This essay takes up an unusual methodological problem in editing a poetic corpus. A group of ten poems by John Scottus Eriugena is supported by two witnesses of roughly the same date. One (siglum R) contains the Latin texts of ten poems, which are complete (with one exception) plus glosses to the numerous graeca in the poems. The other (siglum L) consists entirely of the Greek lemmata and glosses for all ten poems transmitted in R, a selection of lines (with their glosses) composed entirely in Greek that are transmitted by R, plus the texts of seven additional poems written entirely in Greek with their Latin glosses; most of these have inscriptions attributing the poems to John Scottus. The Greek lemmata and glosses in R closely match those in L, allowing for transliteration by R. One suspects that at least poems 1–10 bear witness to a common exemplar. Yet how can one be sure that R did not simply copy its glosses from L, or L take its glosses from R? And how can one be sure that the Greek poems in L are not just a selection of verses that L thought that it would be nice to add?

In two editions of the poems (in which the poems have identical numbering), I put forward the hypothesis that all seventeen poems descend from a common exemplar.Footnote ¹ However, the poems also constitute an anthology compiled by the author or a friend, as first argued by Ludwig Traube, who edited the poems in the Poetae Latini Aevi Carolini series of the Monumenta Germaniae Historica. Footnote ² That the collection of poems was also compiled in dating order was shown by Paul Dutton and augmented in my Scriptores Latini Hiberniae edition.Footnote ³ The poems in that collection, with a few exceptions, concentrate on the activities and the aspirations of Charles the Bald and his wife Irmintrude, beginning with Charles’s war against his brother Louis in 858 and ending with his coronation at Metz in 869, the same year in which Irmintrude died. John was Charles’s chief eulogist, indeed, “the royal poet,” as Paul Dutton called him.Footnote ⁴ John wrote a number of other poems, some of them epigrams, others dedicatory verses to his own writings or translations, but these require a grouping separate from the anthology. I accommodated these divisions in both editions: (1) John’s poems attested as genuine by inscriptions (1–19) and connected by transmission; (2) other genuine poems attested by inscriptions, written in the hand of “i¹” (Eriugena himself), or written as poetic prefaces to genuine works by John; and (3) poems that had some claim to be considered his work, but did not meet the criteria of sections 1 or 2.Footnote ⁵

Two reviews of the CCCM edition by respected scholars challenged its editorial principles and the tripartite division of the poems.Footnote ⁶ The main thrust of the reviews is directed against the hypothesis of an authorial anthology and the existence of an archetype. Another point of discussion concerned the unity of one of the two principal witnesses, siglum R, which was copied into two different manuscripts. Both reviewers appear to agree that the transmission of Eriugena’s poems was a somewhat serendipitous affair. For example, Professor Contreni concludes his review of the Carmina portion of his review article: “The few observations made here suggest that Eriugena’s poems were disseminated in multiple ways in the ninth century and call into question the existence of a putative Ur-collection such as Ω put together by the author.”Footnote ⁷ Professor Chiesa’s conclusion is similar: “… gli unici elementi communi a più di una parte sono sostanzialmente alcuni vocabuli greci, per lo più estrapolati. In una situazione così vaga, è difficile farsi un’opinione sulla consistenza di Ω, che rischia di essere una chimera.”Footnote ⁸ Neither reviewer, however, comes to grips with the methodological issue raised by the recensio codicum, nor addresses the fact that glosses constitute a vital part of the text tradition and must be accounted for in the recensio.

I would not have responded to the reviews (which make some helpful points) were it not for the question of methodology raised, namely, the challenge to the hypothesis that John compiled or approved an anthology of his poems addressed to Charles the Bald composed in the years 859–869. I contend that an archetype for poems 1–17 can be established which presents the poems with authorial glosses in dating order. I shall address this question first by demonstrating the existence of an archetype based on the close agreement of the two witnesses on the glosses to the graeca and whole lines written in Greek and a list of errores communes in the Latin poems 1–10, followed by a demonstration that neither witness copied from the other. Next, I shall show that the glosses to the poems were not compiled by a scribe, but by John himself. Finally, I shall discuss the dating order of the poems.

The poems of group 1 are preserved in two principal witnesses: siglum R, which is divided between Vatican City, Bibliotheca Apostolica Vaticana, Reg. Lat. 1587 and Reg. Lat., 1709, both believed to have been written by the same hand in western France (Fleury?) in the third quarter of the ninth century;Footnote ⁹ and siglum L, the famous Laon, Bibliothèque municipale, MS 444, a Greek dictionary and glossary copied by Martin of Laon before 875 (the year of Martin’s death).Footnote ¹⁰ Fols. 294v–98r of L contain Greek lemmata drawn from John’s poems (including all the poems in R) with Latin glosses; lines and couplets in Greek with their glosses that also occur in the poems transmitted in R; and short poems and prayers written in Greek, also attributed to John with the exception of poem 16. These too contain Latin glosses, some of which are arranged as Latin verses.

Following Traube, I have divided L into L1, the list of glossed Greek words that correspond to the Greek words transmitted in the order in which they occur in R, and L2, which contains whole lines in Greek corresponding to those in R, plus seven poems and prayers written entirely in Greek. Together L (L1 and L2) and R account for all the genuine poems numbered 1–17 (plus 3 fragments) in both editions.Footnote ¹¹ These were all composed between 859 and the end of 869 and constitute what, in my opinion, was an anthology compiled or overseen by John. The two epigrams that are printed in part 1 (nos. 18 and 19) are certified by an inscription and have a remote stemmatic connection to siglum L, but do not belong to the anthology. The poems in sections 2 and 3 are transmitted independently of poems 1–17 (which I call the anthology).

It is highly probable that the Greek lemmata and glosses in Latin poems 1–10 in R and L1 were copied from a common exemplar that contained a few corrections. L1 consistently presents lemmata in Greek characters, while R often transliterates them, albeit accurately. R very occasionally omits glosses found in L1 and adds a few of his own (from another source?), but the vast majority of lemmata and glosses are identical, as will be shown by viewing the complete list of Greek lemmata and glosses provided in the appendix at the end of this article. A careful look at R’s lemmata will reveal that R had an alphabetical knowledge of Greek. His transliterations of the Greek are nearly always correct and occasionally show a knowledge of Greek morphology. For example, the very last entry (10.19) where R renders ΘΙΑΣΟΤΟΝ as thiasotum shows that he recognized the ending -ON (for ΩΝ) as a genitive plural. Crucially, however, R did not copy the Greek lemmata and their glosses from L1. At 2.13 R has the correct ΠΝΥΞ, while L1 gives the meaningless ΤΠΝΥΞ. At 3.25, where the name Isis is used as a metonymy for Egypt, both manuscripts give ΙΣΙΔΑΜ, a mistake for the wanted accusative ΙΣΙΝ. R, however, adds the marginal gloss ΙΣΙΔΑ, a “learned error” based on the supposed genitive — ἰδος. Unless R had advanced to the knowledge of third-declension dental-stem nouns, we may infer that he got his marginal gloss from a correcting hand in his exemplar. Another possible appearance of a correcting hand in the exemplar is found at 8.21. There we have a division between ΑΚΡΟΙ R and ΑΚΡΟΣ L1, glossed summi and summus, respectively. R’s readings are the correct ones, but they are right because he chose the reading of the correcting hand rather than the original. At 3.45 L1 has ΔΑΛΜΑΤΑ, while R has id almata, which is closer to the archetypal ΙΔΑΛΜΑΤΑ, glossed imagines (R L1).Footnote ¹² At 8.34, where both R and L1 read ΟΝΤΑ, R’s gloss is qui est quae sunt, but L1 gives only quae sunt. Footnote ¹³ In the L2 section, we find the full gloss for ONTA, suggesting that L2 returned to the glosses in the common exemplar. Given L’s shaky knowledge of Greek (see below) and the fact that he did not consult R (as demonstrated below), L could not have made this correction independently. A rare omission of a gloss by R occurs at 2.71 ΜΕΛΠΟ, glossed canto in L1. The line belongs to a subscription to the poem in which the author complains of his patron’s parsimony.Footnote ¹⁴ However, the presence of the entire subscription in R and the gloss to ΜΕΛΠΟ only in L1 argue that both manuscripts were copied from an exemplar that contained the authorial subscription with a gloss. Where, then, did L1 find his gloss, since it is not in R?

Further proof that R and L1 were copied from a common exemplar is given by errores communes that occur in the graeca:

These errors can be explained in one of two ways: either both witnesses are copied from a common exemplar, or else one of them is copied from the other. Let us begin with the second possibility. It is impossible that L was copied from R, or to put it more exactly, that L copied its lemmata and glosses from R. In the list of Greek lemmata (L1) many Greek words occur that have no match in R. The omitted words constitute what I label Fragments 1–3, which represent missing poems, and two of the poems preserved in L2 (12a–b and 13). L1 could not have got his glosses from R because the poems whose existence is indicated by Fragments 1–3 are not given in R. Therefore, L1 cannot have been copied from R. R cannot have been copied directly from L1 for the obvious reason that the Latin texts are wanting. But there is a possible tertium quid. Could R have got his Latin text from an unknown source — unless we are to suppose that he wrote the poems himself and claimed Eriugena’s authorship — and taken the glosses from L1? Possible, of course, but then where did L1 get its list of lemmata with glosses, since he did not get them from R? The most logical explanation, and the most economical, is that R and L were copied from a common source that had both text and glosses, that is, Ω. I surmise that there was a correcting hand in Ω from which R entered corrections, and which L1 first ignored, then implemented in L2. I conclude that there existed a common exemplar that explains both the near unity of the glosses and the mistakes common to L and R in the Latin poems 1–10. The same common exemplar was arguably written in the author’s lifetime since the date of L can be reckoned between 869 (date of the last poem) and 875 (death of the scribe Martin of Laon).Footnote ¹⁵

However, we must still account for the absence of the Greek poems 11–17 in R and the poems represented by the glosses preserved by L1 for Fragments 1–3 in L1. R also omits two short Greek poems, the texts of which are preserved in L2, with lemmata and glosses preserved in L1. The following selection of entries in L1 displays the omissions by R in more detail:

As noted, R preserves phrases, whole lines, and couplets in Greek that occur in the Latin poems. It does not, however, preserve any freestanding poem written entirely in Greek. That applies to the poems mentioned above (12a–b and 13, both with graeca in L1), Fragments 1–3, and the Greek poems recorded in L2. Is it reasonable, then, to infer that R also omitted the poems represented by Fragments 1–3 (which occur within the group of Latin poems 1–10) because they were written entirely in Greek? The answer is no. While in one case (poem 12a), it is possible to restore missing lines from Greek entries in L1 that, when read in sequence, construe as verses, this does not hold for Fragments 1–3. Some themes do indeed emerge, but whole verses cannot be extracted from them. Thus, Fragments 1–3 were in all likelihood Latin poems that contained graeca. The total number of entries for the graeca recorded by L1 for Fragments 1–2 amounts to twenty-three and twenty, respectively, suggesting that the poems were of comparable length to some longer Latin poems preserved in R. Why, then, did R omit them?

It will be noted that, with the exception of Fragment 3, the omitted material is clumped together at the end of Latin poem 8, the last poem recorded in the Vat. Reg. lat. 1587 portion of R. Fragment 3, which has but six entries for the graeca, is also missing from the group, but given that it is the last item in L1, it was probably meant to follow Latin poem 10. The scribe writing in Vat. Reg. lat. 1587 left off his copying of the poems at the end of Latin poem 8, resuming his work in Reg. lat. 1709 by transcribing Latin poems 9 and 10. He may have decided to exclude Greek poems 12a–b and 13 on the grounds that they were short and written entirely in Greek, but why did he omit the apparently substantial Latin poems represented by Fragments 1 and 2? The list of Greek words recorded for Fragments 1 and 2 are rich in Eriugenian philosophical terminology (ΟΥCΙΑ, ΘΕΟCΙC) and theological allusion (CΑΡΞ, ΛΟΓΟC). It also mentions the king (ΑΝΑΞ), strongly suggesting that Charles the Bald figures in the work.Footnote ¹⁷ The first two entries of Fragment 2 give the words ΘΕΟC and ΥΠΕΡΑΛΗΘΗC (glossed superuerus), “above or beyond reality.” On fol. 298v of Laon 444, which follows poem 17, Martin wrote: “Vt Dionisius Ariopagita dicit: Deus ΥΠΕΡΘΕΟϹ est, id est superdeus et ΥΠΕΡΑΓΑΘΟϹ, id est superbonus, et ΥΠΕΡΑΛΗΘΗϹ est, id est superuerus.” Clearly the lost poem is a testimony to the influence of Pseudo-Dionysius on Eriugena’s thought.Footnote ¹⁸ The omission of the text of both poems is a serious loss.

The transcription of Latin poems 1–8 in Vat. Reg. lat. 1587 ends on folio 64v. This is the end of the quire. The next page recto-verso is blank except for the inscription on the recto in a humanist hand Cato maior de senectute, and on the verso in a Carolingian hand Cato de senectute cum Macrobio. The scribe had come to the end of his allotted space for the poems. When the same scribe looked for another book in the same scriptorium that offered space to complete the poems, he found it in a booklet of what became Vat. Reg. lat. 1709 between fol. 16v and fol. 18r. He skipped over the short Greek poems 12a–b and 13 and the longer Latin poems represented by Fragments 1 and 2, and entered poems 9 and 10 (a poem of only twenty lines). Poem 10 does not seem to be complete, even though there was space for it on fol. 18r. On fol. 18v, the same scribe, or a scribe with a very similar hand, commenced writing the Augustinian treatise Quomodo Sancta Trinitas et sempiterna atque incommutabilis facillime possit intellegi. Footnote ¹⁹ The scribe of the poems apparently faced the choice of what to include and was reduced to counting lines of works that would fit the space. It is important to note that poems 9 and 10 continue the dating order of poems 1–8. However, if there was not enough space to transcribe potentially weighty Latin poems such as we find represented by Fragments 1–2, it is no wonder indeed that the short Greek poems transmitted only by L2 were omitted.

At this point, we can conclude that poems with their glosses 1–10 in R and L descend from a common exemplar, and, though R alone preserves the Latin texts of the poems, L is a more reliable guide to the contents of the poems dedicated to Charles the Bald and record his actions in the years 858–869. Indeed, the scribe Martin was at pains to distinguish between the Graeca ad uersus on fol. 294v, col. 1, and the Graeca que est [sic] in uersibus Iohannis Scotti (“Greek words which are in the verses of John the Irishman”), which is the starting point of the graeca with glosses of poems 1–10 (L1), the graeca with glosses for Fragments 1–3 (L1), and the transcriptions of poems 11–17 (L2). The graeca ad uersus, a list of 36 Greek words with their glosses, belong to a different group of poems from that of the Graeca que est in uersibus Iohannis Scotti. However, the selected words under the former rubric give the telltale signs of being poems written by John. For example, in col. 1, we find a list of Greek headwords in this order preceding ΒΑϹΙΛΕΥΣ (“king”): ΘΑΥΜΑΣΤΟϹ (“wondrous”), ΕΞΑΡΧΗϹ (as two words, “from the beginning”), ΕΠΙΛΕΚΤΟϹ (for ἐκλεκτός, “elect,” “chosen”?), ΔΙΚΑΙΟϹ (“just”), ΑΓΑΘΟϹ (“good”), ΜΕΓΑΛΟϹ (“great”), ΤΙΜΙΟϹ (“honored”), ΦΡΟΝΙΜΟϹ (“prudent”), and ΑΚΡΟϹ (“exalted”). Some of these descriptors are attested in poems 11–17. There, we find ΜΕΓΑΛΟϹ (12a), ΦΡΟΝΙΜΟϹ (12b), ΕΚΛΕΚΤΟϹ (14.2), ΘΑΥΜΑΣΤΟϹ (17), ΑΓΑΘΟϹ (17), ΑΚΡΟϹ (17), and ΔΙΚΑΙΟϹ (Fragment 2.12). In col. 2 of the Graeca ad uersus we find ΚΛΕΟΡ, a mistake for κλέος (“fame”). This mistake occurs in 14.1 and 15.1, and seems to be a shibboleth of Eriugenian composition.Footnote ²⁰ I think it likely that the entries under Graeca ad uersus point to the existence of other poems written by John that were not intended for the group of poems 1–17. But even if these uersus were not written by John, they point to Martin’s concern for preserving intact only those poems intended for that collection.

We now turn to Greek poems 11–17, preserved only in L2. The notion expressed by one reviewer that Martin collected a miscellany of Greek poems for teaching purposes and not for producing an edition will not bear scrutiny.Footnote ²¹ They are prefaced on fol. 297r by the inscription ΣΤΙΧΟΙ ΙΩΑΝΝΙΣ GLORIOSO REGI ΚΑΡΟΛΟ (“Verses of John to glorious King Charles”), which indicates continuity with the poems preserved in RL¹. However, it remains to examine the question of the archetypal character of the poems preserved in L2. They occur in this order:

Subscription in the hand of Martin of Laon: ΣΤΙΧΟΣ ΠΡΕΠΟΣ ΔΙΔΑΣΚΑΛΟΥ ΜΑΡΤΙΝΟΥ

New inscription: ΤΩ ΚΥΡΡΙΩ ΚΑΡΟΛΩ ΙΩΑΝΝΗΙΣ ΧΑΙΡΕΙΝ

While problems of authenticity affect two cases (poems 15 and 16), the archetypal character of L2 remains to be discussed. Was it simply a grab bag of random poems carelessly put together by Martin for use in the classroom or the continuation of an authorial anthology? I did not reprint in either edition (nor did Traube in his) the couplets in Greek drawn from the long Latin poems that Martin inserted at the head of L2, as their readings and glosses were accounted for (with corrections) in the text and apparatus of Latin poems 1–10. These are two- and three-line Greek verses drawn from poems 2.67–68, 5.49–50, 8.20–22, 8.34–35, and 8.85–86 (numbering according to CCCM edition). Since the Greek lines and their glosses are also attested by R (Vat. Reg. lat. 1507), they are undeniably archetypal.Footnote ²² It should be fair to assume that poems 11–14 are authentic and archetypal, as there is no change of inscription until the “Martin inscription” before poem 16. Poem 11 is authenticated as a poem of John’s by the inscription ΤΟΙ (for ΤΟΥ) ΙΟΑΝΝΟΙ (for IOANNOY) ΤΩ ΚΙΡΡΙΩ ΚΑΡΩΛΩ (“[Verses] of John to the Lord Charles”). Moreover, two poems (12a-b and 13) are verifiably archetypal, as there are entries for both in L1. Footnote ²³ Poem 14 is a prayer for Charles and contains the characteristic misspelling ΚΛΕΟΡ. The authenticity of poem 17 is guaranteed by a new inscription on fol. 298r: ΤΩ ΚΥΡΡΙΩ ΚΑΡΟΛΩ ΙΩΑΝΝΗΙΣ ΧΑΙΡΕΙΝ (“John sends greetings to Lord Charles”). Although they display only the notation prosa, there would be little reason to think that the two prose prayers (15*) — one for Irmintrude, the other for Charles — were composed by someone else. Poem 14, prose 15*, and poem 17 are securely datable. All were written late in the year 869, as will be shown below. Poems 14 and 17 allude to Charles’ ascent to the kingship of Lotharingia in that year and crowning at Metz (9 September 869), while the first of the prose pieces (15*) is a prayer for the salutem of Irmintrude, who died on 6 October 869.

Traube, however, doubted the authenticity of poem 15, which does not address or mention the king and refers to Iohannes in the third person. He also understood the inscription before poem 16 to indicate Martin’s authorship. Accordingly, he printed both poems in an appendix to volume III.2 of the Poetae Latini, which he labelled “Carmina Scottorum Latina et Graecanica,” as XII.3 and XII.5, respectively.Footnote ²⁴ Poem 15 honors a certain Liuddo, who was thought to be a pupil of Martin’s at Laon.Footnote ²⁵ I cite my translation from the SLH edition:

Could Martin have written it? Liuddo? No, it is John’s. The Greek word for “glory” in the first verse is ΚΛΕΟΡ, an obvious error for κλέος (“fame, glory”). John made the same mistake in the first line of genuine poem 14: ΕΡΗΝΗ ΠΙΣΤΩ ΔΗΜΩ ΒΑΣΙΛΕΙ Η ΚΛΕΟΡ ΑΚΡΩ (“May there be peace for a faithful people and glory to its king on high”), where κλέορ should be κλέος and the article should be τὸ. As we noted above, the spelling κλέορ is a Merkmal of John’s composition. Other Greek words in the epigram appear in John’s genuine poems: ΔΗΜΩ (“to the people”) appears in no. 14, ΣΕΒΑΣΤΟΣ (“revered”) at 8.29. Words connected to ΛΑΜΠΕΙ (“shines”) are favored in John’s poetry. ΛΑΜΠΡΟΣ (“brilliant”) occurs at 17.12, ΛΑΜΠΡΟΤΑΤΟΣ (“most brilliant”) at 12a.2 and 12b.1. We have here an example of Eriugenian irony: John yields the field of Greek composition to Liuddo, the brilliant student of Martin’s, contenting himself to be a mere Latinist. But there is more to it. Note that the poem precedes the satire on the doddering teacher with the inscription ϹΤΙΧΟϹ ΠΡΕΠΟΣ ΔΙΔΑϹΚΑΛΟϹ ΜΑΡΤΙΝΟΥ, which is either “the beautiful verse(s) (uersus pulcher) by Martin” or “verse(s) befitting Martin,” which would make Martin the object of the satire. Why does John think Liuddo to be so brilliant? Could it be that it was Martin’s own student who wrote the clever spoof on Martin? Of course, this is not demonstrable.

What I think is demonstrable is that John wrote the gloss to this poem and to all the glossed poems 1–16.Footnote ²⁶ Versions a and b of poem 12 have a metrical gloss, as do poems 14 and 15. If we grant that the near unity of the glosses in poems 1–10 is due to a common exemplar in which Greek glosses were written above the graeca in the Latin poems, and that these glosses are authorial or were at least vetted for accuracy by the author, is it likely, or even possible, that Martin glossed the rest of the poems independently? We get a chance to test Martin’s ability at Greek from his inscriptions. Take, for example, the two consecutive inscriptions on fol. 297r:

Leaving aside orthographical variation such as omega for omicron and vice-versa (common in the remnants of Carolingian graeca), there are elementary mistakes. Both lines exhibit morphological errors. In the first, we see TOY for OI.Footnote ²⁷ In the second, ΤΟΥΤΩ looks like a mistake for HEAYΤΟΥ (ἑαυτοῦ, “his”), and ΑΝΑΚΤΩ should be ΑΝΑΚΤΙ (“to the king,” lit. “leader”). Another inscription by Martin is found at the end of L1 on fol. 296v:

ΕΛΛΗΝΙϹ is a mistake for ΕΛΛΗΝ, the classical form still attested as correct in the Byzantine period, though ΕΛΛΗΝΙϹ is attested for a Greek woman in the New Testament.Footnote ²⁸ ΓΡΑΨΕ is missing its augment, although there are examples of its disappearance in the Volkssprache. Footnote ²⁹ If Martin meant to say that he wrote the foregoing letters, the Greek should be ταῦτα τὰ γράμματα. Note that in the inscription to poem 16 Martin wrote πρέπος for πρέπων. Carlotta Dionisotti makes the following assessment of Martin’s Greek:

When one considers the general accuracy of the glossing throughout the poems, including glosses to lines arranged as Latin hexameters (a notable feat), the author of the glosses was someone fairly competent in Greek and able to write acceptable Latin poetry. I know of no Latin poems written by Martin. John Scottus himself emerges as the compiler of the glosses to his poems. This being the case, even where the authorship of a poem is contentious, as with poem 16, the glosses to it were written by John and justifies the inclusion of the poem in the anthology.

Poem 16, a witty epigram on a doddering teacher, was thought by Traube to be the work of Martin, to whom the inscription ΣΤΙΧΟΣ ΠΡΕΠΟΣ ΔΙΔΑΣΚΑΛΟΥ ΜΑΡΤΙΝΟΥ refers. The poem makes fun of an elderly teacher, whose student is no longer prepared to listen to him:

In fact, a glance at the facsimile of L, fol. 297v (plate II of the CCCM edition) shows that the verse or verses under the inscription were erased. I assumed that given its lowly place at the bottom of the folio in the left margin it had been squeezed in by Martin and was not meant to interrupt the series of verses by John, and therefore the epigram on the teacher, the first item on the following folio (fol. 298r), was John’s own work. However, Carlotta Dionisotti noted the exceptional quality of the Greek poem; indeed, she says, “…the constructing and phrasing, with lively dialogue, humour and vivid imagery in lines 3 and 5 seem to me superior to any other Carolingian Greek poem.”Footnote ³² The wit is worthy of John; however, the Greek is too good for him, never mind Martin. Indeed, there are a number of words that do not occur in any of Johns’s other Greek verses, and there is a nice word-play in the last verse: ΛΗΓΕΙΝ with eta and ΛΕΓΕΙΝ with epsilon: “to cease” and “to speak.” John avoids ‘complicated’ verb forms such as ἐλήλυθα (perfect indicative, active voice) and τιθῶμαι (present subjunctive, middle voice) that occur in this poem (lines 2 and 4, respectively). One must conclude that the Greek poem is by an anonymous writer; the Latin gloss is by John. In general, poem 16 is well glossed, but there are a few errors, notably on the words that do not occur in John’s poems. For example, in line 2 he gives the wrong meaning for ΤΑΧΑ (gl. forsan), here meaning “swiftly” rather than “perhaps.” In line 3, ΚΡΟΤΑΦΩΝ is mistranslated as puerorum (“of the boys”), when in fact it means “of the foreheads”; ΑΠΑΛΩΝ does not mean rudium (“of the uncultivated”), but rather “of the soft” or “tender.” It is possible that John inserted this clever parody into a group of his genuine serious poems as a little prank on Martin.

We may conclude from the above that six of the seven poems given in L2 were genuine and all were glossed by John save poem 17. Nearly all the poems are addressed to the king or his wife, mention them, or relate to them in some way, including even 12a-b, which sides with the king in his dispute with Hincmar of Laon. Two poems (12a-b and 13) are attested by glosses in L1 and, therefore, are archetypal. There is nothing to suggest, as Dionisotti has pointed out, that Martin did anything else than to copy what was put before him, allowing for the inscriptions discussed above. At this point, I believe it fair to say that L (L1 + L2) gives a good account of the number and order of the poems that were in the archetype.

I turn now to the dating order of the poems, which is identical in poems 1–10 in R and L1 (allowing for the poems omitted by R). The dating of poems 1–17 is based on the notes in Traube’s edition, work by Dutton, and my own researches. Poem 1, probably written in 859, celebrates Charles’ victory over his brother Louis the German in 858, thereby preserving his kingdom.Footnote ³³ At the other end is the Greek poem, no. 17, celebrating Charles’ coronation at Metz in 869. Poem 4, Haec nostram dominam, dedicated to Charles’ first wife Irmintrude, can be dated to 864, as it alludes to the donation of an ornamental garment woven by the queen to the church of San Paolo fuori le mura in Rome.Footnote ³⁴ Poem 10, Graculus Iudaeus, marks Charles’ donation of gold and jewels to Saint-Denis and his assumption of the lay abbacy in 867.Footnote ³⁵ I think that we can date poem 8, Si uis ΟΥΡΑΝΙΑΣ also to the year 867. It celebrates two feasts, the Annunciation and the Redemption, that is, Good Friday. The Annunciation always falls on March 25, and, as Easter fell on March 30 in that year, Good Friday would have occurred on March 28. Charles attended Easter that year at Saint-Denis.Footnote ³⁶ Poem 8 is addressed to the king according to the Greek inscription, and the poet strongly implies that he is present. The verbal allusions to the Periphyseon may serve to mark the completion of John’s masterpiece in that year. The fact that the great poem Aulae sidereae (no. 25) does not appear in R and no glosses are recorded by L is an indication that it was written too late for inclusion in the collection. Indeed, it is likely that it was composed in 870.Footnote ³⁷

In addition to poem 17 celebrating Charles’ coronation at Metz, other Greek poems at the end of the collection might also be dated to 869. Poem 14, which begins ΕΡΗΝΗ ΠΙΣΤΩ ΔΗΜΩ (“Peace to your faithful people”), tells that God has set Charles before all other kings. Again, we may have a reference to Charles’ new or impending position as king of Lotharingia and his coronation at Metz as king of both Francia and Lotharingia. The poem concludes with a prayer asking God to subjugate the barbarous races to Charles. After his coronation at Metz, Charles had troubles with both the Saracens and the Northmen (that is, the Vikings). Charles did not come off well against either group and ended up paying a ransom to one and peace money to another.Footnote ³⁸ Another item likely datable to the waning months of 869 are two short prose pieces, which are included in the collection of short Greek poems. I marked them with an asterisk in the CCCM edition to note that they are not poems. The first is a prayer for the safety of Irmintrude and for her eternal salvation. Irmintrude died on 6 October 869; thus, the prayer may have been composed when she was in extremis, or shortly after her death – in which case the Greek soterian would refer to her spiritual safety.Footnote ³⁹ The two versions of poem 12 (12a and 12b) addressed to Hincmar, doubtless Hincmar of Laon, support the king in his dispute with that bishop that began on 1 December 868 and continued into 869.Footnote ⁴⁰ The poems are packaged by two ‘bookend’ pieces: Poem 1 (written in 859), which celebrates Charles’ victory over his brother Louis in 858; and poem 17, which rejoices in Charles’ ascension to the throne of Lotharingia on the death of his nephew Lothair II. John Scottus or a close associate compiled an anthology of these poems arranged in dating order, which John proficiently glossed, and gave to an assistant to be copied. From that copy (Ω), R and L were compiled, reproducing Ω’s errors and making new errors of their own.

A future editor is free, of course, to arrange the poems in a different configuration. For example, as suggested by one reviewer, the epigrams might be grouped together. However, there will continue to be nagging problems of authenticity, such as those surrounding the notorious epigram (now in four recensions) addressed to one of the Hincmars, claiming that the only good thing he did was to die.Footnote ⁴¹ Some compartment of a future edition will be needed to ‘house’ works of unproven authorship until convincing new evidence is found. However, it would be a mistake to disrupt the integrity of the anthology (poems 1–17), which provides a record in verse of Charles the Bald’s triumphs, setbacks, and continuing challenges in the years 858–869 and, perhaps, the best information we have regarding John Scottus Eriugena’s personality and his relationship with his king. The foundation established by Traube remains, on inspection, as strong as it was when it was laid in 1896, and leaves little room for constructing res novae. The editorial situation described in detail above is unusual, possibly unique. However, it does hold a lesson for editors of glossed texts. While not all glossed texts contain authorial glosses, texts transmitted in multiple copies may contain a common gloss. The glosses in each witness need to be collated with the glosses in the others as a regular part of the recensio codicum.