Word-length distribution in modern Welsh prose texts

Although very little Celtic data has yet been examined within the Göttingen project on word-length distributions, one set of Q-Celtic data has already been processed – a set of 31 Scottish Gaelic e-mails, for which the best-fit distribution was the 1-displaced hyperpoisson distribution (Drechsler 2001). This study will add data from a P-Celtic language – Welsh – in order to obtain a preliminary impression of whether Celtic is likely to show the same distribution for all its member languages, or whether there are likely to be differences, perhaps along the Q-Celtic versus P-Celtic dimension.