Discussion

To investigate pitch processing in tonal contexts, Experiment 2 introduced a melodic priming paradigm using in-tune/out-of-tune judgments. In contrast to the harmonic priming studies that used a consonance/dissonance task on chords, the use of melodic sequences allowed us to study the processing of target tones’ pitch (in contrast to processing of target chords’ spectral color, i.e. presence/absence of dissonance/roughness in target chords). The results indicate that tonal relatedness influenced speed of pitch processing: the pitch of an in-tune target was processed faster when tonally related to the melodic context (tonic) than when less related (subdominant). This finding extends the outcome of Experiment 1 to processing speed: in a tonal context, listeners’ tonal expectations are very precise for the pitch of the tonic tone leading to faster pitch processing of a related tonic in comparison to a less-related subdominant. The tonal hierarchy of the context (and its implied key) sets a cognitive reference point (the tonic, see Krumhansl, 1990) and induces expectations for this reference point. The pitch of the target can then be compared to this reference pitch - a comparison that speeds up responses when the target matches expectations (actually is the tonic). Since the subdominant is less expected, listeners’ expectations for its pitch may be less precise and judging the correct pitch is slowed down.

Tonal relatedness interacted with the type of pitch deviation: only in-tune tonic targets were processed significantly faster than subdominant targets. This difference between in-tune and out-of-tune targets is in agreement with the harmonic priming studies that used consonance/dissonance discrimination tasks. In those studies, processing was faster for consonant related target chords than for consonant unrelated target chords, but this difference was either smaller, absent or even reversed for dissonant target chords (e.g., Bharucha & Stoeckig, 1986; Bigand & Pineau, 1997). This difference between consonant and dissonant targets has been attributed to response bias. The weaker tonal relatedness might result in a less strong integration of the target event into the tonal context (creating like a “contextual tonal dissonance”) and this might favor the categorization of the dissonant target as dissonant, thus diminishing or counteracting the tonal priming effect (i.e., decreasing the difference between related and less-related targets). Additionally, response bias might slow down in-tune judgments for less-related in-tune targets, thus reinforcing the tonal priming effect (see Bharucha & Stoeckig, 1986; Tillmann et al., 2006, for a discussion). The interaction between tonal relatedness and pitch deviation observed in Experiment 2 can be explained in a similar way: participants may be biased toward judging less-related target tones as being out-of-tune because they are less well tonally integrated to the context. For less-related out-of-tune targets, this bias would speed up the out-of-tune judgments, thus attenuating the tonal priming effect. The fact that the response bias did not reverse the priming effect for out-of-tune targets (i.e., with faster response times for less-related than related targets) might be due to the subtle tonal relatedness manipulations, as suggested by harmonic priming data: for dissonant target chords, the harmonic priming effect was less pronounced for the opposition of tonic versus subdominant chords (Bigand et al., 1999), while it was reversed for the opposition of tonic versus out-of-key chords (Bharucha & Stoeckig, 1986; Tillmann et al., 1998).