Finding structure in a maze of words: Priming as a guide through the mental lexicon

It is a well-known finding that the presentation of a related word (e.g., cat) enhances processing of a subsequently presented target (e.g., dog) compared to when the preceding word is unrelated (e.g., car). This phenomenon, called semantic priming, has been studied extensively because it is thought to provide insight into the principles by which the mental lexicon is organized. However, several crucial issues remain unresolved and/or hotly debated. In this dissertation, I look for answers and attempt to find structure in the maze of words. A first research line investigates semantic priming with the speeded word fragment completion task. In this task, participants need to complete words from which one letter was omitted, as fast as possible (e.g., lett_ce; lettuce). This paradigm has some interesting qualities in comparison to the traditionally used lexical decision task. Three experiments, described in Chapter 2, showed that this task can capture semantic priming, especially for highly frequent words like money and warm, whereas the lexical decision task failed to find priming for the latter targets. To further examine this promising paradigm, a megastudy was conducted (see Chapter 3). The idea was to gather norms, which are useful for follow-up studies, and to gain more insights into the processes involved in this task. A second research line focusses on the (controversial) notion that semantic priming is at least in part produced by an automatic process such as spreading activation. More specifically, Chapter 4 examines the effect of imposing a working memory load. The results suggested that prospective processes require cognitive resources, whereas retrospective processes are relatively effortless. In Chapter 5, I test whether priming processes are unaffected by intentions (i.e., a second criterion for automaticity). Concretely, I created a context wherein prime processing is detrimental rather than beneficial. Nevertheless, the results showed an inhibitory priming effect, which is consistent with an automatic priming account like spreading activation. A final study, described in Chapter 6, evaluates how priming manifests itself in different tasks (i.e., lexical decision and pronunciation). Task differences and similarities were assessed by correlating item-level priming effects obtained in each task. The correlations were larger when the stimulus onset asynchrony between prime and target was short, which provides evidence for a stable, rapidly operating mechanism such as spreading activation.status: publishe