This paper aims at illustrating a research about the development of software prototypes for automated production of cloze tests (LexMeter) and for the automated modulation of course texts that could match up the real reader verbal competences (Adapter). The verbal competences are defined starting directly from texts’ lexicon and language. We assumed that the virtual reader knows and fully comprehends the meaning of the words included in texts. In this contribution it is discussed what is "behind" the planning of this software, illustrating both its technological and linguistic features. After the definition of the language model, based on lexical and descriptive-statistical aspects, it is described how the automated estimator and the text adapter were designed.
Keywords: Adaptive learning, cloze test, distance education, assessment, reading comprehension, verbal competences.