Introduction

For over 3,000 years, numerous cultures throughout the Near East used a borrowed and adapted version the Sumerian cuneiform logosyllabic writing system to express their native languages. This writing system, in various customized iterations, was used to write Sumerian, various dialects of Akkadian, Hittite, Hurrian, Ugaritic, Elamite, and others. It was employed over a broad chronological and geographical range, from personal names attested in early third millennium Early Dynastic IIIa texts from Tell Abu Salabih and Fara to likely the first century A.D. (cf. Hunger and de Jong [2014] ZfA 104/2, pp.182-94).

Over a long and complex history, the Sumerian values of the hundreds of signs were adopted and adapted by scribes of later languages. Additional values for signs also developed and entered various scribal traditions.[1] The result is a writing system that is very complex when viewed as a single system.

First, any given sign can express various values. These values fall into three basic functional categories: phonogram, logogram, and determinative. This tripartite definition is reductionist, but provides some basic functionality, both for creating and formatting textual data.[2] 

Defining Terms

A phonogram, also called a syllabogram, is a sign that represents a syllable. There are four basic syllable types expressed by phonograms.

V – a single vowel: a, i, e, u

VC – a vowel-consonant combination: aš, ib, il etc.

CV – a consonant-vowel combination: nu, pa, ša, etc.

CVC – a consonant-vowel-consonant combination: bad, zar, tum, etc.

In transcription, phonograms are usually written in lower case, often in italics. A sequence of phonograms in a single word is typically separated by dashes. The sequence of syllables “sounds out” the word. Vowel length is not indicated in the writing system, which leaves some uncertainty. Note in the following examples, the length of the a-vowel is not expressed in the cuneiform writing system and is not added as a level of interpretation to the transcription. Only in the normalized transcription do scholars indicate vowel length.

šar-ru, šarru, “king”

ḫu-ra-ṣú-um, ḫurāṣum, “gold”

Scope

The Standardized Cuneiform Sign List represents an attempt to model a universal and comprehensive list of Sumero-Akkadian signs. The list is standardized in the sense that it represents an attempt to create an architecture into which every sign from every logosyllabic writing system in the Sumero-Akkadian tradition can be recorded. Signs and values used in Old Assyrian are recorded alongside signs and values used in Elamite, Hittite, and various dialects of peripheral Akkadian. As such, this database of signs represents more than a sign list for any given language, dialect, or period.

On a practical level, the sign list is used for importing and performing automated processing of textual data at the data ingestion stage. The import process recognizes the transcribed text based on project-level formatting choices, identifies the signs, and creates the relationships between the epigraphic units and discourse units. In short, when given a text transcription that follows the necessary transcription rules, the import process creates a fully linked and modelled database item, i.e. a text. 

Here are the first few lines of an Akkadian text from Ras Shamra as given to the import tool.

Recto

iš-tu UD.DIŠ.KAM an-ni-im

a-na pa-ni mam-mi-is-tam-ri

DUMU níq-me-pa LUGAL URUú-ga-ri-itKI

Note that the transcription given to the database does not contain additional codes that would serve to markup the document. There are no complex prefixes to set off words or lines. Further, there are no tags or elements to define or classify the content. These cumbersome elements are removed from the interface. Instead, the import tool uses project-level specifications to ingest the formatted transcription. This strategy allows the researcher to import already existing transcriptions from their own files or from other sources, with very little technical formatting. 

The user can load or copy from a formatted document (like Word .docx) and paste into OCHRE. Alternately, transcriptions can be typed directly into a pending content pane from which the import tool will ingest the textual data.

The import tool examines the transcription and creates the following text in the database. Based on a few settings selected at the time of import, the database understands the section heading, the horizontal dividing line, each individual sign, and the relationship between signs and words. The import can assign line numbers automatically. 

A click on any sign in a text reveals the links created on import. In this case, the transcription of the first sign in line one is recognized as the reading iš of the IŠ sign. The script unit that represents that sign is linked to the epigraphic unit. The transcription given at the time of import populates the transcription field. No allograph has been specified.

The Script Unit

In the OCHRE system, a script unit represents the minimal unit of writing. In alphabetic writing systems, the script unit is a letter. In the logosyllabic Sumero-Akkadian writing system, a script unit is a sign.

Each script unit is characterized by readings and allographs. An allograph is an alternate written form of a script unit. Any given script unit in the sign list may have various readings and allographs. Alphabetic writing systems typically have one reading per letter. Some systems, like Hebrew square script, have multiple allographs, like the final letters. Logosyllabic writing systems typically have many readings per sign and can have multiple allographs. In the complex Sumero-Akkadian writing systems, the reading represents the various phonographic, logographic, and determinative values of the logosyllabic sign. Because the sign list represents the writing systems from various locations and periods, the use of allographs is necessary to distinguish the various graphic realizations of signs. The use of allographs also allows the creation and use of language-specific Unicode fonts.

The IŠ sign, for example, is defined by eleven phonographic values and two logographic readings. At the time of this writing, this script unit had been linked to over twenty-three thousand epigraphic units across various research projects. 

Linking Data

Beginning with the smallest observable unit and moving to larger units like words, we model textual data as follows. A script unit, defined by its reading and allograph, is linked to an epigraphic unit. The epigraphic unit is the unit of observation that contains the various details observed for a single sign. It links to the script unit. The epigraphic unit can be described with metadata regarding damage, scribal mistakes, sign placement, and the level of certainty of reading.

Each epigraphic unit is linked to a discourse unit. Typically, a series of epigraphic units is linked to a single discourse unit, making up a word. The discourse unit, when it represents a word, is linked as an attested form in a project lexicon. The attested from is linked to constituent epigraphic units. So, the web of links spans from epigraphic, to discourse, to semantic. 

The epigraphic observation includes multiple types of observations. It records the identification of the sign (IŠ), its reading (iš), and an allograph if necessary. In most cases, the identification of the sign is made by the import tool. Most scholars do not transcribe a text by first identifying the name of the sign. The scholar typically jumps immediately to an interpretive level by transcribing a reading of a sign. The epigraphic observation also includes any metadata specifying the certainty of the transcription.

The discourse observation can include various elements. It can include a phonemic interpretation, sometimes called a normalization. It can include a translation. It can include analytical properties.

Discourse units can be grouped together in larger discourse units, from compound words, to phrases, sentences, and paragraphs. At every level of observation, the user has the option to record properties and metadata. Chief among these properties and metadata are time periods and discourse analysis such as prosopography, geographical, or economic analysis.

The level of semantic interpretation is kept separate from the epigraphic observation. While the import process may identify a sign as a logogram, it remains agnostic about the semantic force of the logogram. The semantic identification is made later. At the level of discourse unit where the discourse unit is a word, various wizards help the user perform grammatical parsing, prosopographical analysis, and geographic analysis. At this stage, the user can address the semantic interpretation of the word.

By performing grammatical analysis with the help of various tools, the user can create a lexicon. Following upon the data model outlined above, the words added to the lexicon link to attested forms in texts, which include links to constituent epigraphic units and script units. In the end, the database create a vast network of textual data. 

Organization and Definition

Each sign in the standardized sign list in OCHRE has a unique name, usually the Sumerian sign name. MZL and ABZ sign numbers serve as aliases. As mentioned above, each sign is defined by readings and allographs. A sign may have a single reading or multiple readings. The readings are classified as either phonograms, logograms, or determinatives. Each reading is unique. There are no duplicated readings across all signs. In the Sumero-Akkadian logosyllabic writing systems, a single script unit (i.e. a sign) can be written in various ways. These various writings of the same sign are allographs. Sometimes an allograph is tied to a specific usage (like a specific phonogram) or period (e.g. Old Assyrian). The allograph tab on the script unit represents the various writings of the sign. Each allograph can have a name, a Unicode character code, and links to other script units where applicable. While each script unit and each reading must occur only once in the sign list, there is no such restriction on allographs. However, it seems unlikely that there will be a duplicated allograph.

Because there are sometimes various signs used as logograms for a single number (like 6 or 60 for example), the user should use the logographic name of the specific sign they wish to record. For example, one may transcribe 6DIŠ or 1ÀŠ for the number six, depending on which signs appear on the tablet. Further, the user can specify which allograph of ÀŠ was used. (There are currently three allographs for this sign, meant to represent the most common forms used by scribes.)

Various scribal practices and various transliteration schemes intersect to create many contradictions. Our system tries to strike a balance between an accurate sign list and a usable input method. We also try to accommodate the Unicode points for the various number signs. Our method allows the user to specify the exact numerical sign they read in the text. For example, for the number 2, the user can transliterate 1MIN, 2DIŠ, or 1AŠAŠ.[4] To take advantage of the built in numerical features, the user must supply a number before the reading, even if that number is 1. This multiplier number cues OCHRE to interpret this logogram as a number. For example, if the text has the logographic value IGI for 1,000 then the user must provide 1IGI to the text import. Providing IGI without the preceding number will create a link to the logographic value IGI, but it will not tell OCHRE to treat this logogram as a numeral.

Numbers 70, 80, and 90 do not have discrete Sumerian signs; however, we have character codes for these three. If a user wants to use these specific characters, they'll need to provide 1SEVENU, 1EIGHTU, or 1NINEU as if these were logograms. If the user doesn't care to use these specific characters,they can provide the normal numeric syntax in the pending content, which would be 7U, 8U, and9U.

Note also that the number 600 is a combined sign, referred to in the sign lists and grammars as GÍŠ-U (and sometimes as GEŠU, which strikes me as inaccurate). Our sign and logogram are GÉŠU. 

[1] Cf. Hoffner (2008): 9-24.

[2] It should be noted that there are various transcription traditions for each type of sign. Typical variations include capitalization and italicization. OCHRE does not impose any specific tradition upon any research. However, the project must decide upon a specific set of rules and adhere to it.

[3] For expediency’s sake, we began by documenting only the commonly used signs, leaving aside for the moment the sign that occur only in rare texts such as vocabulary lists. However, with the data model established and tested, the system is prepared to document all known signs in all dialects. For the OCHRE Data Service, this was more than an exercise, it was a prerequisite for our philology projects.

[4] When importing a text, we cannot allow thedash in AŠ-AŠ, because a dash is a syllable separator.