Table of Contents
Transliteration and the diacritics
Thoughtout the history of Assyriology, cuneiformists have used a wide variety of orthographic conventions for the representation of cuneiform signs in alphabetic script. The set of orthographic conventions known as transliteration consists of a sign-by-sign representation of the cuneiform signs in question: each cuneiform sign is given a reading and when these readings form part of a single word, they are linked by hyphens.
In the figure, note that there are four distinct cuneiform signs. Each of these cuneiform signs has a name, which is generally represented in capital letters and refers to the sign itself rather than one of its interpretations. The name of each sign is immediately below it. The third line in the figure is an example of transliteration. e2 dub-sar-ra means 'the house of the scribe', e2 means 'house', dub-sar means 'scribe' and ra codes the genitive case /ak/ which follows dub-sar. Note that transliteration does not differentiate the logograms from cuneiform signs being used syllabically: both e2 and dub-sar are logoraphic, while ra is used here syllabically to represent the genitive case. It does, however, indicate word divisions and provide a rough approximation of how the sequence of cuneiform signs was pronounced.
Another set of orthographic conventions that should be carefully distinguished from transliteration is transcription. Whereas transliteration attempts to mimic the written form of the cuneiform signs in question, often misrepresenting the linguistic or grammatical structure to some degree, transcription focuses on the grammatical or linguistic structure at the expense of the written orthography. The transliteration in figure 1, for example, seems to indicate that there are two /r/'s in the phrase, the final consonant of dub-sar and the first consonant of the suffix that follows, but this is in fact not the case. In transcription, redundant elements are removed and the morphological structure of the phrase is indicated in one way or another. Sumerian is almost always rendered in transliteration rather than transcription, so there are no conventions for the representation of Sumerian in transcription. In figure 2, I have simply put a morphological gloss in the row for transcription.
The issue that is raised by the contrast between transliteration and transcription in figure 2, namely the interpretation of ra as simply /a/ rather than /ra/ is known as auslaut.
Auslaut is a German word that refers to the sound at the end of a given word or stem. When a Sumerian word ends in a consonant and is followed by a suffix of one kind or another, the auslaut is repeated in most orthographic traditions. The repetition of this final consonant is not meant to indicate that the final consonant is doubled, it is merely an orthographic convention. In figure 2, for example, dub-sar-ra would generally be morphologically analyzed as /dubsar-ak/ with loss of the final consonant of the genitive suffix in word-final position and orthographic doubling of the final consonant of the nominal stem in order to indicate the auslaut. It should also be noted, however, that auslaut phenomena give rise to a variety of ambiguities. Thus, in figure 2, two interpretations of dub-sar-ra are possible: /dubsar-ak/ and /dubsar-ra/, the former with the genitive suffix and the latter with the dative suffix.
Certain consonants, notably "shin," do not use the auslaut system as described above at certain points in the history of the writing system and only show irregular usage in others. In many if not all orthographic traditions, there existed a small amount of variability in the use of the auslaut in representing a nominal suffix or postposition: Jagersma, for example, has noted that both engar-re2 and engar-e exist in the textual record (personal communication 1999), but it must be said that these two forms do not both occur in the same archive: engar-re2 is the conventional representation of /engar-e/ (farmer-Erg) in the Old Sumerian administrative materials, while engar-e only seems to occur in the Ur III materials. One example of such a contrast within the ED IIIb corpus is particularly telling: both agrig-ge and agrig-e seem to occur, but there is only a single occurrence of agrig-e in the ED IIIb corpus and it belongs to an unpublished Leiden text (LB 3) that we only have in a provisional transliteration without an accompanying image. The other twenty-seven occurrences of the word are consistently written agrig-ge.
In general, it is best to assume that in any particular time and place, a particular orthographic rule for the representation of auslaut held sway and that where systematic variation seems to be present, collation and perhaps reconsideration of the archive as such is in order.
In the ED IIIb period in particular, auslauts were quite systematically represented where the stem of the stem of the word ended in a consonant. The primary exception to this is the treatment of sibilants: us2-sa, the nominalized form of the verbal root us2 'to lean; to attach', regularly occurs with the auslaut resumed in the orthography of the nominalization. The nominalization of other verbs, however, such as tusz 'to sit' shows significant variation between tusz-sza4 with orthographic indication of the auslaut of the verbal root and tusz-a with no such indication.
Semantic determinatives and phonetic glosses
Cuneiform signs that aid in the interpretation of a string of cuneiform signs are known as either semantic determinatives or phonetic glosses. Semantic determinatives are signs that identify a semantic class that the referent of the word belongs to, while phonetic glosses help the reader to choose the proper reading or pronunciation of the signs that the phonetic gloss qualifies.