Irrespective of the educational level, grammar is typically tested in a very traditional manner by means of such tasks as multiple choice, gap filling, paraphrasing or translation. The problem with this manner of evaluation is that it mainly taps into learners’ explicit, declarative knowledge that can only be applied when sufficient time is available. In addition, these tasks primarily focus on form, simultaneously ignoring the semantic and pragmatic aspects of grammatical knowledge. The article argues that such traditional tests should be complemented by what is referred to as productive and receptive focused communication tasks (Ellis, 2003), which necessitate the use of a specific grammar structure or at least are designed in such a way that such use is conducive to the attainment of the communicative goal. The application of such tasks allows insights into the implicit, or at least highly automatized, knowledge of the targeted structures (Ellis, 2009; DeKeyser, 2010) as well as shedding light on their form, meaning and use (Larsen-Freeman, 2003). It is suggested that their use not only enhances the validity of testing grammar but can also have a beneficial effect on the ways in which grammar is taught and learned.