The Electronic Text Corpus of Sumerian Literature
| ETCSL Home Page | Catalogue | About the ETCSL | How to use this site |

Editing protocols

On this page, some detailed information is presented on the various editing protocols that have been adopted in the preparation of this corpus.


Work stages

For each composition of the corpus, the stages of the operating procedure are as follows:

  1. collect together edition(s), copies, reviews, discussions, and other information taken from the bibliographical database
  2. create draft electronic file of the composite text, or adapt existing text-file
  3. sub-edit text-file for hyphenation, standardised spelling; partially tag
  4. prepare translation; partially tag; proofread Sumerian composite text and translation
  5. SGML tagging of Sumerian composite text using Author/Editor; validate; proofread
  6. SGML tagging of translation; validate; proofread
  7. update SGML bibliography with items from the database
  8. put SGML files onto the Web

Additions and variants

Additions to a text are defined as words, phrases or lines which occur in one source only, or a minority of sources. These are indicated in text and translation, accompanied by a note such as '1 ms. adds the line: ...'. Whole added lines are numbered with letters (e.g. 90A following line 90).

Variants are classified as substantive if they yield a different translation from the privileged composite text. These too are indicated in text and translation, with the word, phrase or line(s) from which they vary also marked, accompanied by a note such as '2 mss. have instead: ...'. .

Line numbering

For electronic purposes it is necessary to have a single sequence of line numbers from beginning to end of a text. These number the lines of a composite text, not of any individual exemplars (even in passages where the text rests on one exemplar only). Lines attested only in post-OB sources have been omitted. In a few cases, slight revision of the numeration of hitherto standard editions has been necessary.

However, segments of the composite text of a composition separated by breaks of indeterminate length have been labelled as such (e.g. Segment A, Segment B) and lines renumbered from 1 at the beginning of each segment (e.g. B1-65, C1-14).


In the transliterated texts and English translations, the following ten categories of word are tagged:


Preferred readings

To ensure a standardised transliteration, a list of preferred readings is maintained. With few exceptions these are all values recorded in R. Borger, Assyrisch-babylonische Zeichenliste, (Neukirchen-Vluyn, 3rd edn. with supplement, 1986), since this is widely available. Within this, in general 'long' values have been preferred to 'short' values, and values attested in Proto-Ea and Proto-Diri preferred to those attested in later lexical lists. There are certain exceptions to these principles. All homophones are marked with subscript indices (not acute and grave accents), because electronic searching is facilitated thereby.

Words containing /ng/

The Sumerian velar nasal consonant /ng/ has been regularly marked in transliteration as a g with tilde, and a standardised list of words containing this sound is maintained, based on the study of J. Krecher in B. Hruska and G. Komoróczy, Festschrift Lubor Matous (Budapest, 1978), vol. 2, pp. 7-73.

Emesal words

Words in the Emesal dialect are electronically tagged and can be searched for. A standard list of Emesal words is maintained, based on the study of M.K. Schretter, Emesal-Studien (Innsbruck, 1990).


All translations are in continuous English prose, in complete sentences as far as possible. The translations may be divided up into paragraphs for ease of reading. No paragraph is larger than one screen in length. Superscript line numbers are used, usually introducing groups of lines.

Damaged, missing or untranslatable passages from one to several words in length are indicated by the mark ...... (i.e. two ellipses). Damaged, missing or untranslatable whole lines or passages are indicated as follows, on a new line:

(1 line damaged/missing/untlear)
(32 lines damaged etc.)
(unknown number of lines damaged etc.).

Individual words of which the translation is uncertain are followed by the mark (?). Sumerian words and names are systematically normalised in translation (e.g. Inana, Mes-ane-pada).

| Return to the top | Email the ETCSL team | Go to the Faculty of Oriental Studies, University of Oxford. |

Written by Jeremy Black. Updated on 30.iv.99.