Linguistic Modelling of Sequential Phenomena

A number of textual aspects can be represented in form of linear sequences of linguistic units and/or their properties. What frequently remains problematic is the lack of knowledge of the statistical properties of a measure. For a linguistically meaningful and valid analysis of lin-guistic objects, linguistic models are required and the most reliable linguistic models are, of course, laws of language and text. In the present paper, we emphasize the linguistic point of view in the analysis of a chronological corpus, the end-of-year messages of the Italian Presi-dents of Republic, and set up the specific hypothesis that the temporal behaviour of the fre-quency of a word is discourse-specific by means of the Piotrowski-Altmann law