Do local properties function as cues for musical key perception?1

Abstract: A global property (i.e., pitch set) of a melody appears to serve as a primary cue for key identification. Previous studies have proposed specific local properties in a melody (e.g., the augmented fourth, the perfect fifth, etc.) that may function as further cues. However, the role of the latter in key identification is controversial. The present study was designed to investigate what kinds of local properties, if any, function as reliable cues for key identification. Listeners were asked to identify keys for 450 melodies that consisted of the same pitch set, but which differed in sequential constraints. Using multiple discriminant analyses, we evaluated relative contributions of as many kinds of local properties as possible (e.g., single intervals, single pitch classes in each sequential position, etc.). The results showed that, except for the pitch class of the final tone, for which interpretation should be taken cautiously, none of the specific local properties examined contributed significantly to key identification. This finding suggests that, contrary to prior findings, key identification is derived from unidentified properties other than the specific local properties.     

The contribution of category‐based and contextual suppression towards retrieval‐induced forgetting1

Abstract: This study examines whether both context‐related and category‐related memory impairment can be found using a retrieval‐induced forgetting procedure. One group of participants was tested using a standard retrieval‐practice paradigm, whereas for the no‐practice group the retrieval practice was replaced by irrelevant filler activity. The results showed that the within‐participant baseline items were recalled more poorly by the practice group than by the no‐practice group, whereas the practice group showed the standard retrieval‐induced forgetting effect. Thus, retrieval practice impaired recall for both within‐category and baseline items, although it impaired the former to a greater extent. This finding suggests that it is necessary to consider not only categorical relationships between items, but also context information shared by items when exploring retrieval‐induced forgetting.