Using big diachronic language data to feel the pulse of cultural change

Although most ‘big data’ relate to the present and very recent past, advances in data processing power and retro-digitisation mean that the past can be made more accessible to data-led linguistic analysis as well. One example of work attempting to take advantage of this has been the use the Google Books corpus to track frequencies of certain nouns and proper names in order to establish a new discipline of ‘Culturomics’. Based on research on the connection between phraseology (common ways of putting things in speech communities) and culture, this talk shows how a linguistically careful, data-led methodology can ‘feel the pulse’ of cultural change in a way that reveals remarkably intricate facets of social and cultural change in the speech community whose language is investigated. Applying multi-point diachronic mappings between the frequency developments of phraseological phenomena in corpus data and various types of historical data, the method is able to supply results that are both insightful and falsifiable. This is illustrated using a sample of phraseological expressions of 20th century German as used in Switzerland, drawn from the 20-million word Swiss Text Corpus (https://www.chtk.ch/index.php/en). For example, an investigation into the behaviour of the collocation blaue Augen [blue eyes] mapped onto known historical events, strongly suggests that racial ideologies of a certain type were widespread in the speech community until around the middle of the last century, a point that has hitherto been difficult to establish. Using examples like this, a number of key cultural developments over the 20th century are re-traced, revealing new insights into the workings of language, cultural change and the connection between the two