Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Hyphens in Greek Manuscripts Murphy, David J Greek, Roman and Byzantine Studies; Autumn 1995; 36, 3; ProQuest pg. 293 Hyphens in Greek Manuscripts David]. Murphy WORKS ON LATIN PALEOGRAPHY assert that somewhere be­ tween the eighth and tenth centuries, hyphens began to be placed in the right margin to indicate that a word had been divided between syllables when the end of a line had been reached, 1 and this practice seems to have become com- t The following works will be cited by author's name, abbreviation, or short tide: BEEs=N. Bees, Ta XEtpbypacpa trov MEtEropcov I (Athens 1967); BERNARDINELLO=S. Bernardinello, Autograft greci e greco-latini in Occidente (Padua 1979); BICK=1. Bick, Die Schreiber der Wiener griechischen Hand­ schriften (Vienna 1920); CG V =E. Follieri, Codices graeci Bibliothecae Vaticanae selecti (The Vatican 1969); CL=P. Franchi de' Cavalieri and 1. Lietz- mann, Specimina codicum graecorum Vaticanorum 2 (Berlin 1929); CLA=E. A. Lowe, Codices Latini Antiquiores (Pts 1-12 [1934-71], Pt. 22 [Oxford 1972]); CONSTANTINIDES=C. Constantinides and R. Browning, Dated Greek Manu- scripts frol1yCyprus to the Year 1570 (Washington 1993); DOLGER=F. Dolger, Aus den Schatzkammern des heiligen Berges (Munich 1948); GHA = Griechische Handschriften und Aldinen. Eine Ausstellung einlaftlich der XV. Tagung der Mommsen-Gesellschaft in der Herzog August Bibliothek Wolfenbuttel (Wolfenbiittel 1978); HARLFINGER=D. Harlfinger, Specimina griechischer Kopisten der Renaissance. I. Griechen des 15. Jahrhunderts (Berlin 1974); KOMINES= A. Komines, lltvalC£l; XPOVOAo"(11j.l£VCOV llatj.llatcrov lCCOSttcCOV (Athens 1968); LAKE=K. and S. Lake, Dated Greek Minuscule Manuscripts to the Year 1200 I-X (Boston 1934-39); MARTINI and BASSI =A. Martini and D. Bassi, Catalogus codicum graecorum Bibliothecae Ambrosianae I-II (Milan 1906); OMONT (1887}=H. o mont, Fac-similes de manuscrits grecs des XV' et XVI' siecles (Paris 1887); OMONT (1891)=H. Omont, Fac-similes des manuscrits grecs dates de la Bibliotheque Nationale du IX' au XIV' siecle (Paris 1891); RGK=E. Gamillscheg and D. Harlfinger, Repertorium der griechischen Kopisten. 800-1600. I. Handschriften aus Bibliotheken Groftbritanniens (Vienna 1981); II. Handschriften aus Bibliotheken Frankreichs (Vienna 1989); Scritture=G. Cavallo, G. De Gregorio, and M. Maniaci, edd., Scritture, libri e testi nelle aree provinciali eli Bisanzio (=Atti del seminario di Erice, 18-25 settembre 1988 [Spoleto 1991]); VG=M. Vogel and V. Gardthausen, Die griechischen Schreiber des Mittelalters und der Renaissance (=ZbBibl Beiheft 33 [Leipzig 1909]); Xeropotamou=1. Bompaire, Actes de Xeropotamou. Archives de l'Athos III (Paris 1964). 293 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 294 HYPHENS IN GREEK MANUSCRIPTS mon by the twelfth century.2 No satisfactory treatment, however, has been granted to this marginal stroke as it was used by copyists of Greek manuscripts.3 In this paper I shall catalog the different shapes of marginal hyphens, or, to use E. M. Thompson's phrase, "coupling strokes," (in other languages, tratto d'unione, Verbindungsstrich, etc.) as Greek scribes used them, along with the earliest instance I have found of each shape. I hope thereby to construct a rough schema that may be helpful for dating manuscripts. I shall also give reasons why I believe that Greek scribes did not borrow the marginal hyphen from Western practice but rather developed it from a much older mark of Greek prosody, the sublinear hyphen, which we know from ancient grammarians and from papyri. Copyists had long made it a rule to divide words at syllabic breaks when they needed to carry part of a word over to the next line of writing, although they often kept double con­ sonants together in neglect of correct principles of syllabifi­ cation.4 By the twelfth century, they began to place a stroke in the margin to link syllables when a word had been divided by line end. This marginal coupling stroke coexisted alongside the sublinear hyphen in Greek manuscripts into the Renasissance, appearing in diverse forms and positions on the page. Although our focus falls on the marginal hyphen or coupling stroke, it will be necessary to look at the sublinear hyphen as well, for the marginal stroke evolved from it. 2 Cf E. A. Lowe, The Beneventan Script2 I (Rome 1980) 277f; J. Vezin, ·Paleographie et codicologie," AnnEcPratHEt IV" Sect. 110 (1977-78) 590f; B. Bischoff, Latin Palaeography. Antiquity and the Middle Ages, tr. D. Cr6infn and D. Ganz (Cambridge 1990) 171; P. Saenger, ·Coupure et separation des mots sur Ie Continent au Moyen Age," in H.-J. Martin and J. Vezin, edd., Mise en page et mise en texte du livre manuscrit (Mayenne 1990) 451-55, esp. 453. ) Sketchy or inaccurate: W. Wattenbach, Anleitung zur griechischen Palae­ ographie 2 (Leipzig 1877) 35; V. Gardthausen, Griechische Palaeographie 2 II (Leipzig 1913) 407; E. M. Thompson, Introduction to Greek and Latin Palae­ ography (Oxford 1912) 58. 4 F. Bast, Commentatio Palaeographica, in G. Schaefer, ed., Gregorii Cor­ inthii '" De Dialectis Linguae graecae (Leipzig 1811) 859; Wattenbach (supra 0.3) 35. Occasionally scribes put coupling strokes before yap. Oi. 'tI::. etc., as though they were parts of a divided word. Cf. New York Union Theological Seminary 69, a Gospels of ca 1340, probably written in southern Italy: f. 7v yiYPalt'tat -Iya.p; 54r O'uvllYllivCJ)v -10£; etc. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. DAVID J. MURPHY 295 The taxonomy of hyphens outlined below was constructed from examination of almost 800 Greek Mss.5 and some 5,000 published plates. The latter included plates of hundreds of papyri and parchment or paper documents. For hyphens in Latin Mss., I relied principally on published studies, but I also inspected a wide range of published plates as well as sixteen Mss., including eight Latin-Greek bilingual codices, and one six­ teenth-century Italian Ms. Although early Greek type fonts are a different topic, I did inspect various early printed editions. Photographs impose limitations, but for the great majority of microfilms, one can determine from the cut of the pen, the darkness of the ink, and the nature of the writing whether the hyphens were written by the hand that wrote the main text. In several films of Ambrosian Mss., for example, it is clear from a comparison of the hyphens with corrections made in the text that a corrector added the hyphens while correcting. Still, in the case of a few Mss. that I could not consult in situ, some skept­ icism must remain about the source of certain marks in their margIns. I divided each Ms. studied into four sections of fifteen or more folia each, spacing them at equal intervals. In this way 5 By autopsy, 143 Mss. and three documents in these libraries: Vatican, Biblioteca Nazionale in Naples, Beinecke (Yale), Yale Medical Library, Pierpont Morgan, New York Public, and the libraries of Columbia Univer­ sity, Princeton University, Princeton University Art Museum, Princeton Theo­ logical Seminary, and Union Theological Seminary. On film, 640 Mss. from New York Public, Pierpont Morgan, Princeton University, and from four microfilm collections: the Ambrosiana Collection at the the Medieval Institute at the University of Notre Dame (South Bend, Indiana); Knights of Columbus Vatican Film Library at St Louis University; United States Library of Congress, Mss. microfilmed at the Greek Orthodox Patriarachate in Jerusalem and at monasteries at Mt Sinai and on Mt Athos; Plato Microfilm Archive at Yale University. Included in this figure are also microfilms of Vind. suppL gr. 7 and Tub. Mb 14, supplied by their owner libraries, and a microfiche of Messan. gr. 3, supplied by the Institut de Recherche et d'Histoire des Textes. Finally, I examined eleven facsimile editions. Of the Mss. discussed here, I have inspected these by autopsy: Neap. III.E.15, VI.E.43; New Haven: Beinecke 139,251,424, Marston 94, and Yale . Medical Library 35; New York Columbia: Plimpton 3,5, 12, Additional 6, and Smith Western Additional 10; New York Pierpont Morgan 340, 652; New York Public Library 59; New York Union Theological Seminary 59; Princeton: Garret 4, 12, Greek 3, and Scheide 70; Vaticani: Vat gr. 64, Barb. gr. 541, Ottob. gr. 376, 381, Pal. gr. 232. I have also worked carefully in situ on Oxon. Bodl. Clarke 39 and Marc. gr. append. c/. IV.l during previous studies. I am grateful to Ms Sarah Kozodoy for inspecting lerus. Staurou 43 and T_phou 46,57 in Jerusalem. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 296 HYPHENS IN GREEK MANUSCRIPTS some 120 pages would be scrutinized. For very long Mss. I looked at five sections and inspected the entire Ms. when it was relatively short, when numerous scribes had collaborated, or when there was doubt about whether repeated marks were really hyphens. In the case of strokes in the left margin, one needs to examine the pattern of their use throughout the codex to be sure that they are not paragraphoi, obe/oi, or quotation marks.6 In addition, because many stray marks appear in the margins of Mss., it is necessary to observe their placement over a range of pages in order to see whether they consistently divide words and cannot owe their presence to some other cause. The Sublinear Hypen We know from inscriptions, from Aristotle and Isocrates, and from early papyri that Greek possessed marks of punctuation before the Alexandrian period, but it is only with the produc­ tion by Alexandrian scholars of critical texts of poets that we know of systems of punctuation and prosodical marks, i.e., organized patterns of marks placed above, around, or below the letters to assist the reader in making out the sense of the text and in reading it aloud.? To the grammarian Dionysius Thrax (2nd c. B.C.) is attributed a catalogue of ten prosodical marks, among which are the so-called 7ta8Tl: the apostrophe, the 6 For this last mark, cf the wavy, horizontal strokes in the right margin of Glasg. Hunt. 408 (V.3.6), copied in 899 (RGK I.C 148). In Ambr. C 178 inf., copied 1324, a wavy stroke in the left margin indicates a quotation from Scripture, while double apostrophes indicate quotations from Gregory of Nazianzus. On the other hand, a double apostrophe indicates a scriptural quotation in Ambr. C 186 inf., copied on the island of Chalce in 1073, and words of Jesus are denoted by a straight stroke in the left margin in Ambr. B 56 sup. (cf below on Stage 5). On diacritical marks in margins of papyri, cf K. McNamee, Sigla and Select Marginalia in Greek Literary Papyri (=Papyrologica Bruxellensia 26 [Brussels 1992]). 7 Cf. W. Flock, De graecorum interpunctionibus (diss.Bonn 1908) 4-9; Thompson (supra n.3) 6lf; Gardthausen (supra n.3) 382-414; R. Pfeiffer, History of Classical Scholarship I (Oxford 1968) 178ff; J. Moreau-Marechal, -Recherches sur la ponctuation," Scriptorium 22 (1968) 56-66; E. G. Turner, Greek Manuscripts of the Ancient World2 (Oxford 1987) 8-13; P. Rafti, - L'lnterpunzione nellibro manoscritto: Mezzo secolo di studio," Scriuura e Ci1Jilita 12 (1988) 239-98. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. DAVID J. MURPHY 297 (hypo )diastole, and the hyphen, i) ucp£v. 8 This word, coined from the phrase ucp' tV ("into one" or "together"), is feminine because it understands 1tpo(Jcp~ia. as its substantive. The hyphen was used to show the reader that two words normally pronounced separately were to be spoken together as one.9 The need for the hyphen, as well as for the diastole, arose from the difficulty of reading scriptio continua. Ancient Greek and Latin grammarians repeat the same definition and often the same examples. tO One cannot prove, however, that a grammarian invented the sub linear hyphen, for scribes may have already used such strokes on their own volition.tt Although the hyphen should be placed most properly beneath words composed of two parts that are themselves com­ plete or virtually complete words (e.g. lCEAagEcp£l, 1taAly:' a'YPE'tov, VECPEAT\'YEp£'ta, etc. ), it is frequently used beneath com­ pound numbers or words that bear a prefix: cf e.g. ~'Uro~£­ lCa'tT\, 1tR9..1tav, oil1t~1to1:£, in Marc. gr. 454 (s.X), Ms. A of Homer's Iliad. In the medieval world, hyphens would often be placed underneath words that do not truly demand them: e.g. Leon of Reggio, who copied Vat. gr. 1926 in 1124-25, displays a special love of sublinear hyphens, for in addition to putting them on genuinely compound words, he also uses them with words that have prefixes derived from reduplication ('tE'ti9llpOC;, 65v) or with suffixed words (OfglJ.l.£vo'Uc;, 91v; ~llC~(Ji>VT\C;, 99r). 8 The n:aEh} indicate the effect that sounds undergo when they are intoned in the way demanded by the sign; cf. A. Hilgard, Grammatici Graeci 1.3 (Leipzig 1901) 145. Some grammarians did not include them among the prosodical marks or admitted them only as -improperly so-called"; cf. A. Wouters, The Grammatical Papyri from Graeco-Roman Egypt. Contribu­ tions to the Study of the 'Ars Grammatica' in Antiquity (Brussels 1979) 190-97. 9 i1 :Ie uc:pev 'ti8£'tttl. (hav :luo Ail;£1~ tV 'tep ii~a O1CA.10V, 'tl?lXA.£Upa, i~1C£A.£~, etc. Sublinear hyphens also appear in New-Haven Beinecke 139, a Divine Office of s. XV, and in Beinecke 251, copied on Mt Athos in 1585. This mark still appears in modern Greek and is called 'to i>q>EV. The Marginal Hyphen or Coupling Stroke It should be dear that the function of the marginal hyphen or coupling stroke is essentially the same as that of the Alexandrian sublinear mark: to join "in one" (i>q>' EV) syllabic units that a reader may not realize are parts of one and the same word. What differs is the mark's shape and position on the page as well as the coupling stroke's use on divided words that are not compounds. Numerous scribes during s. XIII-XIV use both sublinear hyphens and marginal coupling strokes. But what I want to argue is that the marginal stroke is an outgrowth of the sublinear hyphen. How it arose from the sublinear hyphen can be seen if we layout its developmental stages. Stage 1 comprises the sublinear hyphen discussed above. When the page layout causes the junction of the two parts of a compound word to coincide with the end of a line of writing, the hyphen then appears below the first or the last letter of a line. Were the word in question to fall in the middle of a line, the hyphen would still be placed under the same letters. Although it is occasioned only by chance, I call this position Stage 2, for it represents a step closer to appearance in the margin itself. Examples occur frequently.14 In the twelfth cen­ tury, paired sub linear hyphens begin to appear, i.e., a hyphen is placed on a divided word at the end of the line and again at the beginning of the next line. My earliest example is on fo1. 123v of Vat. gr. 1992 of Gregory of N azianzus, copied in 1104 by Barto­ lomeo da Simeri: Q!\1~v6't\..15 Cf also fol. 191v in Scribe B's 14 E.g. in Paris A of Plato, most of them boat-shaped (fol. Sr 1l!2~JQ.1ta8dc;. These paired sublinear hyphens anticipate paired marginal hyphens (cf Stage 5 below). In Stage 3, the hyphen appears below the line of writing but intrudes into the margin. I have seen only boat-shaped hyphens written in this way. The scribe of Paris A of Plato extends some of his line-end hyphens into the margin, as on fo1. 6v <> 8paa1i, IllaxoC;, or on 13r TtIl!..JIlOX81lpol, where the hyphen bends up in a little hook on the right. Other early examples are New York Pierpont Morgan 652, a deluxe copy of Dioscurides copied ca 890, fo1. 294v En:PQ)YEvrc;; and Ambr. B 106 sup. of 966-67, fo1. 98v in the scholion: 'to ~l-l'tOq>ropo>.17 As a sublinear hyphen would belong underneath the omicron-phi, it is clear that the scribe was not trying to write that traditional mark; he was signalling instead that a divided word continued on the next line. Thus, a true coupling stroke has emerged. The Ms. carries one other, a small, straight stroke on f. 5v: Ouva-I'tal (cf PLATE 2). In the same century, patterns of coupling strokes begin to appear in the margins of a few codices. The earliest dated Ms. in which I have found a pattern of marginal hyphens is J erus. Taphou 57, a collection of patristic writings copied by Gera­ simos of the Nea Mone Monastery on Chios ca 1182.18 In the left margin, in the hand of the first scribe, hyphens appear consistently on proper names: e.g. f. 4r aJ.LI-q>lA,0XLOV; 8r ot am) CJE'\)", I-po'\); 30v xpo I -lCAo'\)' £q>palJ.L ... etc. The last hyphenated name I have seen in the codex appears on f. 93v xpo<; CJou I - lCEVCJOV, although Gerasimos continues the codex as far as his subscription on f. 287v. As scribes often do, Gerasimos seems to have dropped an element of his system of punctuation as he proceeded through the codex (cf PLATE 3). Other twelfth-cen­ tury Mss. with hyphens in the left margin are: Jerus. Saba 140, Lives of the Saints, a hyphen about every five folia; Jerus. Staurou 47, Discourses of Symeon, a hyphen about every ten folia, all but two on proper names; Princeton Garrett 4, a Gospels, a hyphen every five or six folia; New York Pierpont 17 Cf. A. T. Papadopoulos-Kerameus, 'h:POOOA:uj.lltlri! ~l~A.loEnl1CTl ~'tOl KU'ta­ A.o10C; 'troY tV 'tu'ic; ~l~A.loei!KUlC; ... EA.A.llVllCrov KrolHKrov I (St Petersburg 1891) 126. The catalogue'S twelfth-century date is consistent with the Ms.'s light brown ink, small accents, mixed styles of breathing marks, degree of evolution from the canon of Perlschrift, and the letter forms, all of which find counter­ parts in Mss. of s. XII or even earlier, although distinctive later forms are absent. The crowding of the lines of writing and the awkwardness of the hand suggest a provincial origin. Sarah Kozodoy reports that the hyphens in this Ms. and in ferus. Taphou 57 mentioned below were made by the first scribes. !8 See Papadopoulos-Kerameus (supra n.17) 138-42; on the scribe, RGK I.A 65, II.A 50. The date is deduced from internal evidence, but it cannot be off by much; cf. P. Canart and L. Perria, • Les ecritures livresques des xi" et xii e siecles," in D. Harlfinger and G. Prato, Paleografia e codicologia greca (=Atti del II Colloquio internazionale, Berlin-Wolfenbuttel 17-21 ottobre 1983 [Alessandria 1991]) 72 n.16. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. DAVID J. MURPHY 303 Morgan 340 (s. XII-XIII), a Gospels, four hyphens in 304 folia (plus one hyphen in the right margin); Scribe e of Princeton Garrett 14 (s. XII-XIII, ff. 276-311), Homilies of John ehrysostom, four hyphens; and Lond. Add. 37486 (s. XII-XIII), a New Testament copied by Xenophon, hyphens on proper names (RGK I.e 334). In addition to Taphou 46, hyphens appear in the right margin in the two Gospel folia that constitute New York Columbia Plimpton 12, s. XII, all written by the nrst scribe. In another group of twelfth-century Mss., hyphens· appear in either margin and at other times as 'paired' hyphens in both margins simultaneously on the same word: Ambr. E 97 sup., Praxapostolos; ferus. Taphou 43, a Praxapostolos; Vat. Ottob. gr. 381, Acts-Epistles-Gospels, dated 1200, with numerous sublinear hyphens that extend into the margin; ferus. Saba 166, a Gospel Ms.; Saba 244, likewise a Gospels, which offers in addition sublinear, boat-shaped hyphens that intrude into the right margin to span page breaks, as though sub Ii near hyphens in the margin perform the same function for compound words as hyphens at the height of the letters rerform for other words (e.g. f. 100r 1CaO£u~a'te 'tQ.)1 A.outOV and . 227r J,1£OQ...JIV'U1C'tlo'U ).19 The same theory underlies the habit of Theodore tou Kazan­ opoulou, who copied an Aristides, Ambr. G 66 sup., in 1294. Alongside coupling strokes in the left margin, we nnd that on divided, compound words, a sublinear stroke at line end is added to the coupling stroke in the left margin. That the line­ end position occasions the creation of these paired hyphens is revealed by contrast with the many words formed from two roots that receive no hyphen in mid-line: e.g. f. 117r eqllAavOpro1tEUoav'to and f. 134r (cf PLATE 4), where ~llcato­ auvTI in mid-line lacks sublinear hyphens but receives them when divided at line end. An indication that the marginal coupling stroke is a logical outgrowth of the sublinear hyphen is furnished on f. 15r, where the sublinear hyphen, correctly placed on J,1£yaAo.-I-'!1'Uxla, not only intrudes into the right margin but turns upward to the height of the letters. The same combination of an upturning, sublinear hyphen on the right with a horizontal coupling stroke on the left occurs in Ambr. e 100 sup., a palimpsest of s. XIII. 19 'to Aoutov not uncommonly receives a sub linear hyphen in Ms. of the eleventh century and later, as do expressions like /)ijAov on. 'to au'to. etc. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 304 HYPHENS IN GREEK MANUSCRIPTS The late date of the two latter Mss. means that we cannot use them as a proof that marginal hyphens developed from their sublinear predecessors. What we can say is that the practice seen in these two Ambrosiani illustrates the way in which the marginal coupling stroke and the sublinear hyphen perform a kindred function: to join word elements. This intuition pro­ vides the theoretical basis, so to speak, for supposing that the former could grow out of the latter. Many scribes, of course, used both marks, as in Vat. Ottob. 381 mentioned above. From this beginning in s. XII, hyphens in the margin become more and more common. Of some 183 whole or partial Mss. of the twelfth century that I have examined (including eighteen that could be s. XI or XII), I have found twelve to contain some hyphens, or 7%.20 Of 214 thirteenth-century scribal productions (including twenty-three that could be s. XII or XIII), fifty-nine, or 27%, contain hyphens. For the next three centuries, data on texts with hyphens total as follows: s. XIV, 52 of 114 (46%); s. XV, 123 of 176 (70%); s. XVI, 89 of 97 (92%). On the other hand, of 120 items from the eleventh century (including twenty-three that could be s. X or XI), only Ambr. B 56 sup. can possibly be considered a 'hyphenated' Ms., and as we have seen, it contains only one hyphen. I found no exam­ ples in thirty-eight pieces from s. X (including seven that could be s. IX or X), none among seven from s. IX, and no examples among one Ms. of s. VII-VIII, one of s. V-VI, one of s. V, and one of s. IV-V. I have seen no marginal hyphens in papyri. 21 Among dated Mss. or Ms. portions, the following proportions of Mss. with hyphens to the total obtained: s. IX: 0:1; s. X: 0:3; s. XI: 1:34 (Ambr. B 56 sup.), 3%; s. XII: 2:48 (Taphou 57, 1182, and Vat. Ottob. gr. 381, 1200),4%; s. XIII 1: 0:23, 0%; s. XIIJ2: 9:22,41 %; s. XIV1: 10:26,38%; s. XIV2: 5:10, 50%; s. XV1: 37:52, 71%; s. XV2: 52:72, 72%; s. XVJ1: 13:16,81%; s. XVJ2: 17:17, 20 In compiling these statistics and those following on 'regular' usage of hy­ phens, I count scribal productions of at least eight folia rather than codices as such, and in the case of Mss. earlier than s. XII, of at least fifteen folia, for several hands often are found in one Ms. I have counted multiple hands in this way for ninety-five codices in the total sample of 795 Greek Mss. On the other hand, when volumes of a multi-volume Ms. are given separate numbers but are the work of the same scribe, as e.g. Ambr. E 49 info and E 50 info (s. IX), I count them as one item. 21 Hyphens in printed editions of papyri are modern printing conventions. Many papyri contain horizontal filler signs to justify the right margin, but these should not be confused with coupling strokes. Cf e.g. P.Oxy. LI 3614.6 a. R. Rea, ed. [1984] 36 with PI. V), or P.Oxy. L 3557 ([1983] PI. XIII). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. DAVID J. MURPHY 305 100%. Hyphens thus become more frequent in the second half of the thirteenth century. The genre of the work copied affects the likelihood that a scribe will use hyphens. Liturgical Mss., whether of prayers or music, virtually never contain them until the sixteenth century, and even late ones like New York Columbia Plimpton 5 (s. XVIII med.) or New York Public Library 126 (1731 in part) may lack hyphens. The absence of hyphens in the majority of liturgical Mss. may seem strange, for the hyphen would help the lector or cantor to phrase the words aloud, but it is explicable both as a result of the conservatism that ruled the production of this kind of codex and as a consequence of the tendency of the Byzantine chant to build itself upon syllables rather than words. In the case of lectionaries, a lector's familiarity with the tradi­ tional pericopes would tend to make hyphenation unnecessary. Lexicons most often limit entries to a single line, so that oc­ casions for hyphens are few. 22 Most types of poetry do not contain lines that end in the middle of a word. (this never happens, for example, in Homer), and even when lyric meters demand that a line end in mid-word, Mss. often reflect a misap­ prehension of strophic structure, so that lines are wrongly made too long, with words less often divided. I have found only a few, late instances of hyphens in Mss. from Mt Athos,23 a fact perhaps explained by the conservatism and isolation of that monastic milieu. On the other hand, among scribes of s. XIIP, those who worked on texts of classical prose authors seem more likely to have put frequent hyphens in the margins. Hy­ phens appear often in Mss. of authors like Aristotle, Simplicius, Aristides, and Plato, which were copied by scholarly hands. Frequency of Hyphens in the Margin As we have seen, in the earliest Mss. that display a pattern of hyphens, one hyphen appears about every five to ten folia. Their use may be motivated by the appearance of potentially 22 One lexicon that does have hyphens is New Haven Beineeke Marston 94, copied ea 1460 in Verona. 23 Athos Esphigm. 78, copied in 1417 (ef E. Lamberz, -Die Handschrif­ tenproduktion in den Athosklostern bis 1453," in Seritture I 65, Taf. XIX); New Haven Beineeke 251, copied by Cyrillus in the Monastery of Simopetra in 1585 (with one two-stroke hyphen on p.1330 in a portion copied in 1651 by Joasaph of the same monastery). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. HYPHENS IN GREEK MANUSCRIPTS puzzling words like proper names (as with Taphou 57, Staurou 47 [so XII] and Ambr. C 124 sup. [so XIII]) or words in which the syllables might be misread as words themselves (as with Ambr. D 57 sup. [so XIII], fol. 39r EV Kat aY....LJt 1tOAA&V). Often, however, no clear pattern emerges that will explain why the copyist chose to hyphenate some few words aad not the others, and this is not surprising, given the centuries-long vagaries of scribal habits of punctuation (cf Turner [supra n.7] 11). It is not useful to attempt to create overly fine distinctions between stages of the relative frequency with which hyphens occur, for the increase in their number is steady but gradual. After 1250 one begins to find several hyphens on a page in certain codices, e.g. Vat. Reg. gr. 63, 1259-60; Par. gr. 857, 1261 (Omont [1891] PI. LVI); Vat. gr. 64, 1269-70; Par. gr. 1115, 1276 U. Munitiz, Scriptorium 36 [1982] PI. 4); Patmos 14, 1280, frequent but not regular hyphens (Komines PI. 25); Neap. III.E.15 (337), Scribe C, s. XIII p.m., frequent in both margins. When hyphens begin to appear commonly on all words divided by line end, their use has become rationalized. This is what I call 'regular' employment of hyphens. The earliest ex­ ample known to me occurs in the authenticated copy, made at the same time as the original, of a chrysobull of Michael VIII Palaeologus, Jan. 1259 (Archive Iviron no. 61; cf J. Lefort etal., Actes d'Iviron II [Paris 1990] no. 58, PIs. X-XI), where hy­ phens appear in the left margin on all divided words. My earliest Ms. with regular hyphenation is Guelf 35 Gud. graec., frag­ ments of Lives of Thucydides, copied in the last quarter of s. XIII, probably in Constantinople (GHA Abb. 10), with hy­ phens on the left. Other early examples of 'regular' hyphens: fol. 2r-3r of New York Columbia Smith Western Add. 10, which leaves were copied by George (later Gregory) Chioni­ ades, hyphens on the left (cf PLATE 5); 24 Par. gr. 234, section of 24 These leaves contain the end of Chioniades' discussion of division of sciences (the beginning has fallen out of the Ms.) and his introduction to St John of Damascus. Chioniades' marginal scholia, which were written at about the same time (the main scribe leaves room for him to construct astro­ nomical diagrams on fo!' 112r-115v), contain frequent but not fully regular coupling strokes on the left, as well as sub linear hyphens. The bulk of the codex contains works of Damascenus and was copied by Symeon Chartophylax, who hyphenates frequently but not consistently on the left. His name appears in a monocondylion on f. 273r. No scribe listed in VG or RGK I-II can safely be identified with him. His subscription on fo!' 271 v reads ~"vl (JE7t't{EJ.I.~picp) iv(1)\lc'tlrov)o~ \' £'to\)~ ~roE (?), i.e., 1296 (not 1297, against D. Pingree, The Astronomical Works of Gregory Chioniades, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. DAVID J. MURPHY 307 Scribe A, who collaborated with Romanos Anagnostes in 1318 on Cyprus, hyphens on the right (Constantinides PI. 40); the second part25 of Vat. gr. 256, ca 1321, almost regular paired hyphens with a few singles on the right. During the fourteenth century, many Mss. display frequent but not regular hyphens, i.e., three to six per page. 'Regular' hyphens do not begin to become common until the Italian Renaissance, and then they rest in the right margin. Single strokes are l'laced consistently on divided words in Flor. Rice. 54, a copy of Plato's Protagoras and Golden Sayings of Pythag­ oras copied in 1405. In the roundness and relative size of its letter forms, the scribe's hand looks like that of a poorly edu­ cated Italian rather than a native Greek copyist. Indeed, on f. 58v, where the scribe writes" dolor .... in the margin, one can see him forming his Greek and Latin letters with the same movements, so that he appears to be a Westerner. The colo­ phon on fo1. 70r seems to indicate that the Ms. was copied in Constantinople,26 perhaps by an amanuensis who had gone there in the service of an Italian collector. Among 176 scribal productions of s. XV that I have examined, twenty-seven display a pattern of regular hyphens (15%),27 and fifty of 97 from s. XVI have-regular hyphens (52%). Of 332 scribal produc­ tions represented in plates from s. XV (cf. Omont [1887], RGK I-II, Bernardinello, Harlfinger, and GHA), forty-three show a pattern of regular hyphens (13%), and of 283 plates from s. XVI, 117 show regular hyphens (41 %). As the fifteenth century advances, marginal hyphens or coupling strokes are employed 'regularly> by more and more scribes. Indeed, discounting eight plates that are not dated more precisely than to the second half of the s. XV, twenty-two of the thirty-nine remaining fifteenth-century plates that show regular hyphens I.i: Corpus des Astronomes Byzantines II [Amsterdam 1985] 18). The last numeral is apparently the lower half of a closed, minuscule epsilon, the upper half of which has faded in a blot on the page. On the Ms., cf. also Pingree's findings as reported by L. G. Westerink, "La profession de foi de Gregoire Chioniades," REByz 38 (1980) 237f. 2S Cf. G. Mercati and P. Franchi de' Cavalieri, Codices Vaticani graeci I: Codices 1-329 (The Vatican 1923) 337; CL Pl. 40. 2£ As transcribed by G. Vitelli, "Indice de' codici greci Riccardiani, Maglia­ bechiani e Marucelliani," StIt 2 (1894) 506: "die vi Oct. finivi. 1405. Constan." 27 Some earlier examples are: Lond. Hart. 5587, s. XV init. (RGK 1371); Vat. gr. 1334, ca 1420-27 (Harlfinger spec. 10); Vind. phil. gr. 151, 1427 (Bick PI. XXXII); Par. suppl. gr. 192,1439 (RGK II 246). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 308 HYPHENS IN GREEK MANUSCRIPTS date after 1475. Of twenty dated Mss. of s. XV in whose margins I found regular hyphens, fifteen date from 1465 or later. Although hyphens become increasingly commonplace, some Mss. even of the sixteenth century and later have no marginal hyphens. Even some Italian scribes totally eschew their use. 28 Therefore, the absence of hyphens cannot be used to date a codex. This caution is especially important for anyone inspect­ ing photographs of isolated folia of particular Mss., because hyphens may appear on some leaves but not others. Shape and Position of the Marginal Hyphen The marginal hyphen was first formed as a single, horizontal stroke in either margin. This was drawn either as a straight line, as a wavy line resembling a tilde (-), or as a straight or wavy stroke that bends up or down on one end. I have not sought to distinguish these shapes as definite types, for scribes often indulge a desire for variation even on the same page. More useful distinctions are those based on number, position, or angle of strokes, for decorative styles flourish in the fifteenth century and after, many of them importations from Latin Mss. Early and late examples are offered below. Stage .5A, single stroke in left margin: Begins s. XII, first regular usage s. XIII ex. (cf above). Latest examples: Vat. gr. 2139, 1498, by Demetrius Moschus Lacon, f. 25r; Vind. phil. gr. 259, 1500, fo!' 176v: no).:\) I - av9pcon(a and 176r £11 -1'\ by third scribe; Scor. gr. n.l.1, 1523, regular (Bernardinello tav. 76) . .5B, single stroke in right margin: Ambr. 56 sup., 1022, then s. XII to present, Par. gr. 234, 1318, first regular usage (cf above), by far the most common type by s. XIV . .5 C, paired hyphen, single stroke in each margin: Begins s. XII (cf above); almost regular usage in Vat. gr. 256, 1321, Scribe B Latest examples: Mut. Esten. gr. 19 (III A 5), 1519, (Constan­ tinides PI. 156); Marc. gr. 554, 1538 (RGK I.C 173); New Haven 28 E.g. Damiano Guidotto, who copied Columbia Plimpton 3 ca 1500, as well as other Mss.: cr. P. Eleuteri and P. Canart, Scrittura greca nell' Umanesimo Italiano (Milan 1991) 89 and Tav. XXX, with bibliography. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. MURPHY PLATE 1 Ms. Jerus. Staurou 43, A.D. 1122, f.3v (Courtesy Library of the Greek Orthodox Patriarchate, Jerusalem) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. PLATE 2 MURPHY Ms. Jerus. Taphou 46, s. xii, f. 5v (Courtesy Library of the Greek Orthodox Patriarchate, Jerusalem) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. MURPHY PLATE 3 Ms. Jerus. Taphou 57, ca A.D. 1182, f.42r (Courtesy Library of the Greek Orthodox Patriarchate, Jerusalem) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. PLATE 4 MURPHY .. ..,jl • '. :!I ~,' ..:""""""", ",#.l> "'V"':tt~ Ms. Ambrosianus G 66 sup., A.D. 1294, f. 134r (Courtesy Ambrosiana Collection, University of Notre Dame) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. PLATE 5 MURPHY Ms. Columbia Smith Western Add. 10, A.D. 1296, ff. 2v-3r (Courtesy Rare Book and Manuscript Library, Columbia University) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. DAVID 1. MURPHY 309 Yale Medical Library 35, 1585, copied by Andrea Darmario, one instance. 5D, horizontal two-stroke on right, like an 'equals' sign (=), in certain documents long before appearing in Mss.: Archive Xeropotamou 35, Prostagma of Andronicus II, October 1295, lCa't£= IXeaeal (Xeropotamou PI. XXXIII). Also Xeropo­ tamou PIs. XLIV-XLV, 1346, and PI. XLVIII, 1351 (=Dolger 19); New Haven Beinecke 424, 1541, by Valeriano Albini da Forli in Venice, two-stroke and single strokes; Vat. Reg. gr. Pii II 53, s. XVI2, regular usage (RGK I.e 144); Marc. gr. 11.62, 1600, hori­ zontal and slanted two-stroke hyphens, regular (RGK I.e 11); Patmos Mss. of s. XVII-XVIII (Ms. 295, 354, 503, 596, 623, 684, 754, Mae. na'tll. 143; cf. plates in Komines); Prostagma of John VII Palaeologus, copy made in s. XVIII-XIX (Dolger 45' II [PI. 45a]). 5E, paired horizontal two-stroke, like paired 'equals' sign: Vat. gr. 1892, fol. 181vim, s. XVI?; Vat. Regin. gr. 121, s. XVI, with other types of hyphens; Letter of Mathaios Barsarampas, 21 March 1642 (Thesaurismata 15 [1975] PI. A, letter 3), paired and single hyphens. 5E, slanted single stroke on right (/): Scor. Y.I.13, Plato, s. XIII", ff. 6r, 8r; Ambr. R 50 sup, 140S?: Latin text of Aristotle with Greek around it by same scribe; Latin hyphens complete, Greek sporadic; appears to be slanted on right on 3r, 7v; Vat. Urb. gr. 32 and 33, 1444, by Leonardo Aretino; Par. /at. 13038, ante 1478, Aulus Gellius of s. XIJ2 with Greek quotations added by Pietro da Montagnana: regular on right in both languages (Scriptorium 30 [1976] PI. 16); Vat. gr. 693, 1566 (R. Riedinger, ByzArch 12 [1969] PI. 11); Oxon. Can. 52, 1600, r~gular (RGK I.e 262). 5G, slanted single stroke on left: Scor. X.I.13, s. XIII-XIV, f. 260r, 264v, plus other styles; Vat. gr. 933, s. XIII, f. 313v, 333v (where hyphen runs into first letter). 5H, slanted two-stroke on right, like slanted equals sign, adopted from Latin Mss. (cf. Vat. Pal lat. 330, 1400: many slanted two­ strokes on right): Oxon. Seld. B 45 (40), 1482 (R G K I.C 157); Oxon. Corpus Christi 106, 3 May 1495 (R. Barbour, Greek Literary Hands [Oxford 1981] PI. 92); Cant. U. L. Add. ,4173, s. XVIl, regular usage (RGK I.e 340); Lond. Add. 40755, 1600, faint two-stroke, almost two-dot, regular (RGK I.C 245); Athen. Meteor. 217, 1634-35 (Bees, PI. XXXI); New York Pierpont Morgan 656, s. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 310 HYPHENS IN GREEK MANUSCRIPTS XVII, Psalter by Matthew of Buzau in Romania: scattered usage, small size, along with two-dot hyphens (cf 5K below). 51, paired slanted two-stroke: Marc. gr. II.93 (colI. 562), s. XVI, autograph of Meletios Pigas; ovoll6.1/ I I/'E'tat oOPKI/ I 1/6.0t etc. (Thesaurismata 16 [1979] PI. 18'); A rchiv. Stat. Yen. Collegio III, Secreta Lettere Cardinali, Patriarchi, Vescovi b. 2, 15d, letter of Meletios Pigas, 26 Septem­ ber 1590 (Thesaurismata 11 [1974] PI. B). 5J, slanted two-stroke on the right below the line, adopted from La­ tin humanist hands (cf New York PubL Lib. 59, s. XV ex., Latin translation of Aristotle's De virtutibus by Georgius Her­ monymus: slanted one-stroke hyphens below the line): Vat. Pal. gr. 392, 1564 by Andrea Darmario, ff. 82-87, with earlier parts done in s. XVI; both first hand and Darmario use this type. 5K, two dots on the right, like a colon: Par. gr. 3033, s. XVI med.: two-stroke, occasional two-dot, regular usage (RGK II.C 227); Lond. Add. 37010, s. XVI ex., up: IXE1tI.OK01tOU, etc., regular usage (RGK I.C 207); Archiv. Stat. Yen. Collegio III, Secreta Lettere Cardinali, Patriarchi, Vescovi b. 2, 15d, letter of Meletios Pigas, 26 September 1590 (cf, type 51 above); Princeton Greek 3, ca 1780, musical Ms. copied by Petros Byzantios Protopsaltes, two-dot, regular usage in titles and rubrics; Siegelbrief of February 1791, regular (Dolger 92); Princeton Garrett 12, s. XIX init., musical Ms. with two-dot hyphens used regularly in titles and rubrics. The Genesis of Coupling Strokes Hyphens in the margin were common already in Latin and Anglo-Saxon Mss. at the time when they begin to be seen in Greek codices. A few Latin scribes prior to s. X 2 try to show word division by using an occasional virgula 29 or a high point. 30 29 Vind. lat. 418, written in Salzburg 785-821, uses horizontal virgulae; cf. E. A. Lowe, Codices Latini Antiquiores: A Palaeographical Guide to Latin Manuscripts Prior to the Ninth Century X (Oxford 1963) 1478 (no examples on the plate, however). Vezin (supra n.2: 590) refers to the discussion in Nouveau traite de diplomatique of smal~ upright virgulae in Par. lat. 1732, s. VIllI, but the authors of the Nouveau traite report (III 478) that the same mark is used in cases when words are not divided, so that it is not clear whether it is a real coupling stroke. This virgula came to be called the virgula iacens, i.e., horizontal, in discussions of punctuation; cf. Vezin 591. 30 Monac. CLM 14513, s. VIII-IX (CLA IX 1302); Thompson (supra n.3) 58. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. DAVID J. MURPHY 311 Straight strokes in the right margin used to divide words are attested by s. X2 in the British Isles,31 and they become general on the Continent by the twelfth century. Horizontal strokes dominate at first,32 but a consistent slant prevails by s. XII. 33 Two explanations of the appearance of this mark in Greek Mss. are the most likely: either it developed according to the 'internal logic' of the sublinear hyphen, or it was borrowed from Latin codices in some place where copyists of Greek had access to them. Follieri has posited a link between punctuation marks shaped like a nail, which appear in Greek codices of southern Italy in the tenth to twelfth centuries, and similarly shaped marks in Beneventan MSS.,34 and Ambr. B 56 sup., our earliest Ms. to show a coupling stroke, was copied in this same region, viz. Campania. Hyphens in twelfth-century Greek Pales­ tinian Mss. might be taken as witnesses to the influence in that region of Latin scribal habits such as those visible in Queen Melisende's Psalter, which was copied in Jerusalem ca 1149 and 31 N. R. Ker, Catalogue of Manuscripts Containing Anglo-Saxon (Oxford 1957) xxxvf; J. Vezin, «Les manuscrits dates de l'ancien fonds latin de la Biblio­ theque Nationale de Paris," Scriptorium 19 (1965) 86. Saenger (supra n.2) finds the coupling stroke even in Mss. of the eighth century. 32 Cf e.g. Oxon. Keble Coll. Ms. 22, copied at Salisbury in s. XF (M. B. Parkes, «Punctuation, or Pause and Effect," in J. J. Murphy, ed., Medieval Eloquence. Studies in the Theory and Practice of Medie'/lal Rhetoric [Berkeley 1978J 141); Neap. VI.E.43, 1097-99 in Beneventan script, with horizontal or slightly slanted hyphens (inspected by autopsy; cf E. A. Lowe, Scriptura Bene­ ventana. Facsimiles of South Italian and Dalmatian Manuscripts from the Sixth to the Fourteenth Century II [Oxford 1929J PI. LXXXI); Vat. Lat. 3833, 1099-1118 (inspected on film; cf also F. Erhle and P. Liebaert, Specimina codicum latinorum Vaticanorum2 [Berlin 1932J PI. 35). 33 Vezin (supra n.31) 86. Early examples of oblique hyphens: Lond. Cotton Nero A.VII, s. XI ex. (M. P. Brown, A Guide to Western Historical Scripts from Antiquity to 1600 [Toronto 1990J PI. 24); Bern. Burgerbibl. 386, s. XI­ XII, with slanted hyphens used regularly in text and scholia (L. Holtz, «La typologie des manuscrits grammaticaux latins," Re'/IHistText 7 [1977J PI. III); Tal. Catedral Ms. 14-3, 1105 (F. M. Gimeno Blay, -De Scriptura Gothica. Algunos ejemplos a rrop6sito de sus inicios en la peninsula iberica," Scrip­ torium 47 [1993J P. 20). Oblique hyphens rest below the writing line in Durham, Dean and Chapter Library B.11.22, s. XI ex. (M. B. Parkes, Pause and Effect. An Introduction to the History of Punctuation in the West [Hants 1992J PI. 53). 34 Cf E. Follieri, «Attivita scrittoria calabrese nei secoli X-XI," in Calabria bizantina. Tradizione di pieta e tradizione scrittoria nella Calabria greca medie'/lale (Reggio Calabria 1983) 119-23. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 312 HYPHENS IN GREEK MANUSCRIPTS carries such hyphenated words as femi-I neo sexu. 35 But even in Constantinople, large numbers of Westerners began to make their presence felt as trade and cultural contacts increased in the eleventh and twelfth centuries.36 On the other hand, an occasional Western scribe would write Greek, most often in a Latin Ms., and it is conceivable that Greek-speaking scribes may have come across some of their productions and have gotten the idea of marginal hyphens from themY Some Greek Mss. make use of parchment taken from Latin codices, either indi­ vidual pages that serve as flyleaves 38 or even entire Mss. reused as palimpsests,39 and marginal hyphens could have been imitated from examples in these. The same speculation may arise from 35 Lond. Egerton 1139, fol. 202v. Cf H. Buchthal, Miniature Painting in the Latin Kingdom of Jerusalem (Oxford 1957) 1-9, Pis. 17b, 18-19; L. A. Hunt, "A Woman's Prayer to St. Sergius in Latin Syria: Interpreting a Thirteenth­ Century Icon at Mount Sinai," ByzModGrSt 15 (1991) 114f and fig. 18. On the contact between Latins and Greeks in Palestine, cf R.-J. Lilie, Byzantium and the Crusader States 1096-1204, tr. J. C. Morris and J. E. Ridings (Oxford 1993) passim, esp. 195-252; A. P. Kazhdan and A. W. Epstein, Change in Byzantine Culture in the Eleventh and Twelfth Centuries (Berkeley 1985) 174f. 36 Cf W. Berschin, Greek Letters and the Latin Middle Ages. From Jerome to Nicholas of Cusa, tr. J. C. Frakes (Washington 1988) 201-42; Kazhdan and Epstein (supra n.35) passim, esp. 181-201. For a case of Greek-Latin collaboration and mutual influence in the late eleventh or early twelfth century, cf works by N. G. Wilson: "A Mysterious Byzantine Scriptorium: loannikios and his Colleagues," Scrittura e Civilta 7 (1983) 162-76; "New Light on Burgundio of Pisa," Stlt 79 (1986) 113-18; "Ioannikios and Burgundio: A Survey of the Problem," in Scritture II 447-55. Unfortunately, no clear examples of hyphens appear in the plates of the 1983 article. 37 Cf w. Berschin, "Griechisches in der Domschule von Verona," in Scrit­ ture 1221-34, and (supra n.36) 37ff. Oxon. Canon. gr. 97 was written 1239-43 in Paris, and in Corpus Christi Coli. 108 a Western scribe of similar hand copied a medical note using a hyphen: crcpu1C£A.i~ov,tE<;· cr7tucr- I,.HOOT] 7tT\oiJ­ I1U'ta; cf Greek Manuscripts in the Bodleian Library: An Exhibition held in connection with the XlIlth International congress of Byzantine Studies (Oxford 1966) no. 37, Pl. XV. 38 In Ambr. C 60 sup. from the early s. XIII, ff. II-III have been taken from a Latin Ms. of about the same age. Cf also e.g. Ambr. C 135 inf., ff. 387f from a Roman Missal of s. XIII, used as backing sheets; Ambr. Q 74 sup., ff. 1£, 26M from a Latin breviary of s. XIV-XV 39 E.g. Ambr. C 100 sup., a thirteenth-century palimpsest from Calabria written over a Gregorian missal of the previous century. Hyphens appear in the Latin but are lacking in the Greek rescript text. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. DAVID J. MURPHY 313 the production of bilingual codices in southern Italy or even the Greek East. 40 All the above only argues for the plausibility of borrowing. The only datum that can be taken to support the 'borrowing hypothesis' is fo1. 3v of Staurou 43 (ef. above on Stage 4B and PLATE 1), for the virgula served as a coupling stroke in some Latin Mss. (el supra n.29). Working as a scribe and lector of the Shrine of the Anastasis, which was then under the control of the Latin Patriarchate, Basileios may have seen Latin Mss. with such marks. 41 The mark on 3v, however, cannot prove that Greek conventions of marginal hyphens in general came from Western models, for Basileios is only one copyist, this mark is the only such in the codex, and it is not formed like the virgulae that divide words in Latin Mss. Against the 'borrowing hypothesis', the location of early hy­ phens in either or both margins, with a preference for the left, reflects the practices observable in the placement of sublinear hyphens on divided words in Greek codices. It clashes with the Latin habit of placing hyphens on the right. 42 Second, Greek hy­ phens before the Renaissance lie parallel to the wri ting line, be 40 Unfortunately for this theory, no example I have seen is much earlier than s. XIII, and my earliest bilingual codex with hyphens in both languages is a humanist Ms., Ambr. R 50 sup. (1408?), a Latin translation of Aristotle with the Greek written around it by the same scribe as far as f. 25r. Some bi- lingual Mss. with hyphens only in the Latin are: Lond. B. L. HarL 5786, 1153 (Lake II, PI. 141, trilingual); Oxon. Canon. gr. 63, a Latin-Greek psalter on Western parchment, s. XIII init. (Gk Mss . ... Bodleian [supra n.37] item 42, PI. XVIII); Vat. Pal. gr. 232, Ad7.lersus Latinos of Nicholas of Otranto, copied in a crude, Southern Italian hand of s. XIII-XIV, with an adjoining Latin translation composed by the same copyist; Vat. gr. 329, copied s. XIII-XIV, probably in Ephesus (G. De Gregorio, "Osservazioni ed ipotesi sulla circo­ lazione del testo di Aristotele tra Occidente e Oriente," in Scritture II,491), with an occasional hyphen in the Latin of Cato's Distichs and Boethius (e.g. f. 64v pe- Icunie); Vat. gr. 760, copied in the Greek East ca 1330-40 (De Gre­ gorio, tav. XVI). Romanus, Abbot of St Benedict of Uliano near Cosenza, used no hyphens in either language in Vat. gr. 1070 (ef CGV 80 and tab. 57) and Vat. Barb. gr. 541, both of 1291-92. 41 Although clergy living in the monastery attached to the shrine were West­ erners organized under the rule of St Augustine (ef Buchthal [supra n.35] xxx), some lower Greek clergy continued to have ties to the Anastasis and celebrated the liturgy there. Cf J. Prawer, ·Social Classes in the Crusader , States: The 'Minorities'," in N. P. Zacour and H. W. Hazard, edd., A History of the Crusades V: The Impact of the Crusades on the Near East (Madison 1985) 73-76. 42 A small number of Latin and Anglo-Saxon Mss. prior to s. XII have 'paired' hyphens. Cf Vezin (supra n.31) 86; Ker (supra n.31) xxxvi. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 314 HYPHENS IN GREEK MANUSCRIPTS they straight, wavy, or upturned on one end, although Latin hy­ phens in s. XII and later tend to be written as oblique strokes. Of course, different conventional practices do not of themsel­ ves prove different sources. 43 But the weight of convention is considerable. An example of how a copyist could keep Greek and Latin practices distinct is provided by Vat. Pal. gr. 232 (cf supra nAO), whose scribe uses sublinear hyphens only in the Greek, and marginal coupling strokes only in his Latin trans­ lation. Unless more evidence emerges to show that Greek scribes did derive the marginal hyphen from their Latin cousins, it seems prudent not to press this admittedly tantalizing hypothesis. And it is not yet possible to establish where Greek scribes first began to use coupling strokes, although a Pales­ tinian connection is worth tracing. Only with the appearance in the Renaissance of Italians who wrote Greek, or of Greeks who copied for Italian patrons, do we find distinctive Latin hyphen types beginning to surface in Greek Mss. (c{. supra Stages 5E-J). These are cases of direct influence. And however much it looks as though the marginal hyphen was created when certain Greek scribes moved the sublinear hyphen into one of the margins (cf supra Stages 3-4), the increased hyphenation after 1250 may have been abetted by Latin influences that had spread after the capture of Constan­ tinople. 44 THE NIGHTINGALE-BAMFORD SCHOOL, NEW YORK December, 1995 43 R. Marichal has shown that the diple was placed in the left margin in certain Greek papyri from Herculaneum but within or at the end of the line in the Latin P. HercuL 817; cf "De I'usage de la 'diple' dans les inscriptions et les manuscrits latins," in Palaeographica Diplomatica et Archivistica. Studi in onore di Giulio Batte/Ii (Rome 1979) I 63-69, esp. 67. 44 I wish to thank the staff of the many libraries where I worked for their forbearance and ever patient assistance. I would like to thank also the Nightingale-Bamford School for its generous support of my research through a summer stipend. Finally, I am grateful to Nigel G. Wilson and Francis Newton for their suggestions and criticism. Naturally, any faults in this paper are my sole responsibility. Earlier versions of this paper were presented at meetings of the Classical Association of the Atlantic States in Princeton, April 1993, the Saint Louis Conference on Manuscript Studies, October 1995, and the American Philological Association in San Diego, December 1995.