Early Journal Content on JSTOR, Free to Anyone in the World This article is one of nearly 500,000 scholarly works digitized and made freely available to everyone in the world by JSTOR. Known as the Early Journal Content, this set of works include research articles, news, letters, and other writings published in more than 200 of the oldest leading academic journals. The works date from the mid-seventeenth to the early twentieth centuries. We encourage people to read and share the Early Journal Content openly and to tell others that this resource exists. People may post this content online or redistribute in any way for non-commercial purposes. Read more about Early Journal Content at http://about.jstor.org/participate-jstor/individuals/early- journal-content . JSTOR is a digital library of academic journals, books, and primary source objects. JSTOR helps people discover, use, and build upon a wide range of content through a powerful research and teaching platform, and preserves this content for future generations. JSTOR is part of ITHAKA, a not-for-profit organization that also includes Ithaka S+R and Portico. For more information about JSTOR, please contact support@jstor.org. A NEW CLUE TO THE EMENDATION OF LATIN TEXTS By W. M. Lindsay Everyone knows that minuscule, even early minuscule Latin MSS written in Ireland or England, or at Continental monasteries where Irish or English script was practiced, are full of abbreviation- symbols, such as n, "non," p, "prae," ee, "esse," tm, "tantum," tn, " tamen," oo,j" contra," also that those written in other Continental scriptoriums use abbreviation too, but not to the same extent. For example, of these half-dozen symbols only the first three would be current in Continental script of, let us say, the ninth century. And everyone believes that all ancient MSS (let us say of the fifth century) wrote words in full and eschewed abbreviation. It does not seem to play a part in Latin MSS until minuscule writing comes into vogue (in the eighth century). Only one class of ancient MSS is recognized as an exception, namely legal MSS (before Justinian's famous edict). The fifth- century uncial codex of Gaius at Verona is crammed with abbrevia- tion-symbols, such as n, "non," p, "prae," ee, "esse," qd, "quidem," tm, "tamen," qa, "quia," et, "etiam," Ig, "igitur," qu, "quamvis." These are partly the same as the minuscule symbols, partly different. For example, qd would not denote "quidem" in a minuscule MS, but "quod." The accepted theory of Latin abbreviation fails to account for the difference. It declares the eighth-century minuscule scribes to have revived these old symbols of fifth-century legal writing, symbols that had remained in abeyance for some two hundred years. But if this "resuscitation" theory be true, why should qd, which in the fifth century denoted "quidem," denote "quod" in the eighth? A new theory was broached in Zentralbl. f. Bibliothekswesen, XXIX, 56, that abbreviation had never really fallen into abeyance, but was practiced continuously in non-calligraphic script, developing and altering from one generation to another; that these symbols in the Verona Gaius were not really restricted to legal MSS and should not [Classical Philology XI, July, 1916] 270 New Clue to Emendation of Latin Texts 271 be called "Notae Juris" but rather "ancient Notae." This "con- tinuity " theory was confirmed by the discovery in Egypt of papyrus fragments containing speeches of Cicero written with these ancient Notae (n, "non," ee, "esse," qd, "quidem," tm, "tamen," lb, "tibi," Ig, "igitur," and so on). A new avenue was opened for the emenda- tion of texts. Till this discovery was made it had been thought that it was only editors of legal texts who had to reckon with the possibility of errors caused by abbreviation-symbols of the pre-minuscule period. The Cicero papyrus proved the contrary. Suppose it had fallen into the hands of ninth-century transcribers. They would have had no difficulty in interpreting rightly some of the symbols (such as n, "non," ee, "esse") which had remained in continuous use down to their own day. But they would have stumbled at others (such as ib, "tibi," ig, "igitur") which had gone out of fashion long before. And they would have been led into error by a few (such as qd, "qui- dem," tin, "tamen") whose meaning had changed in the interval (qd now denoting "quod," not "quidem," and tin, "tantum," not "tamen"). The plate which accompanies this article exhibits another example of an ancient (fifth- or sixth-century?) non-calligraphic MS which used these ancient Notae, a MS of the grammarian Marius Victorinus. The plate represents (in natural size) fol. 4 V of a MS in the Vatican Library, Pal. lat. 1753. Not that the ancient MS itself is there. What we actually have is a transcript made at Lorsch not later than the first half of the ninth century. But the ancient exemplar is reflected in the tran- script as in a mirror. The Lorsch transcribers have (more or less thoroughly) transferred the abbreviation-symbols bodily from the exemplar into their transcript, and so have enabled us almost to see the exemplar with our own eyes, without requiring the aid of imagination. This precious relic of the ancient book-world we owe, not so much to the conscientiousness of the transcribers, as to the nature of the text they were transcribing. An early section (from which the plate is taken) deals with spelling. The several letters of the alphabet come under mention. A ninth-century transcriber would soon realize the impossibility of picking his way through the obstacles 272 W. M. Lindsay offered by these letters on the one hand and the puzzlingly similar abbreviation-symbols on the other. Look at the sentence (a little below the middle of the page) on the spelling with C and with QU: "nam concussus QU a quatio habeat originem .... at tamen per C quam per Q scribuntur" (Keil, Gram. laL, VI, p. 13, 11. 21-23). How could an ignorant scribe distinguish QU, the abbreviation- symbol of "quam vis," from QU the letter-group? It is no wonder that, although he attempted to grapple with these symbols in the first page or two (misrendering the ancient "quod" symbol in p. 3, 1. 9K., as "quia"), he soon gave up the struggle and resigned himself to making exact copies of them, as they stood in the exemplar, without troubling his head over their meaning. The exemplar seems to have been an uncial MS, for the symbol tin ("tamen") in the sentence just quoted is written in uncial, not minuscule, letters. Anyone can see the difference between its m (uncial m) and the other ra's on the page. And the qu ("quamvis ") has uncial lettering too, although less easily recognized; for the minuscule on this page is fairly large and regular, not far removed from the half-uncial and uncial types. The transcription of the exemplar was divided among a number of monks. Some of them use much smaller and less regular script; and in their pages a transferred symbol in its uncial lettering stands out far more clearly from the context. Others again are less chary of interpreting the unfamiliar signs. There is (or was) in the Valenciennes Library a MS copied directly from the Lorsch (now Vatican) MS. It was written in the St. Amand scriptorium, and a few of its readings are given in KeiPs critical apparatus. One or two of the ancient Notae have actually survived into this "second remove" from the archetype, especially on the pages where they stood in clear contrast to the neighboring words. But on such pages as that represented in the plate the St. Amand scribe could hardly be expected to realize that he had to deal with anything else than ordinary Lorsch abbreviation, and must be forgiven for his attempts to interpret it. More important is another ninth-century transcript (more prob- ably mediate than immediate) from the ancient exemplar. It is a Paris MS, marked B in Keil's critical apparatus, while the Lorsch transcript is marked A. Where it was written is not known, but we New Clue to Emendation of Latin Texts 273 can guess that the scriptorium (Corbie?) was not so thoroughly under Insular (English or Irish) influences as the Lorsch scriptorium (where Anglo-Saxon script was still practiced at this time). For the ancient Nota of "quod," namely q with an oblique cross-stroke, would denote in Insular script "quia," but in (some) Continental scripts "qui." In the Paris MS we sometimes find the error "qui" whereas the Lorsch transcribers regularly write "quia," as often as they are not content to leave the symbol as they found it. The text of Marius Victorinus, derived from a single archetype which was permeated with this fertile causa erroris — a partially obsolete system of abbreviation — is clearly a case where paleography can help textual criticism. Paleography has already provided a list of all known ancient Notae (see the Index to Studemund's apograph of the Verona Gaius, and my additions in Melanges Chatelain) and a list of all symbols current 1 in (1) Insular, (2) Continental script of the eighth and ninth centuries (see my Notae Latinae). The Vatican MS, however, makes everything easy. It is the key to the gate between the (nearly) correct text of the ancient exemplar and the corrupt text of the Paris transcript. When we have collected all its symbols in uncial lettering we find ourselves provided with a passably full list of the ancient Notae employed by the writer of the exemplar. One or two more can be guessed with fair certainty from mistakes of the Paris or the Vatican MS or both. A rough list will be found on p. xiv ("Additions and Corrections ") of my Notae Latinae. Here I will mention only the symbols most likely to mislead a ninth- century transcriber and will put them in capitals, omitting the abbreviation-stroke above (partly for convenience of printing, partly because ancient scribes often omitted it and contented themselves with a dot on each or one side of the symbol) : AT, "a(u)-t(em)." At the beginning of the ninth century this symbol was known only 2 at a few centers of English script. The 1 Something still remains to be done. Separate lists of the symbols current at Lorsch , at St. Amand, and the other writing-centers would be very handy. 8 1 would now print on p. 13 of Notae Latinae, instead of "and also still shows itself," rather "Where it shows itself .... it may come from an Anglo-Saxon exemplar or one with ancient Notae." In the sentence cited at the beginning of § 14 the Monte Cassino scribe (who found the symbol in his Fulda exemplar) probably meant ater, not autem: "Livy was singularly black and beardless" (!). The Fulda scribe may have meant the same. 274 W. M. Lindsay ordinary transcriber would interpret it either as ater (for t with abbreviation-stroke above denoted "ter") or out (for u was often expressed by a suprascript stroke exactly like the abbreviation-stroke) or at (see below on the "etiam" symbol). 2 (like a large "two" or a form of capital Q), "contra." Since q denoted quae (in Insular script mainly) this would be the natural expansion of the strange symbol by a ninth-century transcriber; though he might, if the sense seemed to require it, substitute some other small word beginning with q. He would never dream of contra. ET, "e-t(iam)." An Irish transcriber might interpret this symbol rightly; others would be likely to take it for et with an apex (or sign of long quantity, a sign often shaped exactly as the abbrevia- tion-stroke) added above. This addition was not uncommon with monosyllables, even with short vowel, e.g., an. I with oblique cross-stroke, "inter." Irish transcribers were quite familiar with the symbol. To others it might appear an unusual symbol of in, but would rather suggest an i which had been written in error and obliterated. They would be likely to omit it in their transcript. Among the ancient Notae were a number of symbols of this misleading shape, such as D (with transverse stroke) for any part of the verb dico, II (with the same) for any case of the noun res. To indicate the plural these symbols were doubled, DD, RR. They would be productive of omissions in transcripts. QU, "q(uam)-v(is)." Absolutely tmintelligible to ninth-century transcribers. They would interpret it as "qum," understanding the stroke above the u to be the ra-stroke, and would put cum (the usual spelling of the word) in their transcript. QS, "q(ua)-s(i)." An Irish (or Insular?) scribe (who expressed the word by "qsi") might interpret this rightly. The scribe of the Paris MS seems to have fallen victim to the delusion that it meant quibus. QA, "q(ui)-a." Absolutely unfamiliar in the ninth century, when it would suggest rather qua than quam. But a lucky guess at quia would not be impossible; for it has the form of a "contraction" (the final letter being part of the symbol), and most mediaeval symbols were "contractions" (as opposed to "suspensions"). New Clue to Emendation of Latin Texts 275 QD, "q(ui)-d(em)." Would mean quod in the ninth century, except to Irish scribes, who had a symbol of their own for quod. Many English scribes used both the Irish sign and the other. TM, "t(a)-m(en)." In the ninth century this was the Insular symbol of tantum. Continental scribes, if quite unaware of the Insular practice, might interpret it as turn, taking the abbreviation- stroke for a suprascript (conventional) u. The list might be extended, but the purpose of this article is to stimulate rather than to carry out investigation. Provided with these nine examples alone, anyone who takes the trouble of reading through a few pages of KeiPs text and critical apparatus will see how wonderfully this clue guides him through the labyrinth. Nine-tenths of the manuscripts' corrupt readings he finds to be misinterpretations of ancient Notae. The Paris MS reads with fatal consistency at for autem, et for etiam, cum for quamvis, and so on. And he can improve KeiPs text on page after page: e.g., on p. 4, 1. 5, add, not eorum, but rerum (RR with transverse stroke); on p. 5, 1. 6, read not quia but quod (KeiPs account of the reading of A is imperfect here); on p. 11, 1. 16, not sed et but sed etiam, and so on. If all the ancient Notae of the exemplar had been interpreted rightly in the extant MSS the text would hardly require editing. There is only a single gate that shuts off the corrupt text from the true. And the Vatican MS is the key that opens the gate. Will not some student of the American School at Rome take for his thesis an investigation of the Vatican MS and publish (1) a fuller list of its ancient Notae than my brief examination in the Easter holidays of last year allowed; (2) details of the corrupt readings due to misinterpretation of them? There must be other Latin authors (perhaps chiefly grammarians) the tradition of whose text has been like that of the text of Marius Victorinus, but for whom a key-codex, like the Vatican MS, is wanting. In editing them we are under the same difficulty as we should be if we had only KeiPs B, not his A, to help us. These details would be useful sign-posts to guide an editor. A thesis by a former student of the school, Shipley, Certain Sources of Corruption in Latin MSS (Macmillan, New York, 1904), contains a study of another Vatican MS, Reg. lat. 762, and has proved of great value to editors. The MS was a Tours ninth-century 276 W. M. Lindsay transcript of an uncial MS of Livy (the Codex Puteanus, now at Paris) ; and the booklet explains in detail the various errors to which the ninth-century transcribers of this fifth-century uncial exemplar (without ancient Notae) show themselves liable. We all realize nowadays the necessity of paleographical knowledge for Latin scholars. Textual emendation at haphazard is going out of fashion. It is true that discoveries in the Fayoum have here and there con- firmed a conjecture which previously had no stronger claim than half-a-dozen rivals. But we may say of conjectural emendations what Cicero says of dreams : Out of so many, what wonder if one or two become true ? Seeing that texts took their corruptions mainly from the hands of mediaeval scribes, common-sense tells an editor that his first duty is to acquaint himself with these obvious inlets of error. Most editors have neither time nor opportunity for a minute study of mediaeval paleography and must look for help to manuals like Shipley's. It shows us how errors crept into the transcription of uncial exemplars, especially of historical texts. The thesis now sug- gested would do the same for grammatical texts and perhaps others. For we may fairly assume that technical treatises — on law, grammar, rhetoric, etc. — would be the most suitable for the non- calligraphic form of publication, in which economy of material was secured by a free use of ancient Notae. Of the wonderful collection of grammatical works amassed by Columban in the library of Bobbio at the beginning of the seventh century, many would in all probability be so written. They are often the archetypes of all existing MSS and have disappeared after being transcribed. A theological work, perhaps by Columban himself, a commentary on the Psalms, seems to have been written 1 with pre-minuscule Notae (see p. 70 of my Early Irish Minuscule Script, Parker, Oxford, 1910). And a St. Bertin transcript of Augustine's Letters appears to come from a similar exemplar (ibid.). Perhaps another grammatical instance is the exemplar of Berne 207. But a re-examination of this MS is neces- sary before one can be sure. Marx's paleographical account (hardly 1 A copy made for private use would be likely to be non-calligraphic. The subscrip- tion at the end of the Marius Victorinus exemplar : ' 'FELICITER UTERE STEPH ANE SCRIPTOR ET LECTOR," suggests at first sight that Stephanus wrote it for his own use. But a comparison of the subscription at the end of Book II makes the suggestion less attractive. tt^cen^l P**' -tyrm^TLeU^utU+ixbl utrc|ii a2>t< *■*** c ° n ^^ * *f ****»'H *r^« f^fEbp e,t% - J3 1 * ° a * t " 1 [n** arf^cw/T^.rcrMpcumerr^ %£t*v H^ • f>t-ttcur^u can xe^ cufur j?^«©fctr lrtr**Y-<*r rcrikeLu~P , £>*&rr^~, arret* *»> f=uer-tr uU>uer-&ftt-c©m pofM^» P ^ u f^w^Mnum^fTctefcuo fc>^k«»jc^oe*ttr . M^cc»uic r\or\pxr> *&' Mr ^ji ef^locirttij- czXo cfu^n'oo , f^ctmir u.reo pc| porta/- U^Aectxr om^ti^^i .dc'coccif *^nk^J^ -a ^^. L. ^ ^tl . [^-*^pc3T^ 1 ^« m a^bMor^^^6V'f q-^fo^a.^/^. New Clue to Emendation of Latin Texts 277 satisfactory to paleographers) of the archetype of the Rhetorica ad Herennium (see the preface to his edition) suggests the possibility of a rhetorical example. But here too re-examination of the evidence is necessary. Mommsen infers from certain corruptions of the text an exemplar of this kind for the Vegetius-portion (Hermes, I, 130) and the Prosper-portion (Chronica Minora, I, 371) of that uncial MS in the Vatican Library (Reg. lat. 2077) which contains the palimpsest Cicero Verrines fragments. The investigation proposed above will also improve both KeiPs apparatus for the text and the text itself. And although Apthonius' work on Metre, which has been incorporated with Victorinus' grammar (pp. 31 med. — 173K.), is rather dull, the preceding pages are of great interest to students of the Latin language. They seem to be notes taken from Victorinus' lectures (see American Journal of Philology of this year). But the chief result of the investi- gation will be to provide a new clue to emendation, especially of grammatical texts. St. Andrews, Scotland March, 1916