Planning scope in spoken and written sentence production

This thesis investigates two questions about the cognitive mechanisms underlying the advance preparation of sentences. First, how much planning does the language system require to begin outputting a sentence and second, how is this scope determined. Previous research has concluded that advance planning embraces less than the sentence, is determined by either content or structure of some minimal linguistic unit, and is subject to variation (V. S. Ferreira & Slevc, 2007). Unlike previous research, the presented hypotheses were evaluated in both speech and writing. This eliminates explanations in terms of mechanisms that are modality specific, and therefore not fundamental to the language production system (see Alario, Costa, Ferreira, & Pickering, 2006). In two series of three experiments I elicited short sentences in speech and writing (keyboard typing). Under controlled conditions I manipulated (a) structural and lexical properties of elicited sentences (first series, Chapter 2) and (b) conceptual properties of the sentence's message (second series, Chapter 3). Hypotheses were evaluated by measurement of the time required to initiate output of the target sentence and of eye movements to referents of this sentence (arrays of simple line drawings) shown on the computer screen. These suggested two main conclusions: (1) Consistent with some previous research advance planning scopes over coordinated noun phrases (A and the B) while lexical content requires planning for the first noun but not beyond (Chapter 2), demonstrating for the first time that this effect replicates in writing. (2) Whether or not noun phrases are preplanned beyond the first noun is determined at a conceptual level, and not at a syntactic level (Chapter 3). These findings are in line with current models of language production (Bock & Ferreira, 2014; Konopka & Brown-Schmidt, 2014) and constitute a first step towards confirming the modality independence of these models