Gelzer's study of Julius Africanus mentions other works where this variant rears its head again centuries later, and of course, in the form of the Great Stemma, it was still being faithfully copied into new manuscripts in Spain in the high medieval period.

What is more, the Liber Genealogus inflates the first Shamgar's time period from one to 20 years (paragraph 505 of Mommsen's edition) while leaving the second One Year of Samera intact (paragraph 521). We do not know what time spans the Great Stemma author conceived for Shamgar and Samera, as this data has been deleted by a later editor, but the Liber is a guide.

The Beta recensions of the Great Stemma, which have been partly corrected using Jerome's Vulgate, disallow Samera (the latter) and only include Shamgar in his earlier position in the series.

3. The last of these peculiar commonalities concerns the supposed judgeship of Puah, the father of Judge Tola. Once again, the history of Tola is only the briefest of passages. Judges 10:1-2 has an unstable textual history, but it certainly never ascribes to the father Puah the character of a judge.

I quote the New English translation of the Septuagint.[*]The English translation is based on the so-called Alexandrinus or A version. As above, the notation B here refers to the Vaticanus version.

Thola son of Phoua, son of his father's brother, a man of Issachar, rose to deliver Israel, and he himself lived at Samaria (B: Samir) in the hill country of Ephraim. And he judged Israel twenty-three years. And he died and was buried at Samaria (B: Samir).

In the Liber Genealogus, the account is quite different:

Deinde Fua filius Charram iudicavit eos annis XX. Hic habitabat in monte Efrem, sub cuius tempus pax abundavit et non fuit bellum in Israel (Then Phua son of Karran judged them for 20 years. He lived in Mount Ephraim. During his time peace reigned and there was no war in Israel). Deinde Tole iudicavit eos annis XX duobus (Then Tola judged them for 22 years).[*]Ulysse Robert's Vetus Latina bible version is: Et surrexit post illum qui salvum faceret Istrahel Tholam, filius Ful, filius Charreon, fratris patris eius, vir Isachar, et hic habitabat in Samaria in monte Efrem (Robert, 132) A fragmentary Vetus Latina version without Charreon can be obtained from Augustine (Quaestionum in heptateuchum libri septem, Iudicum, 47) and is quoted by Sabatier (Bibliorum Sacrorum Latinae Versiones Antiquae): Et surrexit post Abimelech, qui salvum faceret Israel, Thola filius Phua, filius patris fratris ejus, vir Issachar ....

The likely source for the uncanonical name Charran/Charreon can be detected in several so-called miniscule manuscripts of the Septuagint which insert the name Karie into the Greek sequence, stating that Puah was the son of Karie.[*]The miniscule manscripts date from the 9th century onwards, but in some cases they retain very ancient readings which have been lost from all the older manuscripts at our disposal. Lagrange (Juges, 186) quotes the passage thus: χαι ανεστζσεν ο εος... τον Θωλα υιον Φουα υιον Καριε πατραδελφου αυτου. Lagrange says this reading is characteristic of a group denoted as M by Moore, who describes it (Commentary on Judges, xliv-xlv) as a group whose most constant members are four Greek codices.

The Great Stemma, in the form we know it today, makes no mention of Charran, but it does treat Puah and Tola as two successive judgeships, listing them, Fua iudex, Thola iudex. As noted above, the Great Stemma does not dwell on time periods or the parentage of any judge.

How are we to explain the elevation of Puah to the role of a judge? It seems this cannot be based on any corruption of the running scriptural text, since the passage is so short, and so simple, that any reformulation would have had to arise from a complex chain of mistakes to yield such a completely different sense. It is much more likely that the corruption would have happened on a page of brief notes where a reader has failed to correctly decode the positions and the significance of the different names as set down by a previous writer.

Puah might have inadvertently slipped from an upper register, of parents, to a lower register, of judges, if for example, the names had been noted thus:

                      Karie
                           Puah
       Gedeon    Abimelec      Tola     Iair
                             XXIII    XXII

With Karie at the very top, a reader might assume that the pair Karie-Puah had been bumped out of line and might inadvertently think that Puah was part of the main sequence in the bottom row. This is only a hypothesis, but it is consistent with the resulting errors. The transformation of Tola's number of years, properly 23, to 22 or 20 is not in itself odd, since alterations in Roman numerals are among the commonest scribal corruptions, but the evolution from one to two distinct numbers in the Liber Genealogus is strange and is difficult to explain. If the notation above did play a role, perhaps the 22 under Iair (Jair) has been understood as applying to Tola.

4. My tabulation of the genealogy of Luke shows four groups of omissions in common: (1) Kenan, (2) Melea, (3) Maath, Nagga and Esli, (4) Amos and Matthathia. This is a strong indication that the Great Stemma and the Liber share a common origin. It is conceivable that they may have relied on a defective Gospel text.

All of the above peculiarities suggest an affinity in spirit and conception between the two documents. But on the other hand, the two clearly have different purposes.

The Great Stemma offers a visualization of biblical genealogy and the biblical time-space, doubtless to assist in counting and calculating with this data. It would be false to suggest the Great Stemma is a memory aid, but it could well have had a use in exposition, demonstrating to a student or a class how two groups of people inhabited the earth before the Flood and how Christ's dual ancestry was believed to split at the sons of David.

The Liber Genealogus has a different intention: it is most likely to have begun as an ekphrasis, or commentary on the great Stemma, and then to have entered use as a handbook, enabling a student interested in the genealogy to find the many names in one place without having to leaf through many scattered books of scripture, and to find etymological and theological notes on each name.

Beyond this difference in apparent intention, the design of the works clearly differs. Let us start with the broader divergences:

• The Liber and the Origo contain extensive etymological speculation on the meaning of biblical names, none of which is reflected in the Great Stemma.[*].
• The Great Stemma has extensive chronological material from Genesis, which is only echoed in the most cursory fashion in the Liber, but on the other hand it lacks the Liber's precise calculation of each element of the Judges Period. Indeed it seems to omit most year data after Abraham.
• The Liber (but not the Origo) adds historical material for four centuries after the time of Christ, whereas the Great Stemma abruptly stops at Christ's birth.

The above differences all involve supplementary matter, and could therefore be readily explained as evolutions and reworkings of an original, more compact Great Stemma text.

But there are certain other divergences which can be detected which seem to indicate separate development by the two works:

• While the G recension of the Liber contains only the most rudimentary ethnic annotations, a Table of Nations has been introduced into the later L and T recensions of the Liber that appears to be largely based on a scheme that can be documented back to the Liber Generationis II and is thought to have originated in the mostly lost Chronicon of Hippolytus.[*]The various nations named are analysed in detail by Inglebert, Interpretatio, 160-2. The original Great Stemma also seems to have contained very little ethnological data. A Table of Nations scheme obtained almost entirely from the Etymologies of Isidore of Seville was introduced into the Stemma's Alpha and Beta recensions. However the ethnological data offered by Isidore (cribbed from Jerome, who obtained his data from Flavius Josephus) contradicts the theories of Hippolytus. In consequence the Alpha and Beta recensions of the Great Stemma contradict the Liber Table of Nations at half a dozen points.[*]Differences include: Magog said to be the father of the Scitus et Gotos in the Great Stemma but of the Celatae et Galatae in Liber Genealogus 176— Tubal: Iberi (GS) Hettali (LG 192)— Meshech: Cappadoces (GS) Defantes Inlyrii (LG 194)— Gomer: Galate (GS) Cappadoces (LG 170)— Aram: Siricorum metropolis Damascus (GS) Itei (LG 81)— Lud: Lidii (GS) Lazones (LG 79). All of these are found in the T recension of the Liber and two are found in F. It is notable that none are found in G..
• The Liber, from the earliest G recension, lists King David's 24 priests (1 Chronicles 24: 7-18), four bards— Heman, Asaph, Jeduthun (1 Chronicles 25: 1-4) and Ethan (1 Chronicles 15:19, also credited as the author of Psalm 89)— as well as the 24 champions of David (1 Chronicles 25: 9-31). These 52 names are not found in the Great Stemma.
• The Jerusalem sons of David are listed in the Great Stemma, but omitted from the Liber which only counts the Hebron and Bathsheba groups of David's sons.
• Nathan the Prophet has a prominent stemma roundel of his own, emphasizing his separate identity, but is omitted from the Liber.
• The Liber ignores the 12 Minor Prophets, whereas their periods of activity are noted along the Great Stemma's timeline, much in the fashion in which the Chronological Canons of Eusebius marks them with glosses.
• The two works are in conflict over their chronology of the Roman kings, but once again we cannot be entirely certain that the version we now see in the Great Stemma was part of the Stemma at its origin.

We must also not omit to mention that there is a very ancient gloss in the Great Stemma, probably based on a sermon by Origen, discussing the theological significance of Lot's incest. This is discussed in more detail on the theological page. The passage has no counterpart at all in the Liber. This difference must however be treated with caution, as one might question whether the gloss is an original part of the Great Stemma. It is omitted from the Epsilon recension and it does not even seem germane to the diagram's central purpose.

Indeed, in others of the cases of difference which I have mentioned above, one is left with a certain ambivalence. When were these distinctive features introduced to the Great Stemma? By its author? Or have they been seamlessly inserted by a later editor and are we falling into a trap by perceiving them as original?

The question is an important one because I wish to demonstrate that the Great Stemma is not a set of notes compiled from the Liber, but a predecessor and an autonomous creative work. One would be left in a state of doubt about this if it were not for certain fine differences which are particularly striking. They are found in the core material, the genealogy, and they involve the kind of obscure detail that would doubtless be of intense interest to an author, but would scarcely ever attract the concern of a revisor.

When one considers the array of micro differences set out below, one's conviction grows that the Great Stemma, the Liber Genealogus and the Origo Humani Generis were, despite their superficial resemblance and their doubtless dependence on common sources, produced by different authors.

Moreover, even if the works do appear to have emerged from a common tradition of learning and to quote some of the same authorities, their authors may even have consulted Vetus Latina or Septuagint scriptural texts which brought them into mutual conflict:

• The list of the sons of Joktan (Genesis 10:28-29) in the Liber is explicitly counted up 12 persons and ended, whereas the Great Stemma offers at least 13 sons, as in the Hebrew Genesis, with the inclusion of Obal.[*]Obal is now canonical, having been standardized as part of the western church's list by Jerome's Vulgate. The G�ttingen Septuagint's Genesis volume, 139, draws attention to the Great Stemma's inclusion of the 13th name, though the note text, "Lat Prol geneal in hispan mss," seems to have been transposed to the wrong place in the critical apparatus. The Göttingen edition excludes Obal from its standard Septuagint text, while noting the name's steady presence in a substantial minority of recensions, employing a wide variety of Greek spellings: γεβαλ, γηβαλ, γιβαλ, γεβελ, ιαβαλ, ιονβαλ, γεβαρ, εναλ. Fischer, Genesis, 140 lists several early authorities, including Augustine, which exclude Obal from the tally and only accept 12 sons of Joktan. The presence in the Stemma manuscripts of a 14th son, Iodore, suggests not so much a scriptural misquote as some confusion in the early editing process: this item is probably a scribal duplication of the name of another brother, Hadoram.[*]Aduram and Iodore are sufficiently different words that an editor would be hard pressed to recognize them as mere variants of οδορρα. Fischer's Vetus Latina Genesis takes the standard Old Latin evolution to be Odorrem, while quoting below it all of the following variants, to be found in evolutions of the Hippolytus Chronicon: Derra, Odor, Adorra, Oderba and Odorem. Iodore and the Great Stemma are not cited until the appendix. The G�ttingen Septuagint's Genesis has no Greek spellings that closely resemble Iodore, but the critical apparatus documents an Armenian form which it transliterates iodoram (139).
• Certain other names that are integral to the Septuagint are omitted from the Liber, yet they are worked into most versions of the Great Stemma: Bozrah (mother of Job), Abishag (companion of David) , Maacah (wife of David), Naama (wife of Solomon).
• Certain uncanonical names are only found in the Great Stemma and never in the Liber, among them: Anazra (mother of Noah in Alpha and Beta); Nicibat (wife of Terah); Oholah and Oholibah (the names of the harlots in Ezekiel 23) in the roles of daughters of Lot and Dinah, the wife of Job. More detail.
• The Liber contains certain distinctive mistakes not found in the Great Stemma, several of which are discussed below. For example, the G recension of the Liber splits the name Iembram in the Levitic genealogy into two, transmitting it as Iebda and Baux, probably through scribal error. Its addition of a second Masiel to the same group probably arose in the same way.
• The Great Stemma excludes two forefathers, Nathan and Heli, from Mary's ancestry in that part of Luke's genealogy which concatenates Melchi, Levi, Joseph.[*]The Beta text contains a curiously botched attempt at emendation where Heli is introduced, but Levi is omitted. This leads one to speculate either (a) that Heli (Luke 3.23) must have been absent from the Gospel text consulted by the Great Stemma author, or (b) that the Heli and Levi were excluded in reliance on the Letter to Aristides of Julius Africanus.[*]Christophe Guignard delivered an important paper at the 2011 Oxford Patristics Conference on why names are exluded from the Julius Africanus list. Publication is forthcoming The earliest, G recension of the Liber restores both Nathan and Heli, so that we have there: Melchi, Levi, Nathan, Heli, Joseph.[*]The text concludes: ... Heli genuit Ioseph, Ioseph genuit Ioachim, Ioachim genuit Mariam (Liber Genealogus, nos. 607-9) The Origo editor makes a most curious emendation, since we find this order in the Origo or T recension: Melchi, Levi, Heli, Joseph.Why this Nathan was initially excluded, whether through consulting a deficient gospel text or by scribal error in the transmission of T, I cannot say, but the matter is in any case not germane to the point I wish to make here.

And what are we to make of those rare instances where we can directly compare how the two authors quote entire sentences of their Vetus Latina text? Here for example are the two authors' paraphrases from Genesis 4:2-5, dealing with God's divergent welcome for Cain's and Abel's sacrifices. The Great Stemma version is at left. At centre is the G recension of the Liber Genealogus. At right is Fischer's Vetus Latina:[*]This is the passage which Ayuso failed to correctly quote from Lagarde, Septuaginta Studien II, 5.

Factus est Abel pastor ovium, Cain autem operarius terre. Et obtulit Abel de primogenita ovium et Cain de fructibus terre. Et respexit dominus in munera Abel, et super munera Cain non respexit.

Abel ... et factus est pastor ovium, Cain autem agricola. Et factus est post dies attulit Cain, de fructibus terrae hostiam deo, et Abel attulit ex primogenitu ovium suarum et ex adipe illarum. Et respexit deus in Abel et in muneribus eius, in Cain autem et in hostias eius non respexit.

Et factus est Abel pastor ovium, Cain autem operabatur terram. Et factum est post dies obtulit Cain de fructu terrae sacrificium domino. Et Abel obtulit et ipse de primitiis ovium suarum (et de adipus earum). Et respexit deus super Abel et super munera eius, super Cain autem et super sacrificia eius non respexit.

Neither author can be using the other's text here as his source. Each has gone back to his Latin text of Genesis to quote from it directly. It is striking that the Great Stemma abbreviates the passage, and preserves the Vetus Latina's literal term for tiller of the soil, operarius terrae, while the Liber leaves nothing out, but offers a freer paraphrase and changes the tiller of the soil into an agricola.[*]The Vulgate text is: Fuit autem Abel pastor ovium et Cain agricola. Factum est autem post multos dies ut offerret Cain de fructibus terrae munera Domino, Abel quoque obtulit de primogenitis gregis sui et de adipibus eorum. Et respexit Dominus ad Abel et ad munera eius, ad Cain vero et ad munera illius non respexit. The notes to Fischer's Genesis, I, 79 quote four uses of agricola in this position, all later than Jerome: Liber Genealogus, a Pseudo-Augustine, Quodvultdeus and Isidore.

It might, as we have said, be argued that even these complex differences arose through subsequent editing. After all, none of our witness manuscripts is closer than 500 years to the period of authorship of either work. But there are so many divergences in their underlying conceptions that one tends to believe that it was the respective original authors who chose to shape the data, each in his singular way. Each clearly had access to prior material that the other did not use. One or both authors may perhaps have even directly consulted and translated Greek works, though apart from the Chronological Canons of Eusebius, we cannot identify any Greek or intermediate Latin work that they may have drawn on.[*]Ayuso draws attention to this point, emphasizing that both the Liber and the Stemma are, at least indirectly, the fruits of Greek rather than of Latin learning.

We now come to the issue of interdependency. If there is any, it is hardly likely that the that the Liber Genealogus is the prior work, since the Great Stemma author would not have simply abandoned the Heli node.

It is also hard to see why the Great Stemma author would have relied solely on an analytical textual work as his source. An existing listing of the succession of judges, kings and prophets might well have helped to reduce the labour of checking the ancestry chain, but the author would almost certainly have required constant direct access to scripture as he worked through the rest of his material, obtaining vital data on the patriarchs to build the timeline and its synchronisms, adding comprehensive details about the minor prophets and discreetly repairing the defects in Matthew's genealogy.

The greater likelihood lies with the proposition that the Great Stemma or a more primitive predecessor was the prior work, brilliant in its own way but narrowly focussed, compiled directly from scripture, perhaps with added input from an epitome of Eusebius's Canons or some other chronological work, whereas the Liber Genealogus is the later, compound work, bringing a more literary flavour to the data, worked up from additional analytical documents, including a handbook of etymologies. Evidence for this is explored on a separate page.

[Note: The above article, which has remained unchanged since 2014, was drafted at an early stage in my research and treats the possible priority of the Great Stemma with a certain diffidence. The arguments which I advanced in 2011 for my Studia Patristica article are in my view so strong that I am now in no doubt whatever that the Liber Genealogus is a work posterior to the Great Stemma.]