CA aims at a descriptive analysis of communication failures, which it views as breakdowns or trouble occurring during conversation. More recently, some authors have proposed that a specific case of communication failure, misunderstanding, be viewed as an integral part of the comprehension process rather than just a breakdown (Dascal 1985; Weigand 1999). In line with such a perspective Bosco et al. (2006) have claimed that all communication failures, not only misunderstanding, integrally belong to the cooperative process (Grice 1957) in which agents are involved during communication. Within a cooperative model of communication, the replies received from a partner provide the speaker with the grounds on which to realize that a communicative attempt has failed. Recognizing that a failure has occurred provides in its turn a starting point for repair. Focusing on the complexity of the cognitive processes involved in failure recognition and repair and following the assumptions of cognitive pragmatics theory (Airenti et al. 1993; Bara 2008), Bosco et al. (2006) have proposed a taxonomy of the types of failures that may occur in communicative interaction. These are failure of the literal meaning, failure of the speaker’s meaning, and failure of the communicative effect (i.e. the partner’s refusal to accept a partner’s communication act). A failure may also involve a combination of two or all of these types. In order to achieve his or her communicative goal, and depending on the kind of failure that has occurred, a speaker may employ different repair strategies. A speaker may simply repeat what he or she said (in the case of failure of the expression act), reformulate what he or she said (in the case of failure of the speaker’s meaning) or change the content of what he or she said (in the case of failure of the communicative effect). This taxonomy allowed Bosco et al. (2006) to generate empirical hypotheses about the relative difficulty of recognizing and repairing different kinds of failure that were confirmed by empirical evidence obtained from 3- to 8-year-old children. In particular, it was found that it was easier for children to repair a failure of literal meaning than a failure of speaker’s meaning, whereas repair of communicative effect was the most difficult.