Minor mode cuing: Do composers signal minor mode sooner than major mode?

Abstract

Western-enculturated listeners use different tonal schemas when listening to versus minor mode (e.g., C. L. Krumhansl and E. J. Kessler, 1982, Tracing the dynamic changes in perceived tonal organization in a spatial representation of musical keys, Psychological Review, Vol. 89, pp. 334-368). To avoid possible tonal ambiguity, one might expect composers to attempt to disambiguate or clarify the mode soon after a work begins. The mode is the most common tonal schema in Western music, so listeners might expect that an unknown work will be in the mode. This suggests that composers might tend to cue listeners sooner when presenting works in the (rarer) minor mode. In this article, we report on 2 studies that test the hypothesis that mode-defining pitches occur earlier for minor-mode works. Surprisingly, the studies do not support this hypothesis: in fact, mode-defining scale tones (mediant and submediant) appear significantly later in notated music than in scrambled versions of the same scores, regardless of the mode. Repercussions of the results are discussed.Keywords: minor mode cuing, tonality, key induction, Humdrum Toolkit, expectationSince about the 17th century, Western music has been dominated by the major/minor system in which two tonal schemas predominate. Throughout the history of the major/minor system, the mode has been more common than the minor mode. Depending on the musical genre, between two thirds and three-quarters of works are written in the mode. For example, in the Barlow and Morgenstern (1948) Dictionary of Musical Themes, 73% of nearly 10,000 thematic passages are designated by the authors as in the mode with the remaining 27% in the minor mode.Huron (2006) reviewed a number of empirical observations suggesting that listeners tend to assume that a newly encountered musical passage will be in the mode. That is, the mode appears to be a listening mode for Western-enculturated listeners. Given the predominance of the mode, this perceptual default is consistent with statistical learning, and more specifically consistent with Bayesian inference. If listeners do indeed tend to assume a major mode default, then the switch to minor-mode listening is probably cued by one or more tonal patterns that fail to conform to the mode. Such disconfirming stimuli are widely regarded as important events for activating appropriate perceptual schemata (Holland, Holyoak, Nisbett, & Thagard, 1986).How listeners infer the tonality of a passage has attracted considerable research attention over recent decades (Brown, 1988; Brown & Butler, 1981; Cohen, 1977, 1991; Cuddy, Cohen, & Mewhort, 1981; Cuddy & Lyons, 1981; Hebert, Peretz, & Gagnon, 1995; Huron & Parncutt, 1993; Krumhansl, 1990a; Krumhansl & Kessler, 1982; Leman, 1995; Smith & Schmuckler, 2004). For monophonic passages, tonality might be inferred through the recognition of mode-defining chroma distributions, through distinctive interval patterns, via chroma-dyad distributions, or through some combination of these or other pitch-related mental representations. Successful key-inference algorithms have been devised using a variety of techniques (Aarden, 2003; Bellman, 2005; Holtzman, 1977; Krumhansl & Schmuckler, 1986; Sapp, 2008; Schmuckler & Tomovski, 1997, 2000; Takeuchi, 1994; Temperley, 2007). However, reliable key inference by machine does not necessarily mean that listeners use an equivalent mental process. The Krumhansl-Butler debate of the 1980s and 1990s (Butler, 1989, 1990; Cross, 1997; Krumhansl, 1990b, 2000; Van Egmond & Butler, 1997) can be interpreted in several ways. However, one interpretation is that the debate arose from disagreement about whether the mental representation for tonal sequences is predominantly pitch related (Krumhansl) or predominantly interval related (Butler). See Huron (2006) for an extended discussion. …