The research we present in this book establishes a unique line of research within the Processing Instruction model by assessing the transfer-of-training effects of this approach to grammar instruction on the way learners make form–meaning connections. In this work we will refer to direct or primary effects as well as to transfer-of-training effects. Each investigation of the effects of Processing Instruction has isolated and targeted for treatment a particular linguistic feature. Learners' knowledge of the target linguistic feature is assessed prior to treatment (pre-testing) and then again after treatment (post-testing). Their increased knowledge of the target linguistic item, resulting from the treatment, is what we refer to as a direct or primary effect. If the treatment also resulted in increased knowledge of some other linguistic item in addition to the target linguistic item, then we would have a transfer-of-training effect. We further categorise the transfer-of-training effects as being secondary effects or cumulative effects. We refer to an effect as secondary if the two linguistic items under focus present L2 learners with a similar processing problem. Word final morphology such as present tense -s and past tense -ed in English present L2 learners with similar processing difficulties. If Processing Instruction on past tense -ed also affected present tense -s, we would call it a secondary effect. Word final morphology such as French imparfait (past imperfective) and word order permutations such as French faire causative constructions present L2 learners with very different processing problems. Processing instruction on these two linguistic items attempts to instil very different processing strategies. Yet, if Processing Instruction on the French imparfait also affected French faire causative constructions, we would call it a cumulative effect.

As detailed in our review of literature in Chapter 2, research on Processing Instruction (PI) has so far focused on measuring only direct or primary effects on learning a specific/targeted linguistic feature. Typically, the research has compared Processing Instruction with traditional output-oriented instruction and/or meaning-based output instruction.