Year: 2020
Type: conference-paper
Abstract: This paper presents the first proper syntactic treebank for Akkadian, an ancient Semitic language which can only be reconstructed from its textual data.We introduce our corpus of early Neo-Assyrian royal inscriptions, present some typical syntactic constructions of this genre and discuss the morphological and syntactic choices we have made.For developing a gold standard for morphological annotations, we tested the manually annotated material against BabyFST, a morphological analyzer of Akkadian.We also tested the reproducibility of the syntactic annotations using the TurkuNLP neural parser.1 Sometimes referred to as ARI (Assyrian Royal Inscriptions). 2 This is a simplification, but a number of royal inscriptions from the eighth and seventh centuries BCE are syntactically much more complicated; consider, e.g., Sargon II's famous Eighth Campaign.3 Grayson (1991), also known as RIMA 2 (references to the volume so below; Q-numbers, also below, refer to Oracc text IDs), contains the inscriptions of Tiglath-pileser I and his successors until Tiglath-pileser II, too, but these kings are usually considered Middle Assyrian.4 Grayson, 1991: 189-397.5 Standard Babylonian is a literary variant of Babylonian dialect (for its use in Neo-Assyrian royal inscriptions, see Frahm, 2019: 144-145); it was never a spoken language.6 Assyrianisms in this corpus were already discussed by Deller, 1957a and b. 7 .By making editions of thousands of cuneiform texts available online for everyone with an Internet connection, Oracc has laid the foundation for Digital Assyriology.8 .9 Partially the problem relates to the wide use of different writing systems among Semitic languages; cf., e.g., Zitouni, 2014: 35.In this context, we are not concerned with the correct interpretation of the syllabic C(onsonant)V(ocal)C(onsonant) signs whose reading values do give Assyriologists some trouble.10 One can find treebanks of these languages at Universal Dependencies ( ).11 Akkadian texts rarely make use of a word-divider or any other device that belongs to the area of punctuation.However, especially literary texts may occasionally leave gaps between words but, as a rule, the original texts do not delimit word boundaries by "whitespace" characters.12 On the size of the Akkadian text corpus, see Streck, 2010.
Institution: University of Helsinki
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
OpenAlex ID: W3097252586