W e assume that evaluative processes in response to musical stimuli can occur spontaneously without explicit demand, and that these responses are important for the emergence of emotions evoked by music. Two versions of the affective priming paradigm served to study spontaneous evaluation of music. In Experiment 1, a lexical decision task (LDT) and in Experiments 2 and 3, an evaluative decision task (EDT) was employed. A total of 20 original four-part, five-chord piano sequences with no specified harmonic resolution were used as primes. During the LDT, congruency in valence of prime-target pairs did not affect response times to the targets. However, for the EDT, significant effects of priming were obtained, indicating that spontaneous evaluations of primes must have occurred. No moderating influences of music expertise or any other person variable on spontaneous evaluation were observed. The diverging results of LDT and EDT point to the possibility that spontaneous evaluative processes are sensitive to context manipulations. Results are discussed with reference to harmonic and semantic priming studies.