It’s only rock ‘n roll (but I like it): Chord perception and rock’s liberal harmonic palette

Abstract

Both music-theoretic accounts and corpus analyses indicate that rock routinely employs chords that deviate from the norms of common-practice music. Yet we know little about how listeners experience the chords that make up rock’s liberal harmonic palette. In the present study, participants in two online experiments rated major chords that followed a short tonal sequence (a major scale + tonic major triad). Liking ratings obtained in Experiment 1 replicated earlier work showing that listeners prefer rock-typical targets—chords that are common in rock, but lie outside the basic diatonic set—to atypical, rarely used targets (Craton, Juergens, Michalak, & Poirier, 2016). Goodness of fit ratings with the same stimuli in Experiment 2 were similar to the liking ratings. High ratings for rock-typical target chords across the two experiments were not an artifact of their register and the pattern of responses was similar across four levels level of music training. In addition, the mean fitness ratings were approximated by simulations conducted with an auditory short-term memory (ASTM) model (Leman, 2000). Considered together, the findings provide evidence that listeners perceive rock-typical chords to be normal components of a key and that they may do so based on information in the auditory signal alone, without recourse to statistical learning mechanisms or representations of tonal knowledge. We speculate that bottom-up processes operating directly on the auditory signal create a perceptual ranking of chord fitness, which provides the harmonic palette from which composers/improvisers in different musical systems may conservatively (common-practice) or liberally (rock) choose. The mean data are available in tabular form for modelers in the online supplemental material.

Cognitive and sensory expectations independently shape musical expectancy and pleasure

“…Although this provides a high degree of ecological validity, it implies that only a subset of possible chord progression combinations has been presented, and the extent to which a chord is surprising or pleasant is only relative to other chords found in this musical style. Previous studies have shown that the same chord in a progression could evoke different expectancy and preference ratings depending on whether the stimuli were composed in the style of common-practice or rock music [ 45 , 93 ]. These results highlight the role of stylistic context in shaping perception and consequent emotional response during music listening, and therefore suggest the need to demonstrate that the current findings generalize to other musical styles.…”

“…As underlying hypotheses and assumptions are formalized and made explicit, the comparison of simulations from computational models with behavioural data provides a direct assessment of the plausibility of the biological or cognitive mechanisms embodied by each model [ 10 , 44 ]. Such an approach has been used to show that sensory expectations as simulated by an auditory short-term memory model could explain many of the priming effects previously taken to support cognitive accounts—even when sensory influences have been accounted for [ 40 , 45 ]. Other studies paint a more nuanced picture, with the comparison of simulations from several computational models against behavioural data generally supporting the influence of cognitive over sensory mechanisms [ 34 , 43 , 46 – 49 ].…”

Cognitive and sensory expectations independently shape musical expectancy and pleasure

“…In support of this, subjective expectations for tones and chords are modelled well by stylistic probabilities given the preceding context (Harrison & Pearce, 2018;Morgan et al, 2019;Pearce, Ruiz, Kapasi, Wiggins, & Bhattacharya, 2010;Sears, Pearce, Spitzer, Caplin, & McAdams, 2019), and listeners find chords that violate syntactic rules of Western tonal harmony unexpected (Koelsch et al, 2019). In contrast, sensory accounts argue that expectancy emerges as an epiphenomenon of how sounds are processed and then perceived Distinct roles of cognitive and sensory information in musical expectancy 4 on the short-term sensory-acoustic level (Bigand et al, 2014(Bigand et al, , 2006Craton, Lee, & Krahe, 2019;Parncutt, 1989). Many findings from existing behavioural and neurophysiological studies on harmonic expectations can be replicated using models that simulate the decay of auditory signals held in short-term memory (Bigand et al, 2014;Leman, 2000), or from the spectral similarity of pitches (Milne & Holland, 2016;Milne, Sethares, Laney, & Sharp, 2011) without resorting to syntactic principles or long-term stylistic probabilities.…”

“…This gives a measure of the similarity of auditory patterns undergoing short-term memory decay based solely on the harmonic and subharmonic features of the auditory input (Leman, 2000). The separate effects of these two models on surprise ratings and the improved predictive performance when incorporating predictions from both IDyOM and PP imply unique contributions of cognitive and sensory information, and speak directly against the proposal that abstract representations of music structure are not necessary for musical expectancy (Bigand et al, 2014;Craton et al, 2019).…”

Probe Tone Paradigm Reveals Less Differentiated Tonal Hierarchy in Rock Music

“…Differentiation of the perception of tonal hierarchy in classical and rock contexts has more recently been investigated in behavioural experiments. Craton et al (2016Craton et al ( , 2019 have reported that participants show preferences for chords found to be common in the rock corpora, but their studies did not include comparing participant responses to ecologically valid stimuli in both the classical and rock contexts. Recently, we reported the results of an experiment in which participants heard a timbre-based style prime (Classical or Rock), followed by either a bVII-I or a V-I cadence (Vuvan & Hughes, 2019).…”