BabyFST - Towards a Finite-State Based Computational Model of Ancient Babylonian

Proceedings of the 12th Conference on Language Resources and Evaluation (LREC 2020), pages 3886–3894 Marseille, 11–16 May 2020 c⃝ European Language Resources Association (ELRA), licensed under CC-BY-NC 3886 BabyFST - Towards a Finite-State Based Computational Model of Ancient Babylonian Aleksi Sahala​1​, Miikka Silfverberg​1​, Antti Arppe​2​ & Krister Lindén​1 1​University of Helsinki, ​2​University of Alberta {aleksi.sahala, miikka.silfverberg, krister.linden}@helsinki.fi, arppe@ualberta.ca Abstract Akkadian is a fairly well resourced extinct language that does not yet have a comprehensive morphological analyzer available. In this paper we describe a general finite-state based morphological model for Babylonian, a southern dialect of the Akkadian language, that can achieve a coverage up to 97.3% and recall up to 93.7% on lemmatization and POS-tagging task on token level from a transcribed input. Since Akkadian word forms exhibit a high degree of morphological ambiguity, in that only 20.1% of running word tokens receive a single unambiguous analysis, we attempt a first pass at weighting our finite-state transducer, using existing extensive Akkadian corpora which have been partially validated for their lemmas and parts-of-speech but not the entire morphological analyses. The resultant weighted finite-state transducer yields a moderate improvement so that for 57.4% of the word tokens the highest ranked analysis is the correct one. We conclude with a short discussion on how morphological ambiguity in the analysis of Akkadian could be further reduced with improvements in the training data used in weighting the finite-state transducer as well as through other, context-based techniques. Keywords:​ Finite-State Transducer, Computational Modeling, Akkadian, Morphology 1. Introduction In this paper, we present a finite-state based general model for Babylonian morphology. At first we give a brief description of the Babylonian and its morphological features. Then we present the modeling principles of our FST approach and measure its performance against a manually tagged corpus. At last, we will discuss potential ways to improve, expand and use the analyzer. 2. Brief Description of the Babylonian Language Akkadian, the language of ancient Babylonians and Assyrians, is known to us by several hundreds of thousands of clay tablets, their fragments and other inscriptions written on various media in cuneiform script. This text material covers a timespan of over two millennia (2400 BCE to 100 CE), making Akkadian one of the world’s earliest written languages, clearly predated only by Sumerian, Elamite and Egyptian (all written already in the fourth millennium BCE). Although a vast majority of the excavated material consists of administrative documents, the Akkadians also left behind a vivid selection of cultural historically significant literary works and inscriptions, including the Epic of Gilgameš and the Code of law of Hammurāpi. Alongside Eblaite, Akkadian is the only known, and the longest surviving member of the now-extinct East-Semitic branch of languages. As a spoken language, Akkadian was gradually replaced by Aramaic after the Persian conquest of Babylonia in 539 BCE, but its prestigious status kept it alive as a literary language until the first century CE. Unlike Sumerian, of which the grammatical description is still widely debated, the understanding of Akkadian morphology and grammar …

Content truncated to 500 words