Recent auditory research using sequentially presented, spatially fixed tones has found evidence that, as in vision for simultaneous, spatially distributed objects, attention appears to be important for the integration of perceptual features that enable the identification of auditory events. The present investigation extended these findings to arrays of simultaneously presented, spatially distributed musical tones. In the primary tasks, listeners were required to search for specific cued conjunctions of values for the features of pitch and instrument timbre. In secondary tasks, listeners were required to search for a single cued value of either the pitch or the timbre feature. In the primary tasks, listeners made frequent errors in reporting the presence or absence of target conjunctions, Probability modeling, derived from the visual search literature, revealed that the error rates in the primary tasks reflected the relatively infrequent failure to correctly identify pitch or timbre features, plus the far more frequent illusOIY conjunction of separately presented pitch and timbre features. Estimates of illusory conjunction rate ranged from 23% to 40%. Thus, a process must exist in audition that integrates separately registered features. The implications of the results for the processing of isolated auditory features, as well as auditory events defined by conjunctions of features, are discussed.

Recent speech research has demonstrated that the presence of prototypes may be reflected in the internal structure of speech categories. Some speech results have found lower discrimination in a vowel P context compared to a NP context, suggesting that a prototype may function as a perceptual magnet, effectively decreasing perceptual distance, and thus, discriminability, between stimuli. The current study examines the function of prototypes in a musical category; another natural, but nonspeech, category. Prototype (P) and nonprototype (NP) sets of major triad stimuli were constructed, with stimuli in the P set being more representative of the category than the NP stimuli. Musically experienced subjects rated the stimuli in each set for goodness as a major triad, with the highest rated stimulus serving as a prototype standard for a subsequent discrimination task. Results from the discrimination task demonstrated better performance in the prototype context. The current nonspeech results appear to follow predictions based on classification and perceptual models (to be discussed), and provide a natural, nonspeech contrast to speech findings. The results will also be addressed in terms of music perception, including aspects such as different temperment systems and the differential importance of chord components.

What role does attention play in the perception of components within a melody? A paradigm used by Palmer and Holleran [Percept. Psychophys. 56(3), 301–312] provides a basis for addressing this question. Their study investigated different influences on the perception of pitch alterations in three-voiced musical passages, where subjects first learned a four-measure standard, then completed a same/different task. ‘‘Different’’ passages were constructed by altering the original standard, with a harmonically related (HR) or unrelated (HU) frequency change made in the lowest, middle, or upper voices of the musical passage. Results indicated that HR changes were less detectable than HU changes and were the least detectable in the middle voice. The current study uses the same procedures and musical materials as the Palmer and Holleran work, but provides the subjects with cues indicating which frequency region might contain the alteration (low, medium, or high). Results will be discussed in terms of the role directed attention plays as a function of harmonic relationship and frequency region.

In the present study, an accuracy, rather than a reaction time, version of the Garner paradigm was used to evaluate the integrality or separability of major chord components. Tuned (prototype, or P) and mistuned (nonprototype, or NP) sets of root position C-major triads were constructed by holding the C constant in all stimuli and varying the E and G frequencies in 2-and 4-Hzsteps. The P stimuli represent small systematic mistunings in the E and G notes relative to an equal-tempered C-major chord. The NP stimuli represent an equivalent range of frequency variation, but relative to a significantly out-of-tune C-major triad. In different experimental sessions, a same-different (AX) task was used to separately evaluate discrimination performance for the E and G frequencies as a function of whether the nontarget frequency (G or E) was fixed or varied in either a correlated or an orthogonal fashion (with the C frequency always held constant). Compared with a fixed baseline condition where only the target frequency changed, both chord components exhibited a significant redundancy gain in the correlated conditions and, to varying degrees, significant interference effects in the orthogonal condition, indicating that the chord components function largely in an integral fashion. Relative to the discrimination of G, discrimination of the E frequency was less influenced by variation in the nontarget (G) frequency, showing that attention, to some degree, could be selectively allocated to the E chord component. In addition, the results were consistent with previous findings that the functional prototype for the major chord category seems to act as a perceptual anchor, rather than as a magnet, and appears to be located in the physiologically defined area of just temperament, as opposed to the more experientially defined area of equal temperament.Historically, classes of speech and, to a lesser extent, music stimuli were often investigated from the perspective of absolute conceptualizations of categorical perception.' With the central assumption that perceptual categories were discrete (with only irrelevant variation within categories), studies tended to focus on the relationship between category boundary locations and heightened discriminability across the category boundary, with discriminability both expected and found to be reduced within categories. This early work is consistent with discrete categorization processes based on feature detectors, where perceptual differentiation is defined primarily in terms of the limits of feature activation, and thus, the category boundary. In contrast to early discrete feature-detector mechanisms, more recent concepts of categorization have been based on the use of prototypes or exemplars located within categories, thus shifting the focus of importance from the category boundary to the nature of perception within categories. Early research based directly or indirectly on discrete concepts of categorical perception could tell us very little about the nature of internal category structure...

Recent speech research has begun to evaluate the internal structure of categories. In one such study, Kuhl (1991) found that discrimination was poorer for vowel stimuli that were more representative of the category (prototype, or P, set) than it was for less representative stimuli (nonprototype, or Np, set). This finding was interpreted as indicating that a category prototype may function as a "perceptual magnet," effectively decreasing perceptual distance, and thus discriminability, between stimuli. The present study examines the function of prototypes in a musical category-another natural, but nonspeech category. Paralleling the Kuhl study, representative (P) and less representative (NP) sets of major triad stimuli were constructed, based on equal temperament. Musically experienced subjects rated the stimuli in each set for goodness as a major triad, with the highest rated stimulus serving as a prototype standard for a subsequent discrimination task. Results from the discrimination task demonstrated better performance in the P context than in the NP context. The current nonspeech results indicate that a prototype functions as an anchor rather than a magnet. In addition to providing a natural, nonspeech standard for comparison with speech findings, the results provide some important insights into the nature of musical categories.Although it has long been conjectured that speech categorization may be based on the use of prototypes or exemplars, most speech perception research has tended to focus on the location of category boundaries. This focus on category boundaries is probably a carryover from the notions of categorical perception which posited absolute recoding of perception in terms of discrete phonetic categories, and which dismissed any within-category perceptual variation as being due to stimulus artifacts (StuddertKennedy, Liberman, Harris, & Cooper, 1970). As a result, little attention has been given to differences in perceptual quality or goodness, which, in theory, should exist within categories that are based on the use of prototypes or exemplars. In contrast to this long tradition ofcategorization studies based on labeling tasks, some recent research has begun to examine the internal structure of speech categories (e.g., Kuhl, 1991;Li & Pastore, 1992;Samuel, 1982; Volatis & Miller, 1992). In general, these studies have found category membership to be qualitatively graded, with Kuhl (1991) providing evidence that quality ofmembership in a vowel category is reflected in a distinct pattern of discriminability between stimuli. Patterns of qualitative grading and related patterns of discrimination This research was supported by Grants F496209310033 and F490693I 0327 from the Air Force Office of Scientific Research to the second author. The opinions expressed are those of the authors and do not necessarily represent those of the granting agency. We would like to thank Cynthia Connine for several helpful suggestions concerning the discrimination models. Correspondence should be addressed to B. Acker, Psyc...