La segmentazione delle parole nei testi di alunni di classe prima: costruzione del corpus e analisi del fenomeno

How do children learn to segment words in their writing? What are the characteristics of the improper segmentations produced by children? Answering these questions implies analysing large corpora of texts written by first grade pupils, which are difficult to analyse using automatic language processing (TAL). This paper therefore intends to make a twofold contribution to linguistic and didactic research: on the one hand, it aims to describe an automatic model for analysing texts with unconventional spelling and, on the other hand, to describe, analyse and understand the phenomenon of word segmentation in order to make the learning-teaching relationship more coherent.