The date of the writing of the Codex Argenteus has been established by false presumptions and insufficient information. The fact that some fragments were found at Bobbio does not in the least entitle us to draw the conclusion that all Gothic documents originated in Italy. The ninth or tenth century sentences in the Alcuin text were written in France; there are other fragments, which were found in Egypt and whose origin in Italy is highly improbable; and the Codex Argenteus, so far as we can trace it, has never been connected with Bobbio or Brescia. Then, the date and provenience is based on the resemblance of the Gothic MS. to the Codex Brixianus. But that is a gratuitous assumption. There is absolutely nothing in the Gothic text or script that gives the slightest clue to its palaeographic dating. The only thing we see is that the letters are made with extraordinary precision and are solid in body, not as was the writing in the fifth century in the Codex Brixianus, but of precisely the same quality as in the imitative art of the Carolingians, who reproduced the fifth century writing in all its details. (36) It is true that the Codex Brixianus had gold and silver letters on purple vellum, but Berger knows of a very large number of MSS. of the chrysographic art in Carolingian times, (37) and the Bible of Theodulphus, the Visigoth, of the ninth century, bears as striking a resemblence to the Codex Argenteus, for it, too, has gold and silver letters on purple vellum, and the exquisite regularity of the script is the same. (38) Indeed, it was through the efforts of the Visigoth Theodulphus that such calligraphy was practiced at Fleury. (39) Hence the identification of the calligraphy of the Gothic Bible with that of the Codex Brixianus is without any foundation whatsoever. But we have a more positive proof that the Gothic Bible could not have been written before the eighth century.

The Eusebian canon is marked on the inside of the page, the nuimber of each verse being enclosed in a calligraphic conventional ornamentation of this type

Heretofore the dating of the Gothic Bible has been determined by a vicious circle. They reasoned as follows: Several fragments of Gospels have been found at Bobbio and Milan, ergo they were written in Italy. If they were written in Italy, they must have been written before the year 552, when the Goths were driven out of the country. Now, the Codex Argenteus has external resemblances with the Codex Brixianus, hence it, too, must have been written in Italy before the year 552. Hence all Gothic documents were written in Italy, and all Gothic literary activity originated among the Ostrogoths. By such reasoning one could prove that all the Carolingian illuminated MSS. were written in Germany, or Italy, or elsewhere in the sixth century. But the Codex Argenteus was not found in Italy; of the learned Ostrogoth activity we know absolutely nothing, while Ulfilas was a Visigoth; we know positively that Gothic was understood in southern France in Carolingian times, and the Gothic calligraphy bears far more striking resemblances to that of the school of Tours. I have not yet a right to claim that I have proved the latter, but the theory of the Gothic scholars is irrevocably exploded, for it rests on the flimsiest of assumptions.

From Weissenburg comes the Codex Carolinus. It contains on four sheets the Epistle to the Romans in Gothic and Latin. Both are written stichdon, i.e., in lines representing clauses, without a separation between the words. The text is superscribed by passages from Isidor of Seville's Liber etymologiarum. Fortunately we possess a reproduction of one page. (43) The editor says that it seems to have been written in Spain. Schöne and Niebuhr (44) had assumed, without good reason, that the Codex Carolinus was in Bobbio calligraphy. However it may be, the dating of the Gothic text is gratuitous. It is quite true that, at first glance, one would identify the Latin column as of the fifth century, but one must again remember Traube's own statement that the Carolingian writers imitated fifth century books down to the minute details. The stichdon writing was by tradition used for the Epistles of St. Paul and did not die until the ninth century, and the writing of stichdon in a bilingual text is attested for the seventh century in the Codex Laudianus. Hence we must have another criterion for the establishment of the date of our MS. Fortunately the page reproduced tells its own story. The palimpsest contains a text from Isidor of Seville, hence it cannot be of a date earlier than the seventh century, and the use of thymologiae for "etymologiae" in the colon shows that it belongs to a much later date. Heinemann thinks that the writing is Visigothic of the eighth century, but there is no reason why it may not be of the ninth. The writing is cursive, but the title of the chapter "depurpureis" is in precisely the same handwriting as the underlying Latin text. If one compares the rounded d, s, and e, the open p and r with the original writing, the identity is immediately obvious. There is but one possible conclusion from this striking resemblance,---the underlying text is not much older than that of the palimpsest, nay, it may have been written by the same hand, and, as the superscribed text is not earlier than of the eighth century, the Gothic is not older than of the same period.

We have a number of Ambrosian Fragments of the Bible with Latin writing over them. One set of such fragments is contained in a quarto Codex of 214 pages, having for its superscription some homilies of Gregory the Great on Ezekiel which Castiglione estimated as of the eighth century. (45) Another Codex, of 156 pages, contains as a superscription St. Jerome's commentary on Isaiah, of the eighth or ninth century. (46) Here, again, there is nothing in the Gothic text to warrant any dating, hence it may be as late as of the ninth century. The remaining five pages of the Ambrosian fragments are apparently of the same date.

I have not touched upon the critical apparatus in all these fragments, because the fact that the Gothic is said to be based chiefly on early Greek sources, instead of the Vulgate, would equally apply to Carolingian times, when Joannes Scottus preferably quoted from the older Greek fathers, (47) and the Visigoth Theodulphus, whose Bibles bear a striking resemblance to the Codex Toletanus, (48) corrected the text in conformity with Hebrew, Greek, and Latin sources. (49) When Ximenes in the eleventh century introduced the Mozarabic Liturgy, he apparently carried out the decree of the Synod of Leon, by abandoning an older, freer Gothic tradition for one more in conformity with the Gallican custom, but that freer Gothic tradition was a survival of an older past which ultimately may go back to Ulfilas, but in the form in which it is preserved to us can represent only the influence of the Gothic writings, and for these we have not been able as yet to find a date previous to Carolingian times.

We now turn to the Skeireins, which will definitely settle the period of the Gothic writings. (50) It is assumed that the Skeireins, a polemical commentary on St. John, is based on that of Cyril of Alexandria (about 400) and that it was, therefore, written not earlier than in the middle of the fifth century (51) while Dietrich insists that the Skeireins may have quoted from the same source as did Cyril, that, therefore, it may still be the work of Ulfilas. (52) Were we to apply this reasoning to the Carolingian commentaries on St. John, we could prove, either that they appeared in the fifth century, or that they were composed by Ulfilas, for Cyril of Alexandria is one of the most frequently and most earnestly quoted authors in the ninth century. Alcuin quotes long passages from him; (53) Agobard refers to him as to a good Catholic; (54) Hincmar cites him. (55) Much is made of the fact by Böhmer (56) that, since the Skeireins is a polemic against Sabellius, who died in 260, and Marcellus of Ancyra, who died in 373, it must represent "an older stadium of the Arian controversy" than offered by these Bobbio fragments. We have already seen that Cyril was considered a good Catholic and that, therefore, his being quoted in the Skeireins precludes that, therefore, his being an Arian polemic. But let us waive this argument for a while, and let us see at what conclusions we shall arrive if the fact that Sabellius is quoted represents an older stadium of the Arian controversy. Alcuin quotes him by the side of Arius as a bad heretic; (57) Hincmar couples him with Arius as one of the two extreme heretics; (58) Joannes Scottus refers to the Sabellian error of confounding the natures of the Trinity. (59) If all that refers to an older stadium of the Arian controversy, then Alcuin, Hincmar, and Joannes Scottus were Arians, and their works must have appeared in the fifth century. It is obvious that the method pursued by those who made out the Skeireins to be an Arian controversy and placed it in the fifth century must be abandoned by a reductio ad absurdum.

It can be shown that the palaeographic proof of the antiquity of the text is based on no firmer foundation. The MSS. of the Skeireins fragments were found in Rome and in Milan and, like all the other fragments of the Ambrosian Library at Milan, came from the monastery of Bobbio, which was founded about 614. Much weight is put on the fact, to prove the Italian origin of the Gothic MSS., but the assumption is at once negatived from the fact that Bobbio and Milan possessed a large number of Spanish MSS. from Septimania, that is, Gothia, of the tenth century. (60) We are, therefore, prepared to find at Bobbio palaeographic documents of the Carolingian type, written by Visigoths. It can easily be shown that at least the parts of the Skeireins contained in the Vatican Codex 5750 were erased by a Visigoth to make place for a Latin text in the ninth century, (61) that, consequently, these parts of the Skeireins are not necessarily older than of the ninth century.

We fortunately possess an excellent reproduction of the whole Codex Vaticanus 5750, (62) which enables us accurately to locate the superscribed text. The Gothic text is contained on pp. 57-62, but it is necessary to discuss the condition of the whole Codex before ascertaining the age of the Gothic script. There are three distinct groups of handwritings to be discerned in the superscribed text, which in the Milan reproduction are given respectively as I, II, III. (63) I, a semi-uncial of the seventh or eighth century, (64) runs pp. 1-4, 13-56, 79-190, 211-274. III, a semi-cursive, of possibly the same date, runs from p. 5 to the middle of page 11. II ocupies half of p. 12, pp. 57-77, 191-210, 275-286. Here majuscule and minuscule letters are mixed. The open a is occasionally found (p. 77), but far more generally it is closed; both the straight and round d are used; e is round, with a horizontal line across; g has both arches open; i does not run under the line, but i-longa is common; m is rounded, occasionally turning the last stroke inwardly; n is sometimes rounded, but far more commonly the majuscule n is used, always in the ligature nt; both the long and the rounded s are used; t sometimes turns the vertical stroke to the right, but far more commonly it has the characteristic Carolingian abruptness; of ligatures we get nt, st, li; f and l are precisely of the form found in Spanish texts. The palaeographer can not help but recognise at a glance that the writing is of the end of the eighth century or of the ninth, and the use of i-longa proves conclusively that the writing could not be older than the eighth century (65) and is of the Visigothic or Beneventan school. In our text the following words, among others, are written with i-longa: In (p. 12), Ipsique (57), Iusserat (195), Interrogari, Iam (197), Incusatus (198), Iuxta (199), Ita (201). Unless the work of Loew can be overthrown, our text represents a Carolingian writing of a Visigothic type.




Endnotes

36. "Die karolingischen Abschriften, die so oft das antike Vorbild auch in allen Aeusserlichkeiten festhalten," L. Traube, Palœgraphische Forschungen, München 1904, p. 20 (Abh. d. k. Bayer. Akad. d. Wiss. III Kl., vol. XXIV, part I). Back

37. S. Berger, Histoire de la Vulgate, Paris 1893, p. 259 ff. Back

38. "On citerait difficilement un plus magnifique monument de la calligraphie du temps de Charlemagne. Nulle part ailleurs je n'ai vu de plus remarquables exemples régularité et de finesse d'écriture. Il n'y a point, à proprement parler, de peintures; mais l'emploi qu'on y a fait de l'or et de l'argent sur des fonds pourprés, l'élégance des inscriptions en grandes lettres enclavées, la pureté et la variété des encadrements de plusieurs pages et des médaillons réservés aux souscriptions finales, suffisent pour constituer une très belle décoration et pour augmenter encore la valeur de la bible, qui forme le plus précieux joyau du trésor de la cathédrale de Puy," L. Delisle, Les Bibles de Théodulphe, in Bibliothèque de l'Ecole des chartes, vol. XL, p. 8. Back

39. "On voit que les artistes employés par Théodulphe ont modifié les formes raides des miniaturistes primitifs, et, en employant l'or et l'argent, ils donnèrent plus de richesses et de reliefs à leurs lettres, qu'entouraient d'abord de simples traits rouges. Est-il étonnant que de tels maîtres aient laissé des préceptes, et que les moines de Fleury, qui ont executé de semblables beautés, aient voulu enseigner aux générations futures le secret de leur art?" Ch. Cuissard, Théodulphe, évêque d'Orléans, in Mémoires de la Societé archéologique et historique de l'Orléanais, vol. XXIV, p. 179. Back

40. Die Trierer Ada-Handschrift, Leipzig 1889, p. 69 f. Back

41. See the illustrations in the Trierer Ada-Handschrift. Back

42. See, for example, plate 11 in Trierer Ada-Handschrift, also plates 6, 7, 9, etc. Back

43. O. von Heinemann, Die Handschriften der herzoglichen Bibliothek zu Wolfenbüttel, Zweite Abth. v, p. 296. Back

44. Sitzungsberichte der Berliner Akademie der Wissenschaften 1902, p. 446 f. Back

45. Ulphilae partium ineditarum in ambrosianis palimpsestis ab Angelo Maio repertarum specimen coiunctis curis eiusdem Maii et Caroli Octavi Castillionaei editum, Mediolani 1819, p. xv. Back

46. Ibid., p. xvi. Back

47. "Sicut in Graeco legitur," Migne, vol. CXXII, col. 298; "nam quod in Graeco scriptum est," 299; "sed si quis intentus Graecum sermonem inspexerit," ibid.; "quod enim in Graeco scriptum est," ibid.; "vel ut in Graeco scribitur," 302; "in quibusdam codicibus Graecorum singulariter sinus patris dicitur, in quibusdam pluraliter," ibid.; "ut in Graeco significantius scribitur," 309; "in codicibus Graecorum anwqen legitur," 315; "sed in Graeco non est ambiguum," 319; and similarly cols. 283, 285, 287, 292, 295. Back

48. Cuissard, op. cit., p. 194 f. Back

49. "Quidquid ab haebreo stylus atticus atque latinus Sumpsit, in hoc totum codice, lector, habes," Carmina II. 1. Back

50. For the history of the text see Streitberg, Elementarbuch, p. 33 f. Back

51. E. Bernhardt, Vulfila oder die gotische Bibel, Halle 1875, p. 617. Back

52. Streitberg, Die gotische Bibel, p. xxx. Back

53. "Videamus quid beatus Cyrillus Alexandrinus episcopus......de hac inquistione senserit," Migne, vol. CI, col. 92 f.; "item beatus Cyrillus.....sic ait," 123; "item Cyrillus....inquit," 175; "tamen S. Cyrillus dicit in illo libello quem contra Theodoretum scripsit," 208 f.; "quidquid beatus Cyrillus Alexandrinae Ecclesiae pontifex synodali autoritate respondit Nestorio, vobis responsum esse absque dubio sciatis," 289. Back

54. "Inter Nestorium haereticum et Cyrillum catholicum," ibid., vol. CIV, col. 36; "ad quod beatus Cyrillus ita respondit," 40; "ait namque praecipuus ille expugnator Nestorianae impietatis doctissimus et beatissimus Cyrillus," 43, et passim. Back

55. Ibid., vol. CXXV, cols. 493, 588. Back

56. Streitberg, Elementarbuch, p. 35. Back

57. "Conticescat Sabellius audiens: 'Ego et Pater,' qui unam personam Patris et Filii prava doctrina disseruit; nam 'ego et Pater,' duae sunt personae. Item erubescat Arius audiens 'Unum sumus,' qui duas naturas in Patre et Filio astruit, dum 'unum' unam naturam significat, sicut 'sumus,' duas personas," Comment. in Joan. X. 29, in Migne, vol. C, col. 894, also col. 883. Back

58. "Quam multi de Trinitate contra Sabellium? quam multi de unitate Trinitatis adversus Arianos, Eunomianos, Macedonianos?" ibid., vol. CXXV, col. 482; "inter insidias horum latronum, Arianorum scilicet et Sabellianorum," 520; "ut beatus Augustinus in supradicto sermone de fide contra Sabellianos et Arianos," 551; "sicuti somnitant Sabelliani...ceu latrant Ariani," 589; "sicut impius Sabellius asseruit," 594; also cols. 567 and 598. Back

59. "Sabelliani quasi multivocum dicebant patrem et filium et spiritum sanctum....Hic videtur quasi tenebras incurrere et labi in errorem Sabellianorum," E. K. Rand, Johannes Scottus, München 1906, p. 39. Back

60. "Or nous avons quelque lieu de penser qu'il y a eu, d'autre part, entre la province ecclésiastique de Milan et la côte orientale de l'Espagne, quelque échange de textes bibliques. Des textes qui paraissent espagnols par leur origines ont été en usage, non seulement dans la Septimanie, mais dans la vallée du Rhône jusqu' à Vienne, et cela jusqu' au Xe siècle: il est fort possible que ces textes aient, eux aussi, franchi les Alpes et se soient mêlés à ceux qui, depuis les temps anciens, étaient en possession de l'autorité religieuse dans ce grand et riche pays," S. Berger, Histoire de la Vulgate, p. 410. Back

61. Thus determined by Massmann (Skeireins aiwaggeljons þairh Johannen, München 1834, p. 55). Reifferscheid (Die römischen Bibliotheken, in Sitzungsberichte der Wiener Akad. d. Wiss., vol LXIII, p. 618), without entering into a discussion of the problem, proclaimed it to be of the seventh or eighth century. Back

62. M. Cornelii Frontonis aliorumque reliquiae quae codice Vaticano 5750 rescripto continentur, Mediolani 1906. Back

63. The editor of the reproduction (p. 19 f.) makes two important mistakes in crediting p. 12 to III, though it is distinctly in the handwriting of II, and in crediting 77-195 to I, although 77, 191-195 are distinctly of the hand II, while 78 is not superscribed. Back

64. Thus determined by the editor (p. 21). As I am not studying this text, I do not vouch for the date. Back

65. "If we consider on the one hand the utter absence of i-longa in the oldest Latin MSS. in uncial and semi-uncial, and its gradual and tentative entrance only into uncial and semi-uncial MSS. of the recent type, i.e., of the 8^th and 9^th centuries," E. A. Loew, Studia palaeographica, in Sitzungsberichte d. k. Bayerischen Akad. d. Wiss., München 1910, p. 4. Back