Computational approaches to orthographic variability in cuneiform bilinguals: The case of scriptio plena in Hattic

The issue of textualizing different languages into cuneiform in the archives of Hattusa, and the spelling strategies adopted to represent their distinct phonetic and phonematic realities, is well known among Anatolian scholars. In this paper, we aim to provide a quantitative approach to this issue, focusing on the phenomenon of scriptio plena in Hattian documents, both Hattian and bilingual with Hittite. Our goal is to complement the traditional philological approach with quantitative data derived from computer-based analysis. We concentrate on spelling strategies, always taking the written syllabogram as our unit of analysis rather than the syllable it represents, whether alone or in combination with others. After a brief status quaestionis (§2), we detail how to structure data collection and parsing using automated strategies (§3) and apply statistical methods to our corpus (§4), first investigating the possible correlation between scriptio plena of vowels and scriptio geminata of consonants (§4.1) through Pearson correlation and odds ratio; then analyzing the syllabographic environments in which scriptio plena is more likely to appear in Hattian, with a comparison to Hittite (§4.2). Our analysis shows that in Hattian, scriptio plena may function as a lexico-prosodic marker, especially at word boundaries (§5). All the raw data and tables are available in the ‘Supplementary materials’ folder on Zenodo.