Measuring Change in Irish Literature

University College Cork (Ireland)

Objectives: Computational techniques make it possible to identify authorship with high confidence levels on the basis of linguistic features that fall below the level of consciousness. Using techniques for cultural analytics like stylometry, one can pursue questions about the centrality of subject matters to Ireland’s literary canon, authorial language and individual style through time. Rachel will: 1) conduct a high-level semantico-stylistic study of the oeuvres of a large dataset of Irish authors; 2) track and understand semantic and conceptual change in the Irish national literary canon over time; 3) explore intra-generational and inter-generational similiarities and differences in meaning and style; 4) expose the lasting influence of ‘ad hoc’ (external, historical) events on language production (war, religion, etc.). Rachel will present research at D7.3 and D7.4.

Expected Results: 
1) Best practice recommendations on methods for detecting concepts and semantic change in relatively small, highly specific samples of language use.
2) Case studies on the impact of ad hoc (external, historical) factors on language variation and use.
3) A rich analytical description of the relationship between conceptual change at community and individual levels in the dataset. 4) Innovative application of word embedding methods for stylometry and computer-assisted literary analysis. 

Planned Secondments: University College Dublin (Insight Research Centre for Data Analytics; Year 2, 3 months; researching nation, gender & genre in Irish fiction). VICESSE (Year 2, 3 months; analysing stylistic changes in the discourse of security reporting in newspapers and policy reports). Also visiting: USFD (NLP innovation incl. word embedding), USAAR (expertise in text type analysis, historical English linguistics).