The dawn and twilight of Old Irish scholarship

ABSTRACT

Even though the publication of Johann Kaspar Zeuss’s monumental Grammatica Celtica (Zeuss 1853) marks the beginning of the modern, scientific study of the Old Irish language, short excerpts of the most important textual witnesses of Old Irish, the so-called Old Irish glosses, preserved in 8th–9th-century manuscripts on the European Continent, had appeared in print since the early 18th century. This article gives an overview of these early publications from the 18th and early 19th century by Johann Georg von Eckhart, Domenico Vallarsi, Lodovico Antonio Muratori, and Vittorio Amedeo Peyron, and assesses their role as early trailblazers in the study of Old Irish.

In the history of scholarship, the beginning of the modern study of the earliest surviving manuscript witnesses of the Old Irish language, namely the so-called Old Irish glosses and related material, is inextricably linked with the work of one man: Johann Kaspar Zeuss, the historian from Kronach in Franconia.¹ His monumental Grammatica Celtica from ¹⁸⁵³ initiated modern scientific research into the history of the Celtic languages and established Celtic Studies as an independent and legitimate academic discipline on equal footing beside more time-honoured philologies such as the Classical languages Greek and Latin or Sanskrit.² This is the point in the history of the language sciences at which Old Irish scholarship emerged, as it were, into the full light of day. Old Irish is the name for the oldest stage of the Irish and Scottish Gaelic language for which rich written documentation is available; in absolute chronology it corresponds to the 8th–9th centuries a.d (Thurneysen 1946, 1; Stifter ²⁰⁰⁹, 55; Griffith & Stifter forthcoming). Old Irish is preceded by the more patchily attested Archaic or Early Old Irish (7th century). The 4th–6th centuries, which are only accessible in a small number of epigraphic testimonies in the unique ogam script,³ are called Primitive Irish. The periods after Old Irish are Middle Irish in the 10th–12th centuries and Modern Irish since 1200.

Zeuss worked on the basis of the sensible, but at the time radical-sounding, principle that in order to understand the grammatical character of a language in diachronic and comparative perspective, it is necessary to make use of its earliest extant sources.⁴ In order to establish the grammatical profile of Irish, Zeuss therefore went back to the Old Irish glosses of the 8th and 9th century, preserved in libraries on the European Continent. He published selections of these with Latin translation in the appendix to Grammatica Celtica (Zeuss 1853, 964–1076; Ebel ¹⁸⁷¹, 978–1051). The central position of the glosses in the study of the Old Irish language derives from their large number (over 16,000 in total) and their lexical, morphological and syntactical variety,⁵ which makes it possible to build the grammatical description of Old Irish almost exclusively on them, even though they represent only a small fraction of the surviving Irish texts that go back to that period. They have in fact served as the foundation for all the authoritative grammars of Old Irish (Zeuss 1853; Ebel ¹⁸⁷¹, but especially Thurneysen ^{1909, 1946}) in the past two centuries. The crucial advantage of the glosses over all other texts is that they are contained in manuscripts from the very period in which they were written, that is to say, they authentically represent the language of their time.

In contrast, the large bulk of Early Irish narrative and poetic literature that has survived in Ireland itself and that is much more diverse and interesting from the point of view of its content than the glosses, is only found in manuscript copies from later, sometimes much later, times. In their transmission, those texts have typically undergone countless reworkings, adaptations and modernisations in their language, so that their genuinely Old Irish core is often very difficult to extract from underneath a much younger linguistic veneer. Some texts only survive in copies that were made a millennium after their original composition.

Among the Old Irish texts in contemporary manuscripts of the period, three large corpora stand out: the glosses from Würzburg (Wb.), Milan (Ml.), and St Gall (Sg.), each consisting of thousands of glosses, i.e. single words, sentences or, very rarely, longer passages. Beside those three big, glossed corpora, there is a comparatively large number of ‘minor’ glosses, i.e. Latin manuscripts with occasional Old Irish notes, ranging from one to a few hundred. All Old Irish texts, glosses, and verse preserved in contemporaneous manuscripts known by the year 2013 are conveniently catalogued with concise bibliographical information in Bronner ²⁰¹³.

In the decades following Zeuss’ pioneering work and his premature death in 1856, various editions and collections of Old Irish glosses were published, for example by Stokes ^{1866, 1872} and ¹⁸⁸⁷, Nigra ^{1869, 1870–1872}, Ascoli ¹⁸⁷⁸, Zimmer ^{1881, 1886}, and Windisch ¹⁸⁸⁴, to name but the most prominent.⁶ All of these works were ultimately superseded by the monumental Thesaurus Palaeohibernicus (= Thes., Stokes, & Strachan ¹⁹⁰¹) at the beginning of the 20th century. Its editors, Whitley Stokes and John Strachan, assembled all contemporary Old Irish material known at the dawn of the 20th century in this still seminal collection. A small number of Old Irish glosses and other texts have been discovered since the publication of Thesaurus Palaeohibernicus, but none of these additions has changed the big picture of the Old Irish language that had been arrived at in the early 20th century.

For a century, Thesaurus Palaeohibernicus remained the virtually unquestioned reference point for Old Irish Studies. However, as part of a wider trend towards digitisation in the humanities, several research projects in the early 21st century have been and are dedicated to re-evaluating these earliest witnesses of Old Irish (cf. Griffith, Stifter, & Toner ²⁰¹⁸, 9–18). One slightly earlier work that straddles the watershed between the print-based and the digital epoch in Old Irish Studies is the exhaustive lexicon of the Würzburg glosses by Kavanagh ²⁰⁰¹, the fruit of a life-long labour throughout most of the 20th century, edited posthumously by Dagmar Wodtko. It is a printed dictionary running to 900 pages, accompanied by a CD-ROM (a technology that is now all but obsolete) with searchable pdfs. Very occasionally, Kavanagh’s readings improve on those in the Thesaurus.

The genuinely digital projects are the Milan Glosses Database (Griffith & Stifter ²⁰¹³), the Parsed Old and Middle Irish Corpus (POMIC = Lash ²⁰¹⁴), the Priscian Glosses Database (Bauer 2014; later integrated into Hofman, Moran, & Bauer ²⁰¹⁸), the Würzburg Irish Glosses (Doyle 2018), and the digitisation of the Vienna Bede fragment (Gloss-ViBe = Bauer ²⁰²³). These projects have occasionally led to substantial revisions of the received text of the Old Irish glosses (e.g. Bauer ²⁰¹⁷; Griffith & Stifter ²⁰¹⁴). Several of these editions, augmented by more texts, have been combined in the database Corpus PalaeoHibernicum (CorPH; Stifter et al. ²⁰²¹) as part of the ChronHib project (2015‒2021). This database will be undergoing a revision and expansion as part of the DiAgnostic project (2023–2027), both projects led by the present author (see footnote 1). A related project is Pádraic Moran’s GLOSSAM project (Moran ^2022–2026) that, among other things, will create a framework for the digital presentation of main text and paratext.

The effects of these digital editions have been making themselves felt slowly, but steadily, in the past years. In the long perspective, they will have a lasting impact on how research in Old Irish Studies is conducted. Because of the possibility of getting fast and reliable quantitative results about the distribution of forms and complex constructions, the way in which Old Irish is grammatically described is bound to change in the coming years. In particular, it can be expected that the syntactic description of the language will take a more prominent place in future grammars than the one that it occupies in currently available handbooks. Given the progress that has been made in Old Irish Studies since the time of Zeuss and Thurneysen, and the much better understanding of the phonological, morphological, syntactical and lexical distinctions between the chronological stages of Irish in general, future grammatical projects will also have to pay more attention to sources aside from the glosses.

While it can thus be seen from this short history of research that the progress in the study of Old Irish preserved in written sources from that time is well documented since Zeuss’ time up to the present, the dawn or twilight of Old Irish Studies (or its stone age, depending on one’s preferred metaphor) is little known. It will therefore be worth the while to delve into the deep history of scholarship in the field. What is overlooked is that the first hesitant studies of Old Irish glosses were already undertaken before the founding fathers of modern Celtic Studies. The aim of the following account is to draw attention to the earliest printed excerpts from Old Irish glossed manuscripts, especially from the Milan manuscript. Some of these printed specimens predate the beginning of Celtic Studies as we know it by more than a century. Zeuss (1853, xx–xxiii, xxix–xxxi), Stokes (e.g. ¹⁸⁶⁶, 1, 17), and Zimmer (1881, xvii–xviii) still made reference to those pioneer publications in the introductions to their own works, and they are referenced in Thesaurus Palaeohibernicus (Thes. i xiv–xv, xxii), but Thurneysen (1909, 1946) no longer mentions them, and it seems that from the beginning of the 20th century these earliest contributions to the field had started to lapse into oblivion. The names and works of these scholars are not recorded in the Bibliography of Irish Linguistics and Literature until 1912 (Best 1913), and these early specimens of printed Old Irish are also missing from the recent Clóliosta (Sharpe & Hoyne ²⁰²⁰).

Johann Georg von Eckhart

The distinction of being the first modern scholar to print extracts from Old Irish glosses goes to Johann Georg von Eckhart (1674–1730). Von Eckhart was long-time secretary to the polymath Gottfried Wilhelm Leibniz (1646–1716),⁷ but had also been Professor of History at the University of Helmstedt (1706–1714) for a short number of years. His methodology of language comparison and interest in the prehistory of languages was strongly influenced by Leibniz.

In 1729, von Eckhart, who by that time had taken up the position as librarian and historiographer to the Bishop of Würzburg Christoph Franz von Hutten, published a two-volume study of Commentarii de rebus Franciae orientalis et episcopatus Wirceburgensi ‘Notes about East Franconia and the Diocese of Würzburg’. The large number of old documents that he brought together for his study includes extracts and a discussion of the Würzburg manuscript M.p.th.f.12, containing the epistles of St Paul with Latin and Old Irish glosses. Another notable text is the first printed version of the Old High German Hildebrandslied. Poppe (1986, 77) remarks that this publication gives ‘good evidence for the interrelation of his historical, philological, and linguistic interests’.

In volume 1, von Eckhart reproduces about half of the Würzburg glosses from folio 1a down to what corresponds to gloss 4a23 in Thesaurus Palaeohibernicus (von Eckhart ¹⁷²⁹, i 847–853), altogether 186 of the 377 glosses edited in Thesaurus Palaeohibernicus. Because of the experience he had acquired under Leibniz, he correctly identified their language as Irish: ‘Ex allatis enim manifestum est, Glossas istas Hibernicas et vetustissimum huius linguae monumentum esse’ (‘From what has been cited, it is evident that these glosses are Irish and the oldest document of that language’; von Eckhart ¹⁷²⁹, 453). Even though his readings contain numerous errors, and his attempts at explaining and interpreting the glosses (von Eckhart ¹⁷²⁹, 452–453) do not stand up to modern scholarship, Zeuss (1853, xx) nevertheless referred favourably to von Eckhart’s work. For more about von Eckhart, see Shaw (1956, 7–8) and Poppe (1986, 72–78, 82–84; who reproduces Eckhart’s first gloss).

The present study, however, is chiefly concerned with the earliest, pre-Zeussian specimens of the Old Irish glosses from the early-9th-century Milan Biblioteca Ambrosiana MS C 301 infra to appear in print,⁸ that is, the selections published by Domenico Vallarsi and Lodovico Muratori in the 18th and by Vittorio Peyron in the early 19th century. Reference to specific glosses will be according to the number system in Thesaurus Palaeohibernicus (Thes.) and to their ID in Corpus Palaeohibernicum (CorPH).

Domenico Vallarsi

As far as I have been able to establish, the first person to bring an extract, namely a single phrase, from the Old Irish sections of the Milan manuscript into print, albeit in garbled form, was Domenico (also Dominic) Vallarsi (1702–1771), a Jesuit-educated Italian priest from Verona, who was active just a few years after von Eckhart. Vallarsi’s interests in antiquities, manuscripts and patristic studies were actively supported by the city and the bishop of Verona, as well as by Pope Benedict XIV. His magnum opus is an edition of the works of St Jerome in eleven volumes, published from 1734–1742, and then republished in a revised and enlarged version 1766–1772. Vallarsi’s edition of the Hieronymica was reprinted in Migne’s Patrologia Latina series 1844–1855, which also serves as the basis for the citations here.

In volume 7 of the Hieronymica, Vallarsi includes an extract from the Milan Codex Ambrosianus C 301 inf. under the title ‘Breviarium in psalmos’, having correctly recognised that, although ascribed to the church father, it was not an authentic work of St Jerome’s. Vallarsi prints specimens of the commentaries on psalms 1, 3 and 4 from the manuscript (1845, 815‒822), but he does not include any glosses on that text, nor does he mention the numerous Old Irish glosses at all. Ascoli, who edited the manuscript more than a century later, was critical of the many errors in Vallarsi’s text (1878, xii–xiii). The excerpts are preceded by a long introduction (1845, 801–814), in which Vallarsi gives a comprehensive description of the manuscript. He only notes the presence of writing in a language unknown to him for the initial page of the manuscript. The recto of the first folio, numbered ‘6’, does indeed contain two riddle poems in the Old Irish language (Thes. ii 291–292; Ahlqvist ²⁰¹⁸; Stifter forthcoming-b). The understanding and study of those poems has been hampered both by their genre ‒ they are riddles for which no solutions are provided in the text ‒ and by the fact that the first page has suffered considerable wear and tear over the centuries, aggravated by the use of reagent (Best 1936, 16), probably sometime before the middle of the 19th century.

In his discussion of the manuscript, Vallarsi quotes a few words from the first page. Vallarsi’s style of writing is not only difficult to follow because of his elaborate Latin syntax, but also because of several printing errors. For instance, the nonsensical filio ‘son’ in the first sentence is evidently a mistake for folio ‘folio’. The following extract, cited after the reprint in volume 26 of Migne’s Patrologia Latina, is concerned with the first page of the manuscript (Vallarsi 1737 = 1845, 813–814):

Codex sub C littera uno supra trecentesimus numero prænotatur, estque oblongæ, ut vocant, in filio [sic!] formæ. Olim ad S. Columbani de Bobio monasterium pertinuit, ex quo omnium pretiosissima eo importata sunt antiquitatis monimenta. Charactere descriptus est, ut vocare antiquariis placet, cursivo vetustioris formæ, et quid ad eam proxime accedat notarum figuram, quam in Ægyptiis papyris hodienum [sic!] cernimus. Quamobrem et est sæpe lectu perquam difficilis, ut notis internoscendis multa exercitatione opus sit: et litteris interdum exesis vetustate, aciem oculorum diffugit, ut salius non uno in loco duxerimus, pati lacunulas aliquot, quam ex ingenio supplere, si fieri id tuto non posset.

Prior charta, qua ad codicem prætexendum librarius est usus, Latinis illa quidem litteris tota describitur, iisque fere similibus reliquo codicis characteri, sed quam sonant linguam, cl. vir cui summa omnia tribuo, ignotam dixit, aut veterem Illyricam. Ego, si hoc ipsum lectoris interest scire, Hebraicam esse monuerim: sunt enim ejus rei indicio, quæ possint, plerisque aliis oblitteratis, pauca legi Hebraicæ terminationis, et soni vocabula: cum primis vero isthæc, quæ ad vocem hiruzech apponitur ad libri oram interpretatio, urbs fortitudinis nostræ. Ita nimirum quæ vocem illam componunt verba, hir, Hebr. עיר, urbem: uz, Hebr. עו [sic!⁹], fortitudinem sonat: affixum denique ch, Hebr. ו, pronomen est, tuum. Rescribendum igitur, inquies, Latine erat, urbs fortitudinis tuæ, non nostræ, ut Hebræo responderet, quod non diffiteor; verum ita sentio, non ejus vocis in Latinum explicandæ gratia, sed ut paulisper diversam ab illa significaret esse Latinorum lectionem, fuisse ab studioso aliquo notam appositam. Is vero quicunque fuerit, Origenis industriam est imitatus: quem enim ille Hebræum textum Græcis litteris sibi descripsit in Hexaplis, hic Latinis reposuit: nimirum uterque vernaculis. Fortasse etiam illum Origenis apographum, non Hebræum archetypum Latine repræsentavit: idque causæ fuit, cur peculiares primigeniæ linguæ sonos, et quibus scatet aspirationum modos, idque genus alia de Græcis non usque adeo ad rem aptis elementis, ad Latina minus fortassis commoda per vim detorta Scriptura sæpe non referat, et dare sine mente sonum videatur.

‘The manuscript under letter C is marked one number above the three-hundredth [i.e. 301], and it is, as one says, a folio of oblong shape. Formerly it belonged to the monastery of St. Columbanus of Bobbio from where the most valuable ancient documents of all have been brought there [i.e. to Milan]. It is written in old cursive style, as the antiquarians call it, and which (?) resembles most closely that shape of characters that we see today in Egyptian papyri. Therefore it is often rather difficult to read, so that a lot of experience is required to distinguish the characters: and since the letters are sometimes worn off by age, it scattered the sharpness of the eyes so that not just once did we consider it wiser to leave lacunae rather than to come up with something to fill them, if it was not possible in a safe manner.

The first page, which a librarian used as front cover for the manuscript, is all written in Latin letters which are fairly similar to the style of the rest of the manuscript. However, in what language they resound, was declared unidentifiable by a gentleman, by whom I set the highest store of all, unless it be Old Illyrian. In my opinion, if the reader is interested in it, I would say it is Hebrew: evidence for this is the little that can be read of Hebrew endings and vocabulary, although it is for the most part obliterated: first and foremost, however, the translation urbs fortitudinis nostrae “the city of our strength” that has been added on the margin to the word hiruzech. The words that make up this compound expression are doubtlessly hir, which means “city” in Hebrew, and uz, Hebrew for “strength”: finally, the affix ch is the pronoun “your”. One may object that in Latin urbs fortitudinis tuae should have been written, not nostrae, in order to correspond to the Hebrew, which I do not deny. However, I believe that this comment has been added by a student, not in order to explain the expression in Latin, but to mean something slightly different from it. But whoever it was, he imitated the zeal of Origines: the Hebrew text that the latter wrote for himself in Greek letters in the Hexapla, the former represented in Latin letters: both of them, doubtlessly, for native speakers of these languages. He probably represented not the Hebrew archetype, but Origenes’ copy in Latin: for that reason it was why a script often does not represent the peculiar sounds of the original language, and the types of aspirations of which it abounds, when it is forced to represent other matters for which Greek is already not suited, but even less so proper for Latin, and that’s why it would seem to offer meaningless sounds’.

Notes:

The anonymous gentleman’s Illyrica ‘Illyrian language’ is the New Latin name for the Croatian language. Vetus Illyrica, therefore, has nothing to do with the elusive Illyrian language of antiquity (for which see Eichner ²⁰⁰⁴; Matzinger ²⁰¹⁶, 10–22), but corresponds to Old Church Slavonic, or Vallarsi’s informant perhaps thought of an early stage of Croatian. Needless to say, this identification is wrong.

The only Old Irish phrase that Vallarsi quotes is the misread hiruzech, for what is correctly Old Irish hisa tech ‘into the house’. Vallarsi wrongly identifies it as Hebrew, but his etymological explanation may in fact shed an indirect light on why an earlier scholar, a 15th-century hand according to Best (1936, 11), had written the quote u[r]bs fortitudinis nostrae from Isaiah 26:1 in the margin beside it. This line, urbs fortitudinis nostrae Sion ‘we have a strong city in Zion’, forms the basis of a Gregorian antiphon for advent. Perhaps that person had made the same, wrong etymological connection that informs Vallarsi’s own analysis.

Origenes’ Hexapla is a critical edition of the Hebrew Bible for comparative purposes, made around 245 a.d. in six versions, arranged in six parallel columns (Field 1875). The first column contained the Hebrew text of the Old Testament, followed by a Greek transliteration of the Hebrew, and four Greek translations, including the Septuagint. Vallarsi likened his assumed reading hiruzech in the Codex Ambrosianus to Origenes’ second column, i.e. a Hebrew text transliterated into a classical alphabet, but here Latin instead of Origenes’ Greek. Vallarsi was aware of the fact that the Latin and Greek alphabets are not suited to express all the graphic (and phonological) distinctions of Hebrew.

Lodovico Antonio Muratori

After his ordination in Modena in 1694 and a short stint at the Biblioteca Ambrosiana in Milan (1695–1700), the catholic priest Lodovico Antonio Muratori (1672–1750) was appointed archivist and librarian at the Biblioteca Estense, the Ducal library in Modena, a position he held until his death. A leading scholar of his time, he was a prolific writer on the history of Italy. Shortly after Vallarsi, he published Antiquitates Italicae Medii Aevi, a collection of seventy-five essays on historical themes in six volumes (1738–1742) as an elucidation and supplement to his work on the sources of Italian history.

The third volume of the Antiquitates (Muratori 1740, 857‒871) contains a short description of Codex Ambrosianus C 301 inf. Like Vallarsi, this includes excerpts from the commentary on the psalms (pp. 859–871; the commentaries on psalms 1 and 2 in extenso, followed by short extracts of several others), but without the accompanying glosses. Even though Muratori (1740, 857) cites the Latin titular inscription on page 1 in the beginning, he makes no mention of the much more prominent Old Irish poems on the same page:

… reperi ego in prælaudata Ambroſiana Bibliotheca Codicem, characteribus tantæ vetuſtatis exaratum, ut mihi videretur ætatem mille annorum attingere. Titulus quidem vetuſtiſſimus, ſed literis non adeo antiquis, ita ſe habet : In hoc Volumine continetur Hieronymi Presbytheri Expoſitio ſuper Pſalterium, non tamen a primo Pſalmo prius, ſed quoſdam alios indirecte prius exponere videtur. Deinde ad Pſalmorum ordinem, ideſt a primo incipiens & demum ſubſequenter uſque ad finem Pſalterii.

… I have found in the aforementioned Biblioteca Ambrosiana a Manuscript, which is written in a hand of such antiquity that it would seem to me to come close to an age of a thousand years. The title, which is very old, but not in such ancient letters, goes thus: In hoc Volumine continetur Hieronymi Presbytheri Expositio super Psalterium, non tamen a primo Psalmo prius, sed quosdam alios indirecte prius exponere videtur. Deinde ad Psalmorum ordinem, idest a primo incipiens et demum subsequenter usque ad finem Psalterii.¹⁰

At the very end of his treatment of the manuscript, after the specimens of the commentaries, Muratori provides a few examples of the glosses, whose language he was the first to recognise as a form of Irish – an identification which, being unfamiliar with the language himself, he put forward only with great caution. In total, there are twelve, mostly very short, randomly chosen examples of the glosses from the first third of the manuscript. Most of them are heavily garbled when compared with modern editions. His specimens of Irish also include the beginning of the comparatively long narrative passage, written on a slip of vellum, that has received the gloss number 52×00 in the Thesaurus Palaeohibernicus (Thes. i 164):

Interea neque illud prætereundum, in antiquiſſimo iſto Codice Ambroſiano difficiliora verba per gloſſam interlinearem interdum explicari; hoc eſt, per Linguam Septentrionalem, quæ num Scotica ſit, Eruditi Britanni decernent. Exemplum dabo : Verecundiæ, ſupra ſcribitur féle : Huc illucque, innunn hille : Homo a Deo ſuſceptus, annanno & Deudit donucht : Tanto honore afonuitmit dodia : Piaculi, inchuil : Anxium, dubach : Manifeſta isfoll : Ad vendicandum, bediachti : Convenienti testimonio, hond foncul immamcidiu : Supplicandum, dunduil : Triumphale, budath &c. Accipe etiam continuatum ſermonem : David niderb linnt in ſendias canone dunaith menadanis intrailſoadit maſued fonaith mentur and Dialvid for longars có. jadomdu tco am mondu re Saúl brethe hoſvidiu mondu ſectub doabi meleach hiterfodinaic manbtha David &c. Sed jam progrediendum.

‘At the same time, it must not be passed over that the more difficult words [in the Latin of the commentaries] are explained in this very old Ambrosian Manuscript by interlinear glossing, namely in a Northern Language. British Scholars will decide if it is Scottish. I will give an example: above Verecundiae, féle is written; huc illucque, innunn hille; Homo a Deo susceptus, annanno & Deudit donucht; tanto honore, afonuitmit dodia; piaculi, inchuil; anxium, dubach; manifesta, isfoll; ad vendicandum, bediachti; convenienti testimonio, hond foncul immamcidiu; supplicandum, dunduil; triumphale, budath etc. Take also this continuous narrative: David niderb linnt in sendias canone dunaith menadanis intrailsoadit masued fonaith mentur and Dialvid for longars có. jadomdu tco am mondu re Saúl brethe hosvidiu mondu sectub doabi meleach hiterfodinaic manbtha David etc. But we have to move on.’ (Muratori 1740, 871)

Muratori’s Scotica (lingua) ‘Scottish’ is a common Latin word for ‘Irish’. For ease of reference, the glosses are repeated in tabular form below, preceded by their number in Thesaurus Palaeohibernicus, and accompanied by a translation and their reading in Griffith and Stifter (2013 = CorPH Text ID 0006). The CorPH ID for each gloss is given underneath the traditional number in the Thesaurus. For easier readability, understrokes highlight the differences between the readings of the Old Irish words. Where the reading of the glossed Latin text in Griffith & Stifter ²⁰¹³ differs from that in the Thesaurus, it will be cited; otherwise the reading can be considered to be identical. Even though the differences are sometimes very small and seemingly trivial, they give evidence of different approaches to editing manuscript texts.

Thes./CorPH		Latin	Septentrionalis/Old Irish	Translation
15a14	M	verecundiae	féle
S0006–152	G&S	uerecondiæ	féle	‘of propriety’
15c6	M	huc illucque	innunn hille
S0006–179	G&S		innunn hille	‘hither and thither’
16a4	M	Homo a Deo susceptus	annanno &Deuditdonucht
S0006–205	G&S	post resurrectionem Homo a Deo susceptus	anarróetdeachtdonacht [leg. doinacht]	‘when the Godhead assumed Manhood’
17b10	M	tanto honore	afonuitmit dodia
S0006–289	G&S	digno tanto honore	a foraitmit [leg. foraithmit] do dia	‘that God should remember Him’
16c1	M	piaculi	inchuil
S0006–241	G&S		in chuil	‘of the sin’
Ml. 19a7	M	anxium	dubach
S0006–398	G&S		dubach	‘gloomy’
18c16	M	manifesta	isfoll
S0006–366	G&S		is follus	‘it is clear’
23d18	M	ad vendicandum	bediachti
S0006–711	G&S	ad uindicandum	bediachti	‘it should be avenged’
35b12	M	convenienti testimonio	hond foncul immamcidiu
S0006–1663	G&S	conuenienti testimonio	hondforcul immaircidiu	‘by the fitting testimony’
40b14	M	supplicandum	dunduil
S0006–2162	G&S	ad suplicandum	dundaíl	‘for the request’
51b26	M	triumphale	budath
S0006–3087	G&S		buadach	‘victorious’
52x00	M		David niderb linntin sendias canone dunaith menadanisintrailsoadit masued fonaith mentur and Dialvid for longars có. jadomdutco am mondu re Saúl brethehosvidiu mondu sectub doabi meleach hiterfodinaic manbtha David etc.
S0006–3149	G&S		IPsidauidrl.níderb linntrain senchas canone dunaithmenadarisintitul so acht masued foraithmentar and dialuiddauidforlongais có·iadomdu·ł.co ammondu·resául brethaehosuidiu mórdusetaib do abi meleȧch hiterfochraic marbtha dauid· […]	‘Ipsi Dauid rel. We are not certain as to the story of Scripture that he calls to mind in this heading, unless it is this that is recalled there. When David went into exile to the Edomites, or to the Ammonites, before Saul, much treasure was brought from the latter to Abimelech as the price of slaying David. […]’

Muratori’s readings of the Old Irish text exhibit the typical errors of someone not familiar with insular minuscule script, errors which are still very common among beginners in medieval Irish palaeography: aside from the perennial pitfalls of minim confusion (i.e. parsing incorrectly the short vertical strokes that make up i, n, m, u and h and assigning them to the wrong letters), Muratori’s text exhibits the confusion of n for r (35b12: foncul immamcidiu for forcul immaircidiu; the latter also featuring minim confusion), of u for open a (16a4: Deudit for deacht), of confusing c and t (51b26: budath for buadach), reading di for ch (52×00: sendias for senchas), and unrecognised abbreviations (52×00: t for trá; 18c16: foll for follus).

At this point it may be apposite to note that several years before he published his description of Codex Ambrosianus C 301 inf., Muratori already had, in ¹⁷¹³, edited an extensive part of another famous medieval manuscript with Irish associations kept in Milan, namely Biblioteca Ambrosiana MS C 5 inf., commonly known as the Antiphonary of Bangor (Muratori 1713, iv 119–159). In fact, the name Antiphonarium Benchorense goes back to Muratori himself (1713, 121–126). The Latin hymn Benchuir bona regula ‘Bangor’s good rule’ (fo. 30 r), which contains the Old Irish genitive Benchuir ‘of Bangor’ and the phrase Munther Benchuir ‘the monastic family of Bangor’, are printed on p. 156 of his book. Without alerting the reader to the omission, Muratori did not print several pages from the end of this manuscript, which dates to the final years of the seventh century. The omitted pages include fol. 34 r, which contains the only other thoroughly Old Irish phrase in the antiphonary, namely common oróit dún ‘this prayer is common to us’ (for which see Stifter forthcoming-a). On the other hand, Muratori did print the hymn [In] memoriam abbatum nostrorum ‘To the memory of our abbots’ from the very last page of the Antiphonary (fo. 36b; Muratori ¹⁷¹³, 159). This hymn, a litany of Latinised names of Irish saints, has also been included in Thes. ii 282.¹¹

Vittorio Amedeo Peyron

The third and last scholar who left his imprint on the early publication history of Old Irish material from the Milan Codex Ambrosianus, in the years immediately before Zeuss’ groundbreaking Grammatica Celtica, is Vittorio Amedeo (also: Amadeo) Peyron (1785–1870) from Turin. Peyron was professor of oriental languages at the University of Turin from 1815. His research focussed on Coptic, for which he wrote a celebrated dictionary and grammar. In the course of his studies, he published papyri from the collections in Turin and Vienna. Peyron’s interest in Codex Ambrosianus C 301 inf. is to be seen in the wider context of his research into palimpsests that also led him to discover hitherto unknown fragments of the works of Cicero, Empedocles and Parmenides, as well as of the Codex Theodosianus.

In the appendix to his edition of fragments of Cicero’s speeches pro Scauro, pro Tullio and in Clodium, he includes a description of the Codex Ambrosianus (Peyron 1824, 188–192). The following extensive extract is from pages 190–191:

Servatur in bibliotheca Ambrosiana distinctus nota C. 301. ord. infer. Membranaceus saec. viii.; character longobardus, seu, uti reor, Saxonicus ad romanum iam inclinans; folio.

In prima libri charta haec adnotavit antiqua manus Monachi Bobiensis. ‘In hoc volumine continetur Hieronymi presbiteri Expositio super psalterium non tamen a primo psalmo prius sed quosdam alios indirecte prius exponere videtur. deinde ad psalmorum ordinem idest a primo incipiens et demum subsequenter procedens usque ad finem psalterii’ quae cum consonant cum adnotatione Inventarii.

[…]

[p. 189]

Rectissime Muratorio animadversum est codicem scatere glossis lingua septentrionali, fortasse Scotica, descriptis. Praeter exempla a Cl^o viro col. 871. allatis alia proferam

in iurgia                         in  immur  —  obscuratione  solis      dintemul

committit                       dorogaib    —  osanna  in excelsis      slanuigthe

narraverunt ut abscenderent laqueos       sechisdorigensat  son

[p. 190]

 comminatur    dommathi     —   insigniorum curationum    innufertaie

 ab excelsis       honaib idlaib —   in excelsis                      isnaib tel  divib

 impugnatio     ontogail         —   his verbis                      usberum

 percutiebat      nosenned      —   conrasit                         roscaird

 viros  proferre   constituit    —   collundorucht.

Sed praestat longiorem glossam ad psalmum xxxiii. exscribere, utinam recte! neque enim vel syllatam [sic!] intelligo: ‘Ps. dd [quod vocis compendium notat david] rt niderb linnt in sendias canone dunaith menadaris intsailso acht [vel adit] masued fonaith mentur and dialvid dd for longais co. iadomdu. tco am mondu. resaul, brethae hosvidiu mordusetaib doabi meleach hiterfodiraic marbtha [vel manbtha] dauid. conranait side laithe nand iarsin fri dd et ninaithgaiin et leicsi huad airducoras tar dia deilb mordraige et firboith forsinni dauid diadiamlad connach nin geuin inti abimelech ciadud futharcair abas et is ducalugud buide dodia iarsint soiradsin rond so er rogab dd in salmso sis. i. bendicart.’ Huius partem vulgaverat Muratorius, qui haud vidit in prima libri charta haberi poëma hac eadem lingua conscriptum, quod decem et septem strophis constat. En postremas, quas facilius legi.

[¹] la. theglus corm roch

los irna fil act oendo

ros istech ndagsɔr dath

atchi intdichon adorsid

—————

Denuas dotiagas hisatech [vel hisutech]

indichec tegde doichlech

sis iarsindiu segde chludo

Gigar assimmurgu.

—————

[¹] Evanida est littera.

[p. 191]

Seilt insin anmin nimete

inthomnissid cose nas

min. emin hita tegilassa

calchondaresa.

—————

Christianum esse poëma liquet tum ex codicis natura, tum ex nomine ihu, quod in una stropharum legi.

‘In the Biblioteca Ambrosiana, a vellum manuscript with the mark C. 301. infer. of the 8th century is preserved; in a Langobard hand, or, as I think, in a Saxon hand¹² that is tending towards a Roman hand; folio. The ancient hand of a monk from Bobbio wrote the following note on the first page: ‘In hoc volumine continetur Hieronymi presbiteri Expositio super psalterium non tamen a primo psalmo prius sed quosdam alios indirecte prius exponere videtur. deinde ad psalmorum ordinem idest a primo incipiens et demum subsequenter procedens usque ad finem psalterii.’ This accords with the comments in the Inventory. […] Muratori very rightly drew attention to the glosses in a northern language, probably Scottish [i.e. Irish], in which the manuscript abounds. In addition to the examples that that gentleman provided in col. 871, I will adduce a few more.

in iurgia                          in immur — obscuratione  solis  dintemul

committit                        dorogaib  — osanna  in excelsis   slanuigthe

narraverunt ut abscenderent laqueos     sechisdorigensat son

comminatur   dommathi     —  insigniorum curationum   innufertaie

ab excelsis      honaib idlaib —  in excelsis                     isnaib tel divib

impugnatio    ontogail         —  his verbis                      usberum

percutiebat     nosenned       —  conrasit                        roscaird

viros  proferre constituit      —  collundorucht.

But it is better to write out in full a long gloss on psalm 33. Correctly, I hope, since I do not understand a syllable! ‘Ps. dd [this abbreviation stands for david] rt niderb linnt in sendias canone dunaith menadaris intsailso acht [or adit] masued fonaith mentur and dialvid dd for longais co. iadomdu. tco am mondu. resaul, brethae hosvidiu mordusetaib doabi meleach hiterfodiraic marbtha [or manbtha] dauid. conranait side laithe nand iarsin fri dd et ninaithgaiin et leicsi huad airducoras tar dia deilb mordraige et firboith forsinni dauid diadiamlad connach nin geuin inti abimelech ciadud futharcair abas et is ducalugud buide do dia iarsint soiradsin rond so er rogab dd in salmso sis. i. bendicart.’ Muratori had published a part of this, but he did not notice that there is a poem of seventeen stanzas in the same language on the first page. Here are the final [stanzas], which were easier for me to read.

[¹] la. theglus corm roch

los irna fil act oendo

ros istech ndagsɔr dath

atchi intdichon adorsid

————

Denuas dotiagas hisatech [vel hisutech]

indichec tegde doichlech

sis iarsindiu segde chludo

Gigar assimmurgu

————

[¹] A letter has become obscure.

Seilt insin anmin nimete

inthomnissid cose nas

min. emin hita tegilassa

calchondaresa.

That this is a Christian poem is manifest from the nature of the manuscript, but also from the name ihu which I was able to read in one stanza’.

On the following page there is again a synoptic reading of the glosses in Peyron (P) arranged according to their numbers and codes in Thesaurus Palaeohibernicus and in CorPH and in the revised reading of Griffith & Stifter ²⁰¹³ (G&S).

Thes.		Latin	Septentrionalis/Old Irish	Translation
16b9	P	in iurgia	in immur
S0006–229	G&S		in immargala	‘into fights’
16c7	P	obscuratione solis	dintemul
S0006–247	G&S	de obscuratione	dintemul	‘by the darkness’
16c13	P	committit	dorogaib
S0006–253	G&S		dorogaib	‘he commits’
17b15	P	osanna in excelsis	slanuigthe
S0006–293	G&S	osanna	slanaigthe	‘save!’
16d6	P	narraverunt ut abscenderent laqueos	sechisdorigensat son
S0006–262	G&S	narrauerunt	sech is dorigensat són	‘that is to say, they have done’
18c7	P	comminatur	dommathi
S0006–357	G&S		dommathi	‘he threatens’
17c9	P	insigniorum curationum	innufertaie
S0006–313	G&S	insignorum	innafertae	‘of the miracles’
14a9	P	ab excelsis	honaib idlaib
S0006–75	G&S		honaib idlaib .i. huare ishitilchaib ardaib· nobitis adi	‘from the idols, i.e. because they used to be in high hills’
14a11	P	in excelsis	isnaib teldivib
S0006–77	G&S		isnaib telchaib	‘in the hills’
14a13	P	impugnatione	ontogail
S0006–79	G&S	ab impugnatione	óntogail	‘from the destruction’
14a19	P	his verbis	usberum
S0006–85	G&S	his uerbis	asberam	‘which we will say’
2b9	P	percutiebat	nosenned
S0006–24	G&S		nosenned	‘he used to play’
14b2	P	conrasit	roscaird
S0006–87	G&S	conrassit	roscaird .i. rolommar	‘he has stripped, i.e. he has plundered’
2b12	P	viros proferre constituit	collundorucht
S0006–27	G&S	ex quibus.iiii. uiros praeesse constituit cantationibus	occollandorachtdoib	‘auguring to them’
52x00	P		Ps. ddrt niderb linntin sendias canone dunaith menadaris intsailso acht [vel adit] masued fonaith mentur and dialvid dd for longais co. iadomdu. tco am mondu. resaul, brethae hosvidiu mordusetaib doabi meleach hiterfodiraic marbtha [vel manbtha] dauid. conranait side laithe nand iarsin fri dd et ninaithgaiin et leicsi huad airducoras tar dia deilb mordraige et firboith forsinni dauid diadiamlad connach nin geuin inti abimelech ciadud futharcair abas et is ducalugud buide do dia iarsint soiradsin rond so er rogab dd in salmso sis. i. bendicart
S0006–3149	G&S		Psi dauidrl. níderb linntrain senchas canone dunaithmenadar isintitul so acht masued foraithmentar and dialuiddauidforlongais có· iadomdu· ł. coammondu· resául brethae hosuidiu mór dusetaib do abi meleȧch hiterfochraic marbtha dauid· conranaicsidelaithe nand iarsin fridauid⁊ ni naithgeuin ⁊ leicsi huad air du corastar dia deilb mordraige ⁊ firboith forsinnídauid diadiamlad connach ningeuin intíabi melech ciadudfutharcair abas ⁊ is duatlugud buide dodia iarsint soirad sin rondsóer rogab dauid insalmso· sís·.i. ben[e]dicam rl.	Ipsi Dauid rel. We are not certain as to the story of Scripture that he calls to mind in this heading, unless it is this that is recalled there. When David went into exile to the Edomites, or to the Ammonites, before Saul, much treasure was brought from the latter to Abimelech as the price of slaying David. One day thereafter he (Abimelech) met David, and he did not recognize him, and he let him go, for God had put a form of ghostly appearance and of a simpleton on David to disguise him, so that Abimelech did not recognize him, although he desired his death. And it is to render thanks to God after that deliverance wherewith He delivered him, that David sang this psalm below, namely, benedicam etc.

Even more so than Muratori’s selection, the glosses chosen by Peyron do not follow any clear order, except for coming from the first few pages of the manuscript. Sometimes his specimens are considerably shorter than how the complete gloss is read today (e.g. 2b12: collundorucht vs. oc collandoracht doib). His motivation for choosing short, simple phrases may have been the desire to allow experts on Irish – by implication in Britain or Ireland, since there were none on the Continent at that time – to easily identify the words and the language. Peyron’s transcription exhibits the same mistakes as those already discussed in the case of Muratori, but altogether the number of misreadings is conspicuously smaller. Like Muratori, wrong word divisions are due to the fact that he was unfamiliar with the language and therefore had no feeling as to what might be a full lexeme or what was merely a verbal or nominal ending (e.g. 14a11: tel divib for telchaib; 52 × 00: dunaith menadaris intsailso for dunaithmenadar isintitul so).

Peyron realised very perceptively that the text on the first page of the manuscript was in verse, arranged in seventeen stanzas, of which he printed the final three stanzas. He was clearly guided in his identification of the verse character of the text by the seventeen large initials that recur in regular intervals. His arrangement of the stanzas in lines is nevertheless haphazard. From the way they are rendered it appears that Peyron thought that the lines in the manuscript coincided with the lines of the poem. The following correct reading and metrically meaningful arrangement of those three stanzas is from Stifter (forthcoming-b), where they are numbered II.6–8:

Is glae thegdais – torm ro·chlos –

inná·fil ac[h]t óendoros.

Is tech ndagḟir – dath at·chí –

nít díchoím a dorsidi.

De ’núas do·tíagar hisa tech;

ní·dichet teg Dé doichlech.

Sís íar suidiu – ségde chlú –

do·tíagar ass immurgu.

Is ed trá in sin amnin.

Ní méte ní·thormassid

écosc n-aímin – airm hi·tá –

tegdassa ad·chondarc-sa.

‘It is a bright house – a report that has been heard –

into which there is only one door.

It is the house of a good man – the appearance that you see –

its doorkeepers are not lowly.

From above, one comes into the house;

a fool cannot go to God’s house.

Downwards, however, after that – propitious the fame –

one comes out of it.

So, that’s it then.

You hardly will not guess it,

the lovely form – where it is –

of the house that I saw.’

On slip 52, which contains the comparatively long episode about David and psalm 33, a modern hand has written Lingua cambro-britannica ‘Welsh language’. In view of this manifestly wrong identification, the note must pre-date Zeuss and probably also Muratori and Peyron, who, as was seen above, both regarded the language of the glosses as most likely Irish. The identity of the person who left this note must remain unsolved for the moment.¹³

Peyron (1824, 191–192) concludes the chapter on the Codex Ambrosianus with a brief description of three related fragments of manuscripts with commentaries on the psalms.¹⁴ Of the third, he says that it abounds with ‘multis glossis interlinearibus Saxonicis’ (‘many interlinear Saxon glosses’). It remains unclear from his account whether Peyron meant the adjective Saxonicus ‘Saxon’ in the sense of the Anglo-Saxon language or of Insular script. That manuscript is in fact Codex Taurinensis F. IV. 1, fasc. 7, with Old Irish glosses and scholia on the Gospel of St Mark.¹⁵ Peyron does not print examples of those glosses.

Conclusion

Having largely passed into oblivion today, the first specimens of Old Irish texts found in manuscripts preserved in Continental European libraries were printed considerably earlier, namely in the early 18th century, than what is today regarded as the beginning of modern Old Irish scholarship, Zeuss’ Grammatica Celtica from ¹⁸⁵³. Although the editors of those first selections of glosses did not understand the texts and accordingly introduced many, rather typical, misreadings, most of them suspected the language to be a form of Irish or identified it correctly as such. In this way, their small selections must have helped to create an awareness of the presence of those early medieval Irish texts in Continental European libraries. Thus they paved, directly or indirectly, the way for their systematic study first by Zeuss, and then by scholars such as Nigra, Ascoli, Zimmer, Thurneysen, and finally Stokes and Strachan in the latter part of the 19th and early 20th century. These are still the foundations on which we stand in the study of Old Irish, but the piles underneath them go down deeper in time than we are aware.

Biography

David Stifter is Professor of Old and Middle Irish at Maynooth University (Ireland). His research includes the diachronic and synchronic grammar and textual philology of Old Irish and of the ancient Celtic languages.

Disclosure statement

No potential conflict of interest was reported by the author(s).