Topic and Focus marking in an Italian corpus: some results of algorithmic measurements and structural interpretation.

The results of an experiment on a corpus of spoken Italian suggest a partly new hypothesis on how the main prominence may be interpreted by speakers in the marking of Information Structure (IS). A “topologic” concept of Prominence can be conceived of, as endowed with the function of demarcation between units, beside and before their culmination and characterization. Much of the process by which speakers interpret the IS of utterances may rest upon this, the specific intonational contours of IS units being probably motivated by other functions. In addition, many real utterances seem not always to signal the distinction between Topic-Focus and Broad Focus clearly, remaining rather underspecified in this respect, with no serious effects on communicative dynamism in the subsequent discourse. Such results, obtained by measuring Prominence as a complex entity (not only intonational in nature) strikingly follow the law of least effort. The used algorithm receives confirmation by the fact that automatic measurements and human evaluations of IS patterns show a very high percent of coincidence.