Memory for Environmental Sounds: an Investigation Using the Levels of Processing Framework of Memory

Abstract

The processes involved in memory for auditory stimuli (speech, music, voices and environmental sounds) have not been well investigated. Moreover, few of the experiments on environmental sound processing have been theoretically motivated. As a consequence, little is known about the way memory for sounds is affected by variables that have frequently been studied in relation to visually presented stimuli or spoken words. Normative data was collected on variables such as familiarity, imageability and nameability for a set of 140 sounds. A series of experiments was conducted using the levels of processing paradigm to investigate memory for environmental sounds. Experiment 1 used a free recall test and demonstrated a levels of processing effect, with higher recall of sounds by subjects who performed a 'higher-level' semantic categorisation task than by subjects who performed a 'lower-level' pitch-rating task at study. Experiments 2 and 3 investigated levels of processing theory using recognition tests. In addition, these experiments examined the role of semantic information in memory for sounds, by presenting sounds or sounds names at test. Levels of processing effects were observed in both experiments and sounds were better recognised than sound names. An interaction between orienting task and test condition revealed that recognition of sound names was better following a semantic encoding task than an acoustic encoding task, but sounds were equally well recognised by subjects in both encoding conditions. Recognition rates in Experiment 2 were highest for target sounds paired with semantically similar distractors, followed by unrelated then acoustically similar distractors. Experiments 4 and 5 investigated recognition memory for acoustic information using the levels of processing paradigm, by presenting verbalisable and 'non-verbalisable' sounds (i.e. sounds with no meaning). No effect of orienting task was observed for either sound type, and non-verbalisable sounds were less well recognised than verbalisable sounds in both experiments. Targets paired with acoustically dissimilar distractor sounds were better recognised than targets paired with acoustically similar distractors. Experiment 6 tested recognition memory for sounds and sound names after a 24 hour delay. A levels of processing effect was observed, with significantly higher recognition by subjects in the semantic condition than by subjects in the acoustic condition in both sound, and sound-name test conditions. Recognition rates for sounds were significantly higher than the recognition rates for sound names. Collectively, the data provide evidence to support the argument that encoding semantic and acoustic information has additive effects on memory for sounds. The additivity of these codes suggests that semantic and acoustic elements may be processed independently in memory. Implications of these results are discussed in relation to environmental sound processing as well as auditory processing in general.