Digital Practices vs Traditional Methods for the Study of the Ancient Near East and its Neighboring Regions
How do digital practices contribute to the research for the study of the ancient Near East? How the use of digital practices can be beneficial for research by contrasting the results achieved till now with the use of traditional methods. Of course, this does not mean that traditional methods need to be avoided, but rather to show how non-digital and digital practitioners can collaborate in order to improve the research.
In addition to Vanessa Bigot Juloux discussion “Why should ancient Near Eastern scholars care to digital practices?” and to illustrate that, H. Jauhiainen suggests the following practical example.
Heidi Jauhiainen (University of Helsinki)
Word2vec is one of the methods in the field of natural language processing which can map vocabulary units to vectors of real numbers in such a way that words appearing in similar linguistic contexts cluster near each other. FastText is a method build on Word2vec that takes into account the subword information by dividing words into sequences of, usually 3 to 6, characters and adding up the information on all the different sequences of a word. In this poster, we present our first experiments with analysing Akkadian texts with FastText. We provide some examples of how it can be used for analysing the semantic contexts where various divine names are attested in Neo-Assyrian texts in ORACC. We are not suggesting replacing traditional methods of philological analysis with automated methods but, rather, proposing that they might be used to find regularities in large datasets and to raise new research questions.
Keywords: Language Technology, Akkadian, Semantic Domains, FastText, Visualization, Deities