At a very general level, the present thesis may probably be better characterized as a two-folded enterprise: on the one hand, it is concerned with theoretical issues on the nature of language and cognition; on the other hand, its main goal is to carry out an empirical investigation of a circumscribed and clearly defined phenomenon in a specific language. With regard to the former, the starting point of the present study can be found in the conception of language as a cognitive function. Although mainstream paradigms in the fields of language and cognitive science tend to view language as an autonomous, innate faculty governed by domain-specific cognitive principles, we shall follow proponents of Embodied Cognition and Cognitive Linguistics and adopt the opposite starting point, according to which language is an emergent system shaped by the constant interplay of domain-general cognitive principles and embodied experience in the physical as well as sociocultural environment. The more concrete goal of this piece of work is to illustrate an analysis of a class of (rather prototypical) Italian caused-motion constructions: we shall be investigating a small sample of homogeneous utterances to provide a characterization of the various constructions (i.e. form-meaning pairings which can vary in size and level of concreteness) they instantiate and the way these contribute to the overall meaning of the expressions, allowing the language understander to grasp the semantic content of a message correctly. This goal will be pursued adopting the Embodied Construction Grammar (ECG) model, a computational formalism developed within the Neural Theory of Language paradigm, which stands on notions inherited from theories linking language to cognition and, ultimately, the latter to the bodily and experiential nature of the organism continuously interacting with the world. This model, applied to Italian data for the first time, possesses formal structures which allow the analyst to develop computational grammars adequate to investigate linguistic phenomena at different levels, examining both their formal and semantic properties, in order to capture the mechanisms whereby the interaction between the lexical units and the syntactic patterns instantiated in an utterance drives the understander to grasp the correct meaning of the message being conveyed. Overall, the application of the ECG approach to the Italian data analyzed in our case-study proves successful: the formalism can be easily extended to account for Italian data without requiring any substantial change, but just a few minor modifications. More generally, we can reasonably conclude that, due to its interdisciplinary roots which allow the model to include converging information from several different scientific disciplines, ECG proves suitable to exploit (and, possibly, elaborate) a number of notions inherited from the fields of Embodied Cognition and Cognitive Linguistics, endowing them with a rigorous computational dimension by means of its formalism. One of the main reasons of this model's effectiveness rests on its capability to satisfy two requirements which are often difficult to reconcile: flexibility and precision. Being extremely flexible, the formalism adopted in the analysis of our data can be adapted to a number of different situations, favoring the construction of computational grammars adequate to carry out the analysis of a several distinct phenomena in various different languages. At the same time, the level of precision of the ECG formalism allows us to develop rigorous grammars suitable to carry out consistent and plausible analysis of linguistic phenomena, also permitting to test their level of efficiency in real time.