Conventions in Editing. A Suggested Reformulation of the Leiden System

Table of contents :
Conventions in Editing. Suggested Reformulation of the Leiden System

Contents
Foreword

I. Introduction
Apologia
General Definition of Objectives


II. Conventions

Above and Below the Preserved Text
Numberìng of Lines

The Use of Editorial Signs
Restorations : Square Brackets [ ]
Lacunae: Dashes, Dots [--], [. . .]
Spaces Left Blank by the Letterer: Small superscript ’ or or
Doubtful Readings : Subscript Dots
Attrition: Shading
Rasurae: Double Square Brackets [[ ]]
Reading Clear, Interpretation Unknown: Capital Letters
Strokes Clear, Letters Unknown: Broken Capitals
Additions by the Editor: Pointed Brackets < >
Substitutions by the Editor: Pointed Brackets < >
Letters Left Incomplete by the Letterer: Pointed Brackets < >
Suppressions by the Editor: Braces { }
Resolutions of Abbreviations and Ligatures: Parentheses ( )
Note on Pointed Brackets

Editing of Earlier Modern Copies
Editing of Earlier Modern Copies
The Problems of Editing

III. Bibliography
Earlier Systems
The Leiden System
The Wingspread Convention

IV. Restoration in epigraphical texts
The Nature and Magnitude of the Problem
The Two Extreme Conceptions
Difficulties with the Principle of Extreme Freedom
Examples of Free Restoration
The Factors Involved in Restoration
A Further Example
Various Essential Considerations
A Suggested Solution
Proposed Rules

V. Illustrative example

Epheboi in Neronian Athens
Previous Study, Provenience, Description
Commentary
The Inscription as a Whole

Citation preview

GREEK

*

ROMAN * AND * BYZANTINE SCHOLARLY * AIDS

T h e series o f S C H O L A R L Y A ID S is published by the Editors o f G R E E K , R O M A N A N D B Y Z A N T IN E S T U D IE S as supplements to the quarterly jou rn al. F o r the list o f those currendy available and o f the G R E E K , R O M A N A N D B Y Z A N T IN E M O N O G R A P H S still in print, see the inside back cover. Communications to the Editors m ay be addressed to the Senior Editor, W il l ia m H . W il l is , B ox 4 7 15 Duke Station, Durham , North Carolina 27706,

U .S .A . Orders for purchase and subscription should be placed directly with the Circulation M anager, M iss D o r o t h y R ounds , at the address given on the inside back cover.

CONVENTIONS IN EDITING A Suggested Reform ulation o f the Leiden System

STERLING DOW Hudson Professor o f Archaeology H arvard University

GREEK * ROMAN ' AND ' BYZANTINE SCHOLARLY ' AIDS ' NUMBER 2 DUKE UNIVERSITY 1969

The Jo h n Simon Guggenheim Foundation has generously provided a grant o f funds in support o f the publication o f this Scholarly Aid.

M ad e an d printed in G reat B ritain by W illiam Clow es and Sons, Lim ited, London and Beccles

Dedicated to J

ea n n e and

L ouis R o b e r t

in admiration fo r other things and because they have done mostfor good editorial usage

CONTENTS Foreword I

II

vi

IN T R O D U C T IO N

i

Apologia

i

General Definition o f Objectives

2

C O N V E N T IO N S

3

Above and Below the Preserved Text

3

Numbering o f Lines

3

T h e Use o f Editorial Signs

5

[καί] [ . . . ] , [------] m,

Restorations (see also I V infra) Lacunae

5 5

Spaces Left Blank by the Letterer

6

Doubtful Readings

6

llllll

Attrition

7

[[καί]]

Rasurae

7

R eading Clear, Interpretation Unknown

9

καί

1TEON

καί

Strokes Clear, Letters Unknown

10

Parts R ead Earlier, Now Missing

10




Additions by the Editor

11

(

)

Substitutions b y the Editor

11




Letters Left Incomplete by the Letterer

11

{

}

Suppressions by the Editor

11

(

)

Resolutions o f Abbreviations and Ligatures

12

Note on Pointed Brackets

12

Editing of Earlier M odern Copies

12

The Problems o f Editing

13

CO N TENTS

III

B IB L IO G R A P H Y

14

Earlier Systems

14

The Leiden System

14

Official Publication

14

Summaries o f the Leiden System for the User

15

Discussions o f the Leiden System

16

The Wingspread Convention IV

17

R E S T O R A T IO N IN E P IG R A P H IC A L T E X T S

20

The Nature and M agnitude o f the Problem

20

The Tw o Extreme Conceptions

21

The K irchner Principle

21

The Principle o f Extrem e Freedom

22

Difficulties with the Principle of Extreme Freedom

23

Examples o f Free Restoration

24

The Factors Involved in Restoration

25

A Further Exam ple

V

26

Various Essential Considerations

27

A Suggested Solution

29

Proposed Rules

3°

IL L U S T R A T IV E A

th ens

E X A M P L E : E p h ebo i

in N ero n ian

(IG I I 2 1989)

32

Previous Study, Provenience, Description

32

New T ext

34

Plate

facing 34

Commentary

34

The Inscription as a Whole

37

Foreword E x p e r ie n c e has shown that organizational formulations easily result in un­

satisfactory compromises, obscurities, and omissions. Doubtless one person, act­ ing alone, is no less liable to error, but he can at least try to satisfy that one person’s notions, such as they m ay be, o f thoroughness and clarity. In much o f what follows, the expression is unavoidably dogmatic, but like the Leiden pro­ posals, everything is recommendations, not laws; whatever authority inheres in the recommendations must derive from inherent reasonableness, if there is any, and from reasonableness alone. O r rather, since we are after all in the realm o f convention, where ultimately it is usage which alone matters, and alone estab­ lishes law, I have understood “ reasonable” to mean “ reasonable in the light of present practice.” For better or worse, I have sought no official backing. L . Robert and Z. Stewart have read the whole with discernment, but I wish not to involve them or any other person. E. L . Bennett J r , H. Bloch, J . Chadwick, W. V . Clausen, K . M . Clinton, D eC. Fales J r , E. W. Handley, J . H . K roll, P. L . M acKendrick, G. N agy, G. M . Quinn, R . S. Stroud, L . L . Threatte J r , S. V . Tracy, L . B. U rdahl and J . C. W aldbaum also have given valuable assistance. I f what is con­ tained herein leads to useful discussion and agreement, in practice (one would hope) as well as in theory, and also, if it is needed, in organizational action, I for one shall be pleased to conform to the eventual decisions. A ll that matters is to serve clarity, simplicity, and adequacy—not to gratify any one person’ s feel­ ings about any particular sign. Certain limitations o f the present effort should be noted. Textual apparatus is not dealt with. There is nothing here on “ style” : m y efforts in that direction, along with those of others in Am erica, are set forth in the American Journal of Archaeology 69 (1965) 199-206 (with abbreviations). Nor have I attempted to deal with usages outside the classical sphere, e.g. cuneiform tablets and the like. The Leiden and other treatments o f conventions have usually cited examples which were fabricated, and which consequently had (it seemed to me) an air of unreality. I have therefore been at pains, except in a few brief instances, to cite instances from actual inscriptions in their context. Some o f these are from my own publications: in searching for instances which are perfectly certain, one is driven to selecting instances which are familiar.

VI

I

INTRODUCTION Apologia

I

N I 93 I Dos Leydemr Klammer System (full references in the B ib l io g r a p h y infra) came into being. Papyrologists had taken the lead, and initially at least representatives o f other disciplines were not as fully consulted as would ideally have been desirable. Nevertheless, during the next decade and a h alf the conventions agreed upon at Leiden came to be widely adopted for the editing o f epigraphical and papyrological texts. Any Classically trained scholar who knows the usages in one o f these fields can often interpret correctly, without re­ course to tables o f sigla and the like, the conventions now used in editing most o f the texts in the other field. But palaeographical texts— the “ authors” — con­ tinued to be edited somewhat differently, and often differently from one another. It m ay suffice to refer to O. Staehlin’s Editionstechnik, 2nd ed. (Teubner, Leipzig tf)1 -!) and P· M aas, Textkritik (Leipzig 1927; later editions, including the Eng­ lish translation [Oxford 1958] are wholly unaffected by Leiden). Several years ago the Leiden1 system was put to a new test. It was urged by the present writer that the Bronze Age texts from the Aegean area should also be edited in such a w ay that, with a minimum of usages necessarily peculiar to the editing o f the Linear B and A tablets (etc.), the diacritical signs should be intelligible to all Classical scholars. Colleagues learned in these writings willingly agreed, and the Wingspread Convention {infra, B ib l io g r a p h y ) is the Leiden system re-stated with a few, readily-intelligible additions necessary for the tablets. Apparently the Leiden System is here to stay. It has worked— can one not say ? reasonably well. No large alteration has been or is likely to be proposed. There is ample evidence, nevertheless, to show that the Leiden system was not perfect or final. N early everyone has his own pet notions, and no one, if he had to start again at the beginning, would devise a system precisely like that which has been adopted. In fact several minor changes have been suggested. In epigraphy there is no agreement in theory or in practice on the proper status of “ restorations” within square brackets. On the use o f dots, theory is clear, practice often delinquent. On several signs used less frequently, there is no clear statement anywhere, neither is there an up-to-date bibliography o f what has been written about the conventions. Nor does any recent article attempt to cover the same ground critically. In addition, the latest official formulation, itself more than thirty years old, o f the Leiden system is out o f print and unobtainable. It left much unsaid and some matters unclear. 1 As between Leiden and Leyden in English, there is no decisive consideration except usage, which favors Leiden.

·+

I

2

CONVENTIONS

IN

EDITING

Evidently therefore the time has come when some reconsideration and re­ formulation might be useful. The task, it seems to me, is one o f carrying out orders— orders interpreted as intelligently as possible not o f issuing commands. The legislators o f Leiden, for instance, chose to alter the meaning o f < > in Greek epigraphy so that, although in nearly all the past century o f publications

Letters (or numerals, etc.) inserted by the editor to supply letters (etc.) con­ sidered by him to have been intended to be inscribed, but which were omitted by error. IG I I 2 3242 lines 5 (part) and 6: άρχοντο; δέ [Αίολίωνοξ] τοϋ Άνπάτρου Φλυέ[ω5 ν]εωτέρου The commentary states that the stone has ANTTATPOY. The man Αίολίων ’ Αντιπάτρου Φλυεύξ and his father are known from other inscriptions. S u b stitu tio n s b y th e E d ito r : Pointed Brackets < ) Enclosing letters (or numerals, etc.) substituted by the editor for letters (etc), actually inscribed, but considered by him to have been inscribed by error, in­ stead o f the (correct) letters given within the pointed brackets. IG I I 2 386 + E M 12564; new text, HSCP 67 (1963) (47-) 49, lines 3 and 4 (parts) : [ ------έπί τήζ A] [ly]ei< 5)os έκτη; π [ρ υτανεία ;------ ] The stone reads E IΑΟΣ, i.e. it has a clear A where Δ was intended. L e tte rs L e ft In co m p lete b y th e L e tte re r: Pointed Brackets < > Especially in inscriptions where the mason had to change from one chisel to another, parts o f letters were sometimes never cut. Most often, e.g. A was left as A. Editors disagree, but strictly should be printed; certainly not, if the reading and intent are clear, A. The subscript dot should be used solely for doubtful readings, never for editorial corrections. Numerous omitted strokes, clearly read, Hesperia 7 (1938) 2-5, 8-9. Here there are so many instances that no effort is made to indicate them except in the commentary. This is allowable: the intent is usually indubitable. S u p p re s s io n s b y th e E d ito r : Braces { } Enclosing letters (etc.) considered by the editor not to have been intended to be inscribed, but inscribed by error. IG I I 2 1990.4, part: έκ των {εκ των} * Ελλήνων In rare instances incomplete letters, which in the final intention o f the scribe were not meant to be inscribed (or written) at all, were left unerased : the in­ completeness was meant to signal the cancellation of the strokes inscribed: The predicament is illustrated in a papyrus, where the facts were detected and clarified by J . Schwartz and H. C. You tie, ΤΑΡΑ g ì (i960) 256: an epsilon was left incomplete where an impulse to write επτά was rejected. Print {p} and explain in the commentary that an epsilon was left incomplete.

12

CONVENTIONS

IN E D IT IN G

R e so lu tio n s o f A b b re v ia tio n s a n d L ig a tu r e s : Parentheses ( ) Indicating letters added by the editor to fill out an abbreviation to the full form o f the word. Hesperia 1 1 (1942) 67, no. 33, is correctly printed in lines 1- 3 :

’ Αγαθή i

[Τύχηι]

[ε]τπ άρχοντος Τϊ (3 Κλ Λ [ -----------] Μελ(ιτέως) στρατ[η]γοΟν[τος δέ, κτλ.] The abbreviations for Tiberios and Klaudios are so fam iliar as not to need expansion; but the demotic might not be known to all readers. (Although the cognomen cannot be restored, the man’s family is known and the demotic is indubitable.) Parentheses are used also to give the full form of an abbreviation inscribed as a ligature. The ligature should be specifically mentioned if it cannot be reproduced. Where the stone has ΓΡ preceded by a name, print Ίτρ(εσβύτερος) ; or in a date, print ·ττρ(ό), or, if the printer can do so, print iff (ττρό). J . H . Oliver, Hesperia Suppi. V I (1941) 89, no. 7, line 15 (part): the inscribed characters are {Inscrs. Brit. Mus. I l l , p. 150 no. 486) :

ΕΥΤΥΧΕΙΤΕ—iff. E. K. ΟΚΤΟΒΡΙΩΝ for which print

εύτνχεΐτε—Trp(ò) ε' Κ(αλανδων) ’Οκτωβρίων N o te o n P o in te d B ra c k e ts T h e signs < ) are ambiguous, and the Leiden convention (reference infra, p. 14) p. 20 proposes Γ 1 for corrections, but these look too much like [ ], and broken type or poor printing might easily change [ to [. Hence in epigraphy this recom­ mendation has not been accepted. Others (e.g. F. W. Schehl, infra) have pro­ posed other signs. In all o f the above instances except restorations and abbrevia­ tions, the critical apparatus ought to state the facts clearly. There is no other w ay to insure clarity o f understanding. To multiply signs is to burden the editor and the reader without guaranteeing success. Instances are uncommon enough so that the critical apparatus will not be unduly lengthened.

E ditin g o f E a rlier M odern Copies A certain number o f inscriptions survive only in the transcriptions made of them usually some decades ago. T o a greater or less extent these copies are defective, and the modern editor, no matter how conservative (epigraphical editors are, or surely ought to be, the most conservative of all), must emend, often extensively. He must put into his text other letters, different from those he 6nds in the copy.

CONVENTIONS

13

For diacritical signs there are two choices, and practice is divided. One choice is to put a dot under the letter substituted by the new editor for the read­ ing in the copy. In practice the dot used thus has sometimes come to mean noth­ ing whatever, except “ a letter, not having necessarily any resemblance to the letter now proposed, was read in this space.” Editorial extremes o f this sort in editing copies by Fauvel, Baumeister, and Vernon are cited by Pritchett, A JA

59 ( 1 955) 57 · The alternative is to treat the copy as if it were an original text and to enclose in pointed brackets letters substituted by the editor. Pritchett has proposed that this be done whenever there is little or no coincidence o f strokes. On the other hand, if the substituted letter is similar, the proposal is that it should be dotted. Thus iota for sigma would be printed , whereas theta for omicron would be dotted, Θ. This would be an improvement, and yet it is far from being a perfect solution. (1) It makes the editor the judge of whether the resemblance is close; he will tend to decide in favor of his own interpretation, and thus he will favor dots. (2) Pointed brackets occupy more horizontal space and do really mar the page (as M eritt urged) more than any other sign. There is only one solution, and that is to print the copy itself as near on the page as possible to the new edition of it. This can always be done : the copy is not like a photograph. Then the editor, if he wishes, can state that all letters altered by him are dotted. More important is to introduce a general statement, based on positive instances, if there is sufficient material, about the accuracy of the copyist; about which letters of the alphabet misled him, and how often. Depending on the style o f lettering, this m ay vary from inscription to inscription (if there are several), and from inscriptions which evidently were well preserved to those which were not.

The Problems o f E ditin g The main problem, surely, is not to change the Leiden system, but to improve its working. More accurate realization, however, o f what is involved, and sharper formulations, are only a part of the task, and the lesser part. The more important sphere o f effort is not legislation but practice. We are all more or less at the mercy of editors. Photographs must always be printed, but they cannot always suffice. The very areas where difficulties arise are the areas where photo­ graphs are likely, especially when dimmed by reproduction on the printed page and reduced in size, to be inadequate. Avoidable editorial malpractice certainly ought to be stigmatized: no one who reads the astonishing instances collected by W. K . Pritchett in A J A 59 ( 1 955) 55—Gr can doubt that. Discussion o f editorial conventions m ay tend to make editorial practice more accurate. But lack o f skill is also a factor, curable only in part by diligence and long practice, by better squeezes and greater patience.

Ill

B IB L IO GR AP HY h e fo llo w in g b ib l io g r a p h y is intended to be complete, at least for epigraphy, and to be critical. Items no longer worth reading are included for their historical value. I hope they are sufficiently criticized; to be explicit, however, I give here the opinion that under D iscussions the only items o f enduring value are the second half of the J./ L . Robert passage, on publica­ tion, and the article by W. K . Pritchett, on dotting letters.

T

E a rlier Systems The Leiden System was not universally adopted at once. For summaries of earlier systems in epigraphy see e.g. Supplementum Epigraphicum Graecum 6 (1932) p. viii, and earlier volumes; J . Kirchner, Inscriptiones Graecae I I iii i 2, page v. Other discordant systems, G. Klaffenbach, Griechische Epigraphik2 (Gottingen ig66) 103. For a more extensive treatment, with older bibliography (p. 9), see the official publication o f the Union Académique Internationale {infra, ed 2., next section; hereinafter referred to as U A I 2). For a brief and convenient sum­ m ary, see Chronique d’Egypte 7 (1932) 268. History o f the use of dots to indicate uncertain readings, W . K . Pritchett, American Journal o f Archaeology 59 (1955)

55- 57· The Leiden System O F F IC IA L P U B L IC A T IO N

Union Académique Internationale, “ Emploi des signes critiques; disposi­ tion de l’apparat dans les éditions savantes de textes grecs et latins ; Conseils et recommandations.” 19 32: J . Bidez and A . B. Drachmann (Paris: E . Champion, 5 Q uai M alaquais). 46 pages. 1 938 : Same, édition nouvelle by A . Delatte and A . Severyns (Brussels, Secrétariat administratif de l’Union Académique Internation­ ale, Palais des Académies; Paris, Société d’Edition Les Belles Lettres, Boulevard Raspail, 95). 50 pages. Although the effort had been begun some years before, the main feature was an attempt to take into account the conventions adopted by the Papyrological Section o f the 18 th International Congress o f Orientalists, which met in Leiden, 7 -12 September 19 3 2 ; the measure in question was passed unanimously on 10 H

BIBLIOGRAPHY

15

September. The object was to recommend usage that might become uniform for the editing of all Greek and Latin writings : i.e. for papyrology, epigraphy, and texts o f authors. A formidable list of scholars eventually contributed opinions. The present brochure, which was the result, records previous usages and pleads for uniformity on a basis of a few simple signs used with a minimum o f change. The second half o f the brochure is on textual apparatus. There are lists of signs, abbreviations, etc., and a full index. SU M M ARIES O F T H E L E ID E N SYSTEM F O R T H E U SE R

J . Kirchner, Inscriptiones Graecae, I I iii i 2 (Berlin 19 35 : Attica, Dedications) page v. The examples are clear, but the whole is compressed, and the “ erased” 'letters under [[ J d ) ,in ΕΔΟΞ E 3SQJ ΔΗ Μ Ω I, should be printed in brackets, εδοξε |[t o i ] δήμωι. The final entry, A in notandis litteris dubiis, should be altered to state that the letter is partially damaged, and that what remains is compatible with the proposed reading, but that the actual remains, independently o f the con­ text, do not make the reading mandatory. I f e.g. the left stroke only o f a pi is legible, the letter should be printed with a dot, Π ; but if the two strokes forming the upper right corner of a p i are definitely preserved, even though only partially, print TT without a dot. Supplementum Epigraphicum Graecum, ed. J . J . E. Hondius, 7 (Leiden 1934) p. viii (in line 13 for εδο ξι read εδο ξε), and subsequent volumes, e.g. ed. A. G. Woodhead, 18 (1962) p. xii. Same examples as in IG I I 2; the erasures are however properly treated, but correct the part on subscript dots as supra under J . Kirchner, IG I I 2. Μ . N . Tod, Greek Historical Inscriptions (I2 [Oxford 1946] p. xx has the old system), I I (1948) p. viii: Leiden system. Alter [[ J to read : enclose letters deliberately erased in ancient times but still legible; [[[]]] enclose letters believed by the editor to have stood in the erased area. Correct the part on subscript dots as supra under J . Kirchner, IG I I 2. B. D. M eritt, Η . T . W ade-Gery, and M. F. M cGregor, Athenian Tribute Lists, I I (Princeton 1949) 7· The Leiden system, but with reversion to the rectangular enclosure | | for the second (most recent) inscription in a rasura. Apart from lack o f sanc­ tion, this has the disadvantage of not being immediately intelligible: the reader might assume it surrounded letters read from the first inscribing. Dots under letters “ epigraphically doubtful” is also a statement liable to misinterpretation. G. Klaffenbach, Griechische Epigraphik2 (Gottingen 1966) 10 2 -10 3 . Alter as supra on dots. On erasures note that an erasure need not have been made b y the original mason, or by any mason.

6

CONVENTIONS

IN ED ITIN G

American Journal o f Archaeology 69 (1965) 200. Expand to conform to IG ID and SEG supra. The double brackets [[ ]] enclosing erasures should not appear to be o f two different kinds, but should all be the same. A. G. Woodhead, The Study o f Greek Inscriptions (Cambridge 1959) 6 - 1 1 . Page 9: the letters restored in the erasure should be printed not £Όττελλίωι]] but [[[Όπελλίωι]]]. Page g: dots: alter to read as supra. Other­ wise these pages are the best and fullest recent guide. DISCUSSIONS O F T H E 'L E ID E N SYSTEM

Unsigned, “ Essai d’unifìcation des méthodes employees dans les éditions de papyrus,” Chronique d'Egypte 7 (1932) 285-287. The first announcement : a brief account o f the procedure at Leiden, a brief and faulty sum m ary o f the signs, and some good recommendations for papyrological publications. (Program of the session o f 1 1 September, p. 12 9 ; President, M . D. Cohen; his speech, pp. 1 3 1 - 1 3 3 ; seven additional papers on other subjects, after the first of which came the “ Essai d’unifìca­ tion.” The brochure had come before the session o f the previous day, after four papers had been read, and along with three others bearing on publica­ tion. These three are the ones that are summarized next hereunder.) B. A. van Groningen, “ Projet d ’unification des systèmes de signes critiques,” Chronique d ’Egypte 7 (1932) 262-269. Papyrology has a central position, since it deals with documents, as does epigraphy, and texts o f authors, as does palaeography. Discussion of signs: some o f the views set forth have not been adopted. Η. I. Bell, “ Note on Methods of Publication,” ibid. 2 70 -271. Recommendations for ease o f use, concerned with larger aspects than signs: see the official brochure, and L . Robert {infra). A . S. Hunt, “ A Note on the Transliteration o f Papyri,” ibid. 272-274. Concerned with dots and underlinings to indicate doubtful readings. A sound discussion, but no absolutely precise formulation is offered. See Pritchett {infra), who, however, did not know Hunt’s article. U . Wilcken, Archivfiir Papyrusforschung 10 (1932) 2 1 1 - 2 1 2 . Notes thirty years o f uniformity within papyrology; contrasts epigraphy and texts o f authors. Declares that the Greek Corpus will follow the Leiden system, despite difficulties and anxieties. B. A. van Groningen, Mnemosyne, n .s . 59 (1932) 362-365. Sim ilar to the article noted above, but briefer. F. W. Schehl, American Journal o f Archaeology 58 (1954) 23 n.32. Urges < > for corrections by the editor to replace errors in the text; « » for additions by the editor to make up for omissions in the text. This has not been adopted, it is too elaborate. J./ L . Robert, La Carie, I I (Paris 1954) 9 -13 . Criticism o f the Leiden system, in favor o f keeping some o f the old epigraphic conventions; and recommendations, all excellent, for larger aspects of publication. See also R E G 68 (1955) 186.

BIBLIOGRAPHY

I?

G. Klaffenbach, Gnomon 27 (1955) 239-240. This is a brief discussion at the end o f a review o f J./ L . Robert, La Carie·. departure from the Leiden system would be deplorable ; the system adds no printing cost or difficulty; Teubner has adopted it. W. K . Pritchett, “ Dotted Letters in Greek Epigraphy,” American Journal of Archaeology 59 (1955) 5 5 -6 1. This should be read entire. Summary: see the following item, p. 37 1 n .13. W. K . Pritchett, “ Ancient Athenian Calendars on Stone,” University of California Publications in Classical Archaeology, voi. 4, no. 4 (Berkeley 1963) 267402. The section on restorations, pp. 373-382, is a strong statement, with examples drawn from one o f the most difficult fields. This essay alone should suffice to alter practice. (But about one related aspect, stoikheion, pp. 382-384, more consideration is needed o f how masons work.)

The W ingspread Convention Institute for Research in the Humanities (University o f Wisconsin, M adi­ son, Wisconsin 53706, U .S .A .; 1962; two pages; obtainable on request). “ Notae Diacriticae in Edendis Textibus Mycenaeis Minoicisque, a Tertio Colloquio International! Studiorum Mycenaeorum in ‘Wingspread’ con­ vocato, editoribus commentatoribus commendata.” After discussion in a session of the colloquium (7 September 1961), the Convention was drafted, in consultation with various scholars, by a com­ mittee consisting o f J . Chadwick and S. Dow. The effort was directed (a) toward bringing editorial usage into conformity with the Leiden system, and (b) toward suggesting special usages, which might be uniformly adopted in this field, that would make for greater clarity in editions of Minoan and Mycenaean texts (Linear A, Linear B, and earlier scripts). (a) The Leiden system was adopted without change. The note on restorations between square brackets m ay be quoted: “ The Leiden system makes no provision, and scholars have experimented but have never agreed on one, for indicating what restorations are positive and what are conjectural. Restorations o f letters or individual signs which make normal spellings in words which are themselves indubitable, or o f formulae which are amply attested in the given context, are posi­ tive; but where certainty is unattainable, the editor must be free to print other letters, signs, or words, which he feels should be restored to indicate the possible sense. In this situation the simplest courses are to insert a small interrogation point, and to comment in the apparatus ; or to leave space blank and to give the suggested restoration wholly in the apparatus.” (b) Certain exigencies recurring frequently in the tablets led to the adoption o f the following, which is supplementary to the various Leiden usages:

ι8

CONVENTIONS

IN E D IT I N O

[----? — ] Dashes and interrogation point to indicate doubt whether a sign or more than one sign was inscribed in the area. N .B . A space between a bracket and a sign, or a sign and a bracket, indicates that the sign is believed to be the beginning, or in the other case the end, of a word or sign group ; if the bracket immediately precedes or follows the sign, this indicates that the word or sign group m ay or m ay not be complete, e.g. : 50 [ means that no digits can have followed the numeral; 5o[ means that it is impossible to tell whether or not any digits followed. Similarly, — ] te-ra means that there is enough space, or a divider, between the broken edge and the sign te to show that this is the beginning o f the word. — ] te-ra means that there is insufficient evidence to show whether or not the word is complete. — ~\-te-ra means that there is a trace o f an unidentified sign preced­ ing te, and there is reason to think that it is part o f the same word. (In Classical epigraphy positive strokes o f the unidentified sign would be shown with shading. This is not feasible for the syllabaries, with their more elaborate signs.) Areas broken aw ay: sup. mut. (scil. supra mutila) above the first line, or inf. mut. (scil. infra mutila) below the last line, indicate that the beginning, or end, o f the tablet is missing. I f the missing part can be delimited, the text should state, e.g. : sup. mut. desunt 3 versus inf. mut. desunt ca. 3 versus

In the eight years that have elapsed since the Wingspread Colloquium, usage has still varied with regard to hyphens and commas. For illustrations o f the most recent (and I think, most sensible) practice, see J . Chadwick, BSA 58 (1963) 68ff. There are still proposals that need study : , for the divider; so that it shall not be overlooked, leave space on either side, as in pe-da , wa-tu. This is convenient for the editor and, if in fact it leads to no misunderstand­ ings, harmless. But when one thinks o f scholars in other fields trying to use texts in which all the words are separ­ ated apparently by commas, one hesitates. It is deplorable that the normal divider (at least) cannot be conveniently represented.

BIBLIOGRAPHY

19

Accent marks as in ra’ in place of subscript figures as in raz are favored by some. This usage, derived from a convention used in transcribing Hittite cuneiform, is another instance of willingness to risk misleading non-specialists. The sub­ script figure can hardly mislead. ‘Broad’ transcriptions, as contrasted with transcriptions that use the highly accurate diacritical marks, e.g. hyphens, dividers. Broad transcriptions are to be used for a word or two quoted in the course of a sentence in a modern language. Good authorities sanction broad transcriptions, which in m any contexts may be safe. In Classical epigraphy scholars have learned, however, that it is wise to make all quota­ tions accurate: in the best usage, broad transcriptions are not found.

IV

R E S T O R A T I O N IN E P I G R A P H I C A L TEXTS The N ature and M agnitude o f the Problem t h e c o u r se o f the centuries, most Greek inscriptions on most sites have been damaged. The principal causes o f damage, never systematically studied, but informative in many instances, are : breakage, for various rea­ sons; damage by traffic, as in the case o f paving blocks and door-sills; erosion by water, and other damage by weathering, in the case of inscriptions long exposed ; soil damage o f various kinds when inscriptions are long buried ; hard cement and other damage incurred when inscriptions are used for building; damage b y direct human action, as in Attika by Philip V , Sulla, the Heruli, et al. ; and accidental damage in modern museums. Small wonder that an inscrip­ tion o f more than a few lines preserving all its letters is a rarity; and that on an average, individual inscriptions (except columnar grave monuments, the largest class in Attika) survive to the extent o f less than half. M ore letters are lost than survive. Next to providing an accurate version o f what is preserved, the chief and hard­ est task o f the epigraphical editor is usually to restore the letters that are not preserved. In its totality the task is enormous, never-ending, almost omni­ present. No one o f the problems that arose at Leiden is remotely comparable in magnitude to that o f clarifying usage in regard to restoration. In practically all instances, the sign for restorations long has been, and is, the pair o f square brackets [ ]. About what sign to use, there could and can be no dispute. Leiden m ade no change, and no one proposes any different sign. But about what to put inside the square brackets— whether to put only what is vir­ tually certain, or whether it is allowable to put also matter that m ay be entirely unsupported and conjectural— there has been much dispute, and practice varies extremely. In one aspect or another, restoration has been discussed, by L . Robert and others, time and again. Nevertheless I venture to think that there is room, and need, for a treatment which will attempt to be systematic. Extremists are few, perhaps, but they are entitled to a full and fair hearing. I have tried to quote opinions so fully that there will be no danger o f misrepre­ sentation by suppression o f the context; and I have not made use o f invidious italics mine. On the other hand, let lay readers keep firm ly in mind the fact that once a restoration is launched, especially if it is made to appear “ authoritative,” so

N

RESTORATION

IN E PIG R A P H IC A L

TEXTS

21

it m ay be copied and utilized for years, pernicious in itself and setting an example for others. Certain fundamentals are not