Automatic Orality Identification in Historical Texts

Proceedings of the 12th Conference on Language Resources and Evaluation (LREC 2020), pages 1293–1302 Marseille, 11–16 May 2020 c⃝ European Language Resources Association (ELRA), licensed under CC-BY-NC 1293 Automatic Orality Identification in Historical Texts Katrin Ortmann, Stefanie Dipper Department of Linguistics Fakultät für Philologie Ruhr-Universität Bochum {ortmann, dipper}@linguistics.rub.de Abstract Independently of the medial representation (written/spoken), language can exhibit characteristics of conceptual orality or literacy, which mainly manifest themselves on the lexical or syntactic level. In this paper we aim at automatically identifying conceptually-oral historical texts, with the ultimate goal of gaining knowledge about spoken data of historical time stages. We apply a set of general linguistic features that have been proven to be effective for the classification of modern language data to historical German texts from various registers. Many of the features turn out to be equally useful in determining the conceptuality of historical data as they are for modern data, especially the frequency of different types of pronouns and the ratio of verbs to nouns. Other features like sentence length, particles or interjections point to peculiarities of the historical data and reveal problems with the adoption of a feature set that was developed on modern language data. Keywords: conceptual orality, historical data, German 1. Introduction Human language is used for communication in two major forms, written and spoken. Depending on the medium, utter- ances can differ significantly, as both discourse modes place different demands on the language user. While spoken dis- course requires online processing, thus depending on work- ing memory, written discourse may be processed multiple times and at any desired speed. However, linguists recognize that besides this medial distinc- tion there is also a lot of variation within discourse modes. As Halliday (1989, 32) explains, “‘written’ and ‘spoken’ do not form a simple dichotomy; there are all sorts of writing and all sorts of speech, many of which display features characteristic of the other medium.” This variation within discourse modes was termed concep- tual orality or literacy by Koch and Oesterreicher (1985) and it can come in handy, if the historical development of cer- tain linguistic phenomena related to discourse mode should be investigated. For historical time periods, obviously, only written language data is available. Hence, in order to gain knowledge about historical spoken discourse, it is necessary to identify written texts that are close to the oral mode, i.e. conceptually oral or spoken-like texts. In this work, we test on historical German texts a set of gen- eral linguistic features that we have proven to be effective for the automatic identification of conceptual orality in modern language data (Ortmann and Dipper, 2019). The remainder of this paper is structured as follows: Sec- tion 2. gives a short overview of the related work. Section 3. describes the historical data from different registers and Sec- tion 4. introduces the linguistic features used in this study. In Section 5. we use the features to classify the historical texts according to conceptual orality and inspect the results obtained both between …

Content truncated to 500 words