The subject of ROAR in the mind and in the corpus: What divergent results can teach us

The use of different research methods in linguistics invariably leads to questions about the convergence and divergence of research findings. Aiming for convergence, while understandable, may distort our understanding of language phenomena, if convergence is seen as the only publishable result. We suggest a place for diverging results in furthering our understanding of the data techniques used to investigate linguistic phenomena. We illustrate this point through an experimental and corpus-based investigation of the preferred syntactic subjects of the English verb ROAR and discuss how deeper reflection on these diverging results leads to a better understanding of the different data types.