The project investigated the relationship between the semantics of attitude verbs –e.g., believe and want– and the syntactic types of embedded clauses –e.g., declarative, interrogative– they can combine with. The hypothesis pursued in the project is that such combinatorial restrictions are not the result of accidental syntactic specifications but, instead, emerge from the semantic properties of the attitude predicate. Our results substantially advance the state-of-the-art on the topic on three fronts. Empirically, we collected data across 18 languages testing semantic and combinatorial properties, resulting in a publicly available cross-linguistic database on which to test hypothesized correlations. Second, we contributed novel theoretical approaches deriving several cross-linguistically attested correlations between the semantics of attitude verbs and the clause types and subtypes they embed as well as concrete case studies. Third and finally, we addressed existing methodological limitations by defining a more articulated and fieldwork-oriented version of the Focus-sensitivity test and by experimentally testing debated empirical patterns in the literature.