Tuesday March 17, 09:45-10:45

Item-specific and Relational Information in Memory for Actions

Johannes Engelkamp

(Department of Psychology, University of the Saarland, Germany)

It is a widely shared assumption that item-specific and relational information should be distinguished and can be differentially used in explicit memory tests. More specifically, it is assumed that free recall depends on item-specific and relational information, whereas recognition mainly relies on item-specific information. Finally, organizational scores (such as ARC) reflect relational information which is usually manipulated by list structure.

I will discuss the role of item-specific and relational information in memory for action events. When action phrases are to be remembered, it turns out that there is an enactment effect. Action phrases which are performed by the subjects (SPTs) are better retained than phrases which are only listened to (VTs). Self-performing actions (SPTs) usually also lead to better recall than observing the experimenter performing them (EPTs).

In more detail I will discuss two questions.

The first question is whether the enactment is due to item-specific information, to relational information or to both types of information.

The second question is whether there is only one type of relational information.

It will turn out that the enactment effect is due to the excellent item-specific information provided by self-performing actions. Moreover, it will turn out that categorial-relational information as it is provided by list structure is used equally after VTs and SPTs. However, categorial relational information is only one type of relational information. Other types of relational information must be considered as well.

Categorial relational information (as provided by list structure) is rooted in preexperimental knowledge. Besides this type of preexperimental relational information, at least one other type must be considered. I call it episodic relational information. It refers to associations which are newly formed during the study episode. Both kinds of relational information can exist between action phrases and within action phrases.

I will present experimental evidence that, on the one hand, preexperimental relational information is equally used in VTs and SPTs, whereas, on the other hand, episodic relational information is strongly dependent on the type of encoding (VTs, EPTs, SPTs). Moreover, episodic relational encoding is dependent on whether new associations had to be formed intra-phrases or inter-phrases. Inter-action encoding is particularly difficult in SPTs, whereas episodic relational intra-action encoding is particularly efficient in SPTs.

It will be concluded that one must distinguish whether relational encoding refers to preexperimental or new associations and whether the associations refer to inter-action or intra-action relations.