Academic journal article Memory & Cognition

Is Awareness Necessary for True Inference?

Academic journal article Memory & Cognition

Is Awareness Necessary for True Inference?

Article excerpt

In transitive inference, participants learn a set of context-dependent discriminations that can be organized into a hierarchy that supports inference. Several studies show that inference occurs with or without task awareness. However, some studies assert that without awareness, performance is attributable to pseudoinference. By this account, inference-like performance is achieved by differential stimulus weighting according to the stimuli's proximity to the end items of the hierarchy. We implement an inference task that cannot be based on differential stimulus weighting. The design itself rules out pseudoinference strategies. Success on the task without evidence of deliberative strategies would therefore suggest that true inference can be achieved implicitly. We found that accurate performance on the inference task was not dependent on explicit awareness. The finding is consistent with a growing body of evidence that indicates that forms of learning and memory supporting inference and flexibility do not necessarily depend on task awareness.

The transitive inference (TI) task requires one to learn context-dependent relations and to make inferences about untrained relations. In a typical TI task, participants are trained on a set of overlapping two-item discriminations or premise pairs (e.g., A > B, B > C, C > D, D > E; "A > B" denotes a forced choice between A and B in which A is the correct choice). An efficient representation of these simultaneous context-dependent relations is A > B > C > D > E. At test and without reinforcement, the capacity for inference is tested with an untrained pair, B?D. The capacity for correct inference has been observed in numerous nonhuman animal species (Davis, 1992; Gillan, 1981; Weaver, Steirn, & Zentall, 1997) and in humans (Greene, Spellman, Dusek, Eichenbaum, & Levy, 2001; Heckers, Zalesak, Weiss, Ditman, & Titone, 2004; Nagode & Pardo, 2002; Preston, Shrager, Dudukovic, & Gabrieli, 2004).

In humans, inference is possible with or without explicit awareness of the contingencies (Frank, O'Reilly, & Curran, 2006; Greene, 2007; Greene, Gross, Elsinger, & Rao, 2006; Greene et al., 2001; van Elzakker, O'Reilly, & Rudy, 2003). If participants are not informed of a hierarchy at the outset, TI task awareness either is independent of task performance (Greene et al., 2006; Greene et al., 2001) or may occur in the absence of task awareness (Frank, Rudy, Levy, & O'Reilly, 2005). That learned contingencies may support inference without awareness is of particular interest because it suggests a level of flexibility previously believed to require conscious memory processes (Clark, Manns, & Squire, 2002; Reber, Knowlton, & Squire, 1996; Smith & Squire, 2005).

One possibility is that implicit versions of TI are not really demonstrating flexibility, but instead are solved using pseudoinference (Frank, Rudy, & O'Reilly, 2003; Siemann & Delius, 1994; van Elzakker et al., 2003; Zentall & Sherburne, 1994). Consider a five-item TI task (e.g., A > B > C > D > E). If the A item is present, it is always the correct choice; likewise, E is always incorrect. The middle items B-D are each correct only 50% of the time. Van Elzakker et al. argued that end items cause a blocking effect that is responsible for differential associative strength among items. Accordingly, the perfect predictive value of the end items causes them to be treated as atomic discriminations (select A, avoid E), and so blocking leads one to ignore B in the AB pairing and to ignore D in the DE pairing. This has the consequence that B is ignored when it would be an incorrect choice (A > B) and is attended to only when it is the correct choice (B > C). Similarly, D is ignored when it would be the correct choice (D > E) and is attended to only when it is the incorrect choice (C > D). According to this view, the novel pairing at test, B? …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.