• No results found

3,3,1,1 Memory Architecture: Structural Dichotomy

3.3.2 Process Approaches

The intrinsic concept that underscores the majority o f process approaches is the transfer appropriate processing framework, discussed in Chapter 1. This view is also embodied by the proceduralist views that emphasise the role o f the operations and activities engaged at encoding and retrieval in order to define the memory acquired (Kolers, 1973; Kolers & Roediger, 1984), and by the encoding specificity hypothesis proposed by Tulving (1973). All o f these process approaches are predicated on the notion that memorial activity involves processes (Roediger, 1990b), procedures (Kolers, 1973; Kolers & Roediger, 1984), or records (Kirsner & Dunn, 1985) that can be accessed at a later time. However, these approaches, like the multiple memory systems framework, suffer from the problem o f being rather vague in terms o f the level o f theoretical specification, since the principles o f function, rather than the precise specification o f the causal sequences o f mechanisms and procedures responsible for task performance, have been formulated (Brent, 1984; Toth & Hunt,

1999).

The application o f process approaches to understanding age-related differences typically involves seeking differences in the types o f processing operations carried out on encoded events, in the extent to which context influences the specificity o f encoding, and in the ability to recapitulate original processing operations at the time o f retrieval. This overall approach is embodied within two dominant processing dichotomies: perceptual versus conceptual processing and self-initiated versus environmental support. Each o f these will be addressed in the following subsections. In addition, the distinction between activation and elaboration based processes and the components-of- processing view will be discussed.

conceptual processing, which originates from research conducted in the reading literature (McClelland & Rumelhart, 1981), and has been extended to memory processing (Blaxton, 1989; Jacoby, 1983b; Roediger, 1990b; Roediger et al., 1989b). There is also now evidence o f the functionally distinct neurological structures that subserve perceptual and conceptual memory processes (Blaxton et al., 1996; Gabrieli et al., 1994; Schacter, 1994a). In particular, cognitive neuroimaging has revealed that perceptual priming o f verbal materials is correlated with activity in the visual regions o f the occipital cortex and left frontal cortex, whereas the neural correlates o f conceptual priming appear to be located in the left frontal and temporal-parietal cortical regions (Blaxton et al., 1996; Gabrieli et al., 1996a; for a review, see Schacter & Buckner, 1998). Similarly, neurological dissociations have been obtained in adults diagnosed with Alzheim er’s disease, since evidence o f preserved perceptual priming but impaired conceptual priming has been reported ( for a review, see Fleischman & Gabrieli, 1998; Monti et al., 1996; Russo & Spinnler,

1994).

Despite the relative success o f the basic four-cell process-based framework that was described in Chapter 1, violations have been obtained that highlight two basic difficulties with the framework: (1) the lack o f a consistent taxonomy classifying memory tasks as perceptually or conceptually driven; and (2) the difficulty in measuring these processes without confounding the findings with differences in retrieval intentionality. For example, associative word stem completion is argued by some to require semantic elaboration at encoding, as would be expected if it were a conceptually driven task (Srinivas & Roediger, 1990), although this is certainly not universally accepted (Light et al., 1995; Reingold & Goshen-Gottstein, 1996a; Reingold & Goshen-Gottstein, 1996b). These issues will need clarification if progress is to be made. In this regard, the experimental work presented in the thesis that evaluated age-related differences in perceptual-associative word stem completion applied two variables, divided attention and item method directed forgetting, in order to provide additional data concerning the functional properties associated with this form o f memory for new associations.

Additional assumptions may be necessary to increase the inductive pow er o f the perceptual- conceptual processing distinction. For example, it has been argued that perceptual priming is supported by dissociable form-specific and modality-specific processes (Gabrieli, Fleischman, Keane, Reminger, & Morrell, 1995; Tulving & Schacter, 1990), whereas multiple, dissociable forms o f conceptual unconscious processes may mediate performance on conceptual indirect

3. Conscious and Unconscious M emory Processes and Ageing

1996). For example, Vaidya et al. (1997) proposed a largely descriptive model o f two dissociable routes to the retrieval o f conceptual information dependent on whether or not the conceptual retrieval cue instantiates competitive and non-competitive memorial access. Competitive access refers to the relatively unconstrained retrieval processes instantiated when alternatives compete in response to a retrieval cue, whereas non-competitive access refers to the retrieval o f a single semantic entry because the relevant item is directly activated. Further, it was argued that only those conceptual memory tasks that instantiate competitive access are sensitive to conceptual elaboration and divided attention. Accordingly, given that category exemplar generation fosters the competitive retrieval o f items, an effect o f divided attention would be predicted by this model. The experiment reported in Chapter 6 directly addressed this issue.

For the present purposes, the relation between the perceptual-conceptual distinction and conscious- unconscious processes is also o f particular relevance. As discussed in Chapters 1 and 2, the relation between the ‘content’ o f mnemonic processes and conscious and unconscious processes has been characterised as orthogonal (Jacoby, 1996; Jacoby et al., 1997b). However, the attribution o f a greater role for conceptual processes in direct memory tests relative to indirect mem ory tests, independent o f the informational demands o f the test cues (Craik et al., 1994; Roediger, Weldon, Stadler, & Riegler, 1992; Weldon, Roediger, Beitel, & Johnston, 1995), is not able to account for dissociations between conceptual direct and indirect memory tests in any simple manner. By extension, although a role for perceptual processing in indirect m em ory tests has been established, the specification o f how indirect memory tests utilise conceptually based information is less clear (Challis & Brodbeck, 1992).

In addition to the availability o f consciousness, a criterial feature separating the nature o f the information that supports direct and indirect memory tests is the availability o f the spatiotemporal context in which an item occurred (Bower, 1996; Jacoby & Hollingshead, 1990). This condition is able to account for the effects o f divided attention (Mulligan, 1998). Specifically, there is substantial evidence to suggest that divided attention disrupts conceptual and contextual encoding (Cowan, 1995; Craik, 1989; Rabinowitz et al., 1982), rather than perceptual processes. Therefore, an effect o f divided attention on perceptual and conceptual direct m em ory tests would be predicted, because these tests are supported by contextual encoding. Similarly, conceptual direct and indirect memory tests would be expected to be impaired by divided attention because these tests are mediated by conceptual processing (cf. with the attentional view). Nonetheless, the preponderance

processing distinction (Roediger, 1990a), demonstrates the importance o f determining the nature o f perceptually- and conceptually-driven processing, since the utility o f the distinction will be, in part, dependent on understanding the context-dependent nature o f these processes (cf. Craik, 1991; Kolers, 1975).

Specific predictions emerge from the converging operations provided by the perceptual and conceptual processing dichotomy in relation to ageing. For example, conceptual rather than perceptual processing is expected to be impaired in old age (Rabbitt, 1979; Rabbitt, 1982; Rybash, 1996, see also Chapter 1). Understanding the dissociations in performance in older adults as a function o f direct and indirect memory tests will reflect, at a minimum, the interaction between the perceptual-conceptual and the conscious-unconscious processing dichotomies, rather than that between conscious and unconscious retrieval alone. As will become apparent shortly, it is in the context o f the environmental support hypothesis that the conceptual-perceptual processing distinction has been most clearly applied to ageing and memory.