The present study was designed to explore serial position and suffix effects in the short-term retention of nonverbal sounds. In contrast with previous studies of these effects, a probe recognition paradigm was used to minimize the possibility that participants would use a verbal labelling strategy. On each trial, participants heard a memory set consisting of three pure tones, followed five seconds later by a probe tone. Participants were required to indicate whether or not the probe tone had been a member of the memory set. On most trials, a suffix sound was presented 1 second following the third sound in the memory set. Results revealed that tones presented in the first and last positions of the memory set were recognized more accurately than were tones presented in the middle position. Furthermore, recognition of sounds presented in the last position was compromised when the memory set was followed by a postlist suffix of similar pitch, spectral composition, and spatial location.
Empirical investigations of the nature of short-term memory have typically required participants to remember items that could be easily verbally labelled, such as letters (Crowder, 1968; Jolicoeur & Dell-Acqua, 1998; Murray, 1968), digits (e.g., Moray & Jordan, 1966; Salame & Baddeley, 1982; Wickelgren, 1965), words (e.g., Copeland & Radvansky, 2001; Loess, 1967), and pictures of common objects (e.g., Coltheart, 1999; Yuille & Fox, 1973). However, as Greene (1988) has argued, such an approach to developing an understanding of short-term memory may be limited because memory for such stimuli "could reflect more general aspects of memory, and be influenced by mechanisms specific to speech processing" (p. 247). For this reason, explorations of short-term memory using stimuli that are difficult or impossible to label verbally are important because they may provide a more thorough understanding of the nature of short-term memory. Empirical investigation of short-term memoiy for nonverbal visual material has generally shown better memory for the first few items in a memory set (this is typically referred to as a primacy effect) for stimuli such as snowflake pictures (Neath & Knoedler, 1994), and kaleidoscope patterns (Wright, 1994; Wright et al., 1990). Because these types of stimuli cannot be easily labelled verbally, the evidence of primacy obtained for them suggests that their retention is not dependent on verbal labelling. This evidence stands in sharp contrast to the suggestions of some theorists that primacy effects are necessarily dependent on verbal processing strategies (e.g., Rundus & Atkinson, 1970).
A variety of studies have been directed toward determining whether factors known to affect retention of verbal material hold also for short-term memory of nonverbal sounds. Thus, for example, it appears clear that memory for pitch decays steadily over time (e.g., Egan, Schulman, & Greenberg, 1961; Eriksen & Johnson, 1964; Harris, 1952; Massaro, 1972b; Moss, Myers, & Filmore, 1970; Pollack, 1964; but see Cowan, Saults, & Nugent, 2001, for a desccnting view). The possibility that serial position effects may arise in the short-term retention of nonverbal sounds has been the subject of a number of studies. In several of these experiments, a serial recall procedure has been used wherein listeners were required to list the sounds that they had heard in the order in which they were presented (e.g., Foreit, 1976; Rowe, 1974; Surprenant, Pitt, & Crowder, 1993). The serial position effects reported in these studies have varied somewhat with some investigators reporting only primacy (Foreit; Surprenant et al., 1993) and others reporting both primacy and recency (Greene & Samuel, 1986; Roberts, 1986; Rowe; Rowe & Rowe; 1976). Because written serial recall was required in all of these studies, however, it is possible that the serial position effects that were obtained may have been driven by retention of the verbal labels or written symbols that were to be recalled rather than by memory for the acoustic characteristics of the sounds themselves. …