IV. — The Origin of Sigma Lunatum.

BY JOHN HENRY WRIGHT,

HARVARD UNIVERSITY.

THE earliest and indeed the only form of sigma found upon the most ancient Greek papyri1 is the sigma lunatum, or crescent-shaped sigma, a form also referred to in literature as far back as the age of Alexander. In a well-known line of the iambographer Aeschrio, the contemporary and friend of Aristotle,2 the (new) moon is described as the beautiful sigma of the heavens. On the Artemisia papyrus 4 in Vienna, which may still be regarded as the oldest Greek inscribed papyrus (ca. 300 B.C.), the letter assumes an angular form <. This crescent-shaped sigma persisted without essential modification for more than a thousand years. Inscriptions, however, contemporary with the oldest papyri, as well as those of earlier and later date, regularly give us the four-bar Ionic form (≤).

6

Sigma lunatum is usually explained as derived from the Ionic sigma by successive simplifications, due, first, to a rounding of the exterior angles, and, secondly, either to a merging of the outer curves thus obtained into one long

1 Cf. Blass, Griechische Paläographie (I. Müller, Handbuch 12), pp. 304 ff. 2 So at least Suidas (Ptolemy Chennus ?).

3 Mývn тò kaλòv ovpavoû véov oîyμa, ap. Walz, Rhet. Graeci, III. 651 (Anon. Epit. Rhet.). Bergk suggests that vén should be read, 'verbis audacius traiectis' (P. L. G. II1. p. 516). Of course véov does not mean 'a new form of sigma.'

4 Facsimile in Palaeographical Society, II. pl. 141, and in part in Thompson, Handbook of Greek and Latin Palaeography, p. 118. Cf. Blass, Philologus, XLI. P. 746 ff.

5 Thompson, l.c., p. 118. Mahaffy places slightly earlier than the Artemisia papyrus some scraps from Gurob, in which the Labors of Heracles seem to be described (Petrie Papyri, pp. 52 ff., and table of alphabets opposite p. 64). Wattenbach speaks of this alphabet as about the same age as the Artemisia papyrus (Anleitung zur Griechischen Paläographie, 3d ed., p. 9). Cf. Mahaffy, Empire of the Ptolemies, p. 30, note.

Sporadically sigma lunatum appears on the stones from the fourth century B.C. onward: cf. Larfeld, Griech. Epigraphik (I. Müller, Handbuch 12) p. 535.

sweeping crescent (with or without an attempt to preserve a trace of the entering angle), or to a discarding of one of the curves.1

This derivation may be illustrated as follows:

It is the purpose of this paper to show that the more probable derivation of sigma lunatum is from the three-bar Attic letter, of which it was an earlier cursive form. Or:

The evidence upon which this theory is based is drawn in large part from the Athenian vase inscriptions of the fifth and fourth centuries B.C. In palaeographical questions of this sort the evidence of the vases has been too little employed. The vase inscriptions stand midway between epigraphical and palaeographical monuments, sharing the qualities of each between these two stools they fall to the ground. The neglect of them is due in part also to the inadequate character of much of our information concerning the true forms of letters on the vases. Editors of catalogues of vase collections, and other writers on the subject, are frequently satisfied with giving only a few representative forms, from fonts of printed type, and not exact facsimiles from which alone safe inferences can be drawn. The tables of facsimiles, however, in Jahn's Munich catalogue and in Heydemann's Naples catalogue, and the carefully executed facsimiles in Furtwängler's model Berlin catalogue, are happy exceptions. In these facsimiles, in a few others in recent periodical publications, as well as on the vases themselves (in the Boston collection), I have found what appear to me sufficient data for determining the question.

The inquiry is interesting as throwing some new light on the probable forms of the letters used in the autograph manuscripts of the great writers of the fifth century B.C.

My contention, then, is in brief that sigma lunatum is

1 Blass, Griech. Paläographie, p. 304.

a cursive form of the three-bar Attic . It is important, before proceeding further, to define what we here mean by a 'cursive' letter, since there are two classes of forms to which the term might be applied.

There are two classes of cursive forms, which we may designate respectively as normal cursives, and as casual cursives. The former are the shapes that letters regularly assume, when one writes currente calamo. As contrasted with the engraved 'print' form, they are commonly uniform and rounded, instead of angular. The latter, the 'casual' cursives, are the irregular forms into which, in hasty writing, the set, engraved, or print letters are thrown. When in an epigraphic monument - an Athenian decree, for example

we find a rounded form of a letter, instead of the more usual angular form, in the midst of carefully cut letters. retained in the angular form, we may safely assume that this rounded letter is a 'normal' cursive imported into the epigraphical alphabet from the script. Such an occurrence may be used as a test of the normal cursive letters. Inscriptions of all sorts furnish excellent examples of both kinds of cursives. The same letter, at a given date, may have then three distinct kinds of form: (1) the regular epigraphic, monumental form, which may or may not differ from (2) its normal cursive form; (3) one or more casual cursive forms. Thus, for the Ionic sigma we have: (1) ≤; (2) {, as I shall later seek to show; and (3) Ɛ, €, etc. For Attic sigma we

have: (1) ≤; (2) < or C; and (3) S, ≤, ↳, etc.

Of course we must assume that the normal cursive originated in a casual cursive which convenience in writing had made a typical form. For convenience, in the remaining part of this discussion, the word cursive, when used alone, will be employed in the sense of the 'normal' form. And in the application of this criterion to vase inscriptions the utmost caution must be exercised.

That there were true cursive letters in the Old Attic alphabet, or in fact much before the beginning of the fourth century B.C., in short, that there was a script hand differing from the monumental hand in the fifth century, has been either

tacitly ignored or openly denied by scholars.1 It has been assumed that in these times there was no essential difference between the forms of letters as engraved on the monuments and as written on soft wax tablets or on sheets of papyrus. In the earliest period, when there was little writing, this may well have been the case; and the prevalence for the most part of epigraphic forms on the vases of the earlier period proves it. But with the abundant preparation of books and with the teaching of writing at school, it was inevitable that a cursive style should set in, the evidence for which is furnished not only by the vases, but also by the inscriptions on stones.

Now, as I shall try to show, not only were there cursive forms of letters in the earlier period (or before 403 B.C.), but there were cursive forms both for letters of the Old Attic alphabet and for those of the newly adopted Ionic alphabet, alphabets which at least in private inscriptions were concurrently used throughout the larger part of the fifth century. To the much that has been written proving the use of the Ionic alphabet in Athens for literary purposes in the fifth century I will add nothing:2 I wish, however, to emphasize the fact that the Attic was also used for literary purposes. This clearly appears from the language of Theopompus,3 from which we learn that the decree of Archinus promulgating the Ionic as the official alphabet from 403-2 B.C. onward had reference not only to public documents, but also to writing in the schools (τοὺς γραμματιστὰς παιδεύειν τὴν Ἰωνικὴν γραμματικήν). In the schools manuscript copies of literary works, especially of epic and lyric poets, were used; these, then, at least to some extent, must have been previously written in the Attic alphabet. A pretty proof, not only of the use of manuscripts in schools, but also of the

1 Wilamowitz, Philol. Untersuchungen, VII. p. 307. His statement, however, that the vases give a ‘lediglich monumentales alphabet' needs qualification. 2 Wilamowitz, ibid., pp. 286 ff.; Kretschmer, Griechische Vaseninschriften, p. 106. But cf. Blass, l.c., pp. 301 ff.

3 Schol. Dion. Thrac., ap. Bekker, Anecd. 783. 20; Phot. Biblioth. cod. 176. Cf. Usener, Rhein. Mus. XXV. pp. 591 ff. Kretschmer, Griech. Vaseninschriften, pp. 103, 106, note.

4 Plato, Prot. 325 E.

presence in the manuscripts of Attic (as well as Ionic) forms, is accessible in the charming scene on the familiar Duris vase in the Berlin Museum, where a teacher holds in his hand a papyrus roll on which are written phrases from an epic passage in which we find in close juxtaposition Attic and Ionic forms 2 (5, 2, etc.).

Side by side, then, if we may infer from the inscriptions of the fifth century, the Attic and Ionic forms of the same letter were not infrequently used—much as long and short s's occur simultaneously in the same page in an English manuscript. This concurrent use would lead the average writer and reader to forget their difference of origin, and, in due time, of the two forms the one best adapted for writing would supersede the more difficult one. Thus, on my theory, Attic cursive sigma C (from >) superseded, for general literary use, the Ionic cursive sigma { (from ≤). But I anticipate. It remains to be proved that there were in use in the fifth century B.C. distinctly cursive forms as contrasted with the more familiar epigraphic forms.

It is impossible here to treat exhaustively the subject of cursive letters in fifth century Greek writing. But for a period ending not much later than 350 B.C. and running well back into the fifth century, we have, at least, the following forms, which I believe may be safely assumed to have been normal cursive letters:

Є for E.3 3

1 Berlin Vasensammlung, No. 2285. Figured in Mon. dell' Inst. IX. 54; Arch. Zeit. 1873, pl. 1 (Michaelis); Gardner-Jevons, Manual of Greek Antiquities, frontispiece and pp. 309 ff.; Anderson-Schreiber, Atlas of Classical Antiquities, pl. 90. 2, etc. Cf. Kretschmer, l.c., pp. 104 ff.

2 There are many cases of mixture of the Ionic and Attic alphabets on the vases. Cf. Kretschmer, ibid., p. 105. Add Boston, Perkins coll. (Brygus? 9). See also below, pp. 86-9, for a few additional instances, especially of <, >, C.

3 The classical example is in a correction in C. I. A. II1. 17 A, 45 (B.c. 377), where we read MENOI. Cf. also C. I. A. II. 1137, 8 (B.C. 305-4). The vases furnish many examples; eg. LE A^POS, Brit. Mus. B 325 (Walters), or Munich, 48 (Jahn); IME, Naples, S. A. 172 (Heydemann). To be sure on the Artemisia papyrus and in the Heracles fragments (see above, p. 79, note 5), E has the angular form, but the retention of epigraphic forms with cursive forms is not surprising.