FIGURE 3.

Between-context analysis (1): behavioral results for three groups of participants. Speech onset latency-priming effects (novel – repeated syntax) in ms (error bars reflect standard errors), split by Speaker Switch (no speaker switch means production prime – production target; speaker switch means comprehension prime – production target) and Target Structure (active or passive voice). Dark-gray bars with solid fill represent the average priming effect of the MRI participants in the non-communicative condition. Of the two bars with striped pattern fill, the darkest gray bar represents the MRI participants in the communicative context and the lighter gray bar represents behavioral participants in the communicative condition. There were speech onset latency-priming effects in the two communicative as well as in the non-communicative condition. The groups differed from each other in the Speaker Switch – Passive target condition, in that only the behavioral participants in the communicative context showed a significant priming effect here. There was no overall interaction Syntactic Repetition ^* Group: it is not the case that participants in the communicative context show stronger syntactic priming effects than participants in the non-communicative context.