Learning person-specific knowledge about new people.

Abstract

Much progress has been made in the last twenty years or so on understanding the cognitive processes involved in familiar person recognition (see, for instance, Burton, Bruce, and Hancock, 1999). Indeed both Burton and colleagues (Burton et al., 1999) and Bredart and colleagues (Bredart, Valentine, Calder, Gassi, 1995) have proposed implemented models of several aspects of familiar person recognition. Yet models such as these presuppose the prior process of person familiarisation. A satisfactory theory of person processing needs to address not only familiar person processing but also unfamiliar person processing, and most critically, the familiarisation process which bridges the gap between these two categories. There is much evidence to suggest, moreover, that familiar and unfamiliar people are processed by the cognitive system in different ways. Yet studies investigating the gradual transition phase by which a person changes from being unfamiliar to being highly familiar are relatively rare. Some research has been carried out on the perceptual changes which accompany familiarisation (e.g. O’Donnell & Bruce, 2001). Perceptual change, however, is only one component of familiarisation. The other critical component is forming knowledge about the individual (i.e. semantic knowledge). There has hardly been any research on semantic familiarisation to date. The aim of this thesis is to obtain evidence pertaining to the issue of semantic familiarisation which should constrain future models of person processing. I review models of familiar person recognition in chapter one of the present thesis and further identify the gap existing in the literature with respect to semantic aspects of familiarisation. I also offer arguments as to the intrinsic importance of including familiarisation in any theory of person processing. The first experimental chapter investigates, by means of both verification and cued-recall tasks, whether a script-like representation of person-specific knowledge may prove more fruitful than current mechanisms of representing such knowledge in the course of learning about new people. I conclude that an explanation in terms of general memory theory best fits the results obtained in this chapter. The second experimental chapter moves away from the learning methodology used in first chapter to employ a much more ecologically valid learning paradigm. That is, as opposed to using one simple face image and a set of semantic properties I use an initially unfamiliar soap opera with which participants gradually become familiarised. I consider the structure of person-specific knowledge for both a set of famous people and further a set of characters from the soap opera. Results here suggest that occupations are the important category for famous persons but that, for a newly learned set of individuals, relational information is more important. I again examine these questions using both recall and verification tasks in this chapter. I argue that one critical factor in determining preferential representation of a given type of information in the representation of person-specific knowledge for some person is the co-occurrence frequency of encountering that person with that class of information. The final experimental chapter (Chapter 4) attempts to investigate the processes of semantic familiarisation in an on-line manner. I use a self-priming paradigm (see Calder & Young, 1996) to probe the development of participants’ semantic representations at regular intervals across a learning period. I report both within-domain and cross-domain experiments addressing this issue. Results indicate that after a relatively short interval newly learned persons behave comparably to famous persons in several important respects. Existing theories of general memory (e.g. Logan, 1988; McClelland & Chappell, 1998) account well for the perceptual familiarisation which occurs in the reported experiments. Further development is required, however, to account for the results on semantic familiarisation. I summarise and evaluate the experiments reported in the present thesis in the final chapter and also attempt some theoretical development which should provide useful for the creation of a unified theory of person processing.