The capacity to ascribe goals and intentions to others is a fundamental step in child cognitive development. The aim of the present study was to assess the age at which these capabilities are acquired in typically developing children. Two experiments were carried out. In the first experiment, 4 groups of children (age range = 3 years 2 months-7 years 11 months) were shown pictures representing hand-object interactions and asked "what" the individual was doing ("what task") and "why" ("why task"). In the "why task," observed handgrip could be either congruent with the most typical action performed with that object (e.g., to drink in the case of a mug) or corresponding to the act of putting away the object. In the second experiment, children saw pictures showing a handgrip either within a context suggesting the most typical use of the object or its being put away. Results showed that by 3-4 years, children are able to state the goal relatedness of an observed motor act ("what" understanding), whereas the ability to report the intention underlying it ("why" understanding) is a later and gradual acquisition, reaching a high performance by 6-7 years. These results, besides their intrinsic value, provide an important baseline for comparisons with studies on developmental disorders, also highlighting the relevance of distinguishing "what" and "why" understanding.