Our starting point is the very last entry in the target article's Table 4, which indicates that musical expectancy is the only mechanism that depends strongly on musical structure. Fortunately, the field of music theory provides conceptual tools for analyzing music, and this raises the question of what musical structures give rise to expectations and what are the emotional consequences. Our approach is to begin on the musical side and consider how understanding musical processes leads to a somewhat different perspective on musical emotions than that associated with Juslin & Västfjäll's (J&V's) other five mechanisms.

We identify Leonard Meyer's (1956) monograph, Emotion and Meaning in Music, as the most influential theoretical framework for studying musical emotions. Its success stems from his shift from the question “Why does music produce emotions?” to the more tractable question, “How does music produce emotions?” This focuses attention on the music itself and how it is constructed. Another important aspect of Meyer's theory is that it deemphasizes the general mood (such as happy, sad, or peaceful) engendered by passages, movements, or entire musical pieces, and emphasizes instead the moment-to-moment response to the ongoing flow of music.

The theory's essential claim is that music produces emotions because listeners actively generate expectations (mostly unconsciously) for what is to follow. Depending on the relationship between these expectations and what actually happens, listeners experience varying degrees of tension or relaxation. In Meyer's words, “Thus in a very general way expectation is always ahead of the music, creating a background of diffuse tension against which particular delays articulate the affective curve and create meaning” (Meyer Reference Meyer1956, p. 59). The feeling of tension is not necessarily negative, nor is the feeling of resolution necessarily positive. Rather, the response depends on the particular way expectations are fulfilled, perhaps in a particularly artful way or at an unexpected delay.

Meyer emphasizes three different sources of expectation. The first, extra-opus knowledge or style knowledge, refers to quite general patterns in a musical style. These are codified in music theory, and empirical research extensively documents that listeners' knowledge about melody, harmony, and rhythm influences what they expect in a given musical context (e.g., Bharucha & Stoeckig Reference Bharucha and Stoeckig1986; Boltz Reference Boltz1993; Jones Reference Jones1990; Jones et al. Reference Jones, Johnston and Puente2006; Krumhansl Reference Krumhansl1990; Schmuckler Reference Schmuckler1989). This knowledge does not depend strongly on explicit musical training; non-musicians internalize it through passive exposure.

A second source of expectations, called intra-opus knowledge, refers to the listener's experience of a particular piece of music and the expectations that are based on its characteristics. For example, if a piece of music begins with a particular theme, then the listener will expect that the theme is likely to recur later in the piece or reappear in variations.

Meyer also emphasizes the influence of Gestalt principles of perceptual organization on music perception. In this tradition, Narmour (Reference Narmour1990; Reference Narmour1992) proposed what is called the implication-realization model. Its five principles for melodic expectations have been tested using a fairly wide variety of musical styles and listeners in different cultures (e.g., Krumhansl Reference Krumhansl1995; Krumhansl et al. Reference Krumhansl, Louhivuori, Toiviainen, Järvinen and Eerola1999; Reference Krumhansl, Toivanen, Eerola, Toiviainen, Järvinen and Louhivuori2000; Thompson & Stainton Reference Thompson and Stainton1998). The results find support for principles such as: Generally expect small changes in pitch, but if there is a large jump expect a tone that fills the gap.

To study the rise and fall of tension, real-time measures have been developed in which listeners move a device to indicate the amount of tension they experience throughout the course of a piece or a segment of music (e.g., Fredrickson Reference Fredrickson1995; Krumhansl Reference Krumhansl1996; Nielsen Reference Nielsen1983). Theoretical models, especially Lerdahl's (Reference Lerdahl2001) tonal pitch-space model, have been developed to provide a precise account of the degree to which musical structures produce tension (see Lerdahl & Krumhansl Reference Lerdahl and Krumhansl2007).

But how does this relate to the more usual sense of emotion? Several studies suggest music results in changes in emotion physiology associated with real-life emotions. For example, respondents in Sloboda's (Reference Sloboda1991) questionnaire study were able to pinpoint the particular moment in pieces where they experienced, on repeated occasions, a strong emotion, and these coincided with points in the music where an expectation of some kind was violated. Different physiological reactions (such as tears or shivers down the spine) were produced by different kinds of violations.

Real-time judgments of tension in the study by Krumhansl (Reference Krumhansl1997) correlated most strongly with real-time judgments of fear, but judgments of happy and sad also made a contribution. Changes in emotion physiology showed a similar pattern. This suggests that tension is a multivalent attribute influenced by different emotions. Supporting this, Krumhansl and Schenck (Reference Krumhansl and Schenck1997) found that judgments of tension were almost identical with judgments of the overall amount of emotion for both music and dance.

Early event-related brain potential (ERP) studies (Besson & Faita Reference Besson and Faita1995; Janata Reference Janata1995) found correlates of the degree to which expectations are violated, a result replicated in other studies (e.g., Koelsch et al. Reference Koelsch, Gunter, Friederici and Schröger2000). A recent study (Steinbeis et al. Reference Steinbeis, Koelsch and Sloboda2006) bridged the gap between musical expectancy and emotion by measuring listeners' physiological responses to music. Tension, subjective emotionality of the music, an early negativity ERP response, and electrodermal activity (EDA) increased with harmonic unexpectedness.

Blood and Zatorre's (Reference Blood and Zatorre2001) positron emission tomographic (PET) study showed brain responses at specific listener-identified time points with strong emotions, as well as other physiological changes. In an functional magnetic resonance imaging (fMRI) study, areas of secondary auditory cortex were active when listeners heard violations of expectations for pitch and rhythm (Krumhansl Reference Krumhansl2005). Lastly, a study using irregular, unexpected chords (Koelsch et al. Reference Koelsch, Fritz, Schulze, Alsop and Schlaug2005) found that unexpected chords elicited orbital frontolateral cortex activation, an area shown to support emotional processing.

In summary, empirical evidence, using a variety of behavioral and neuro-cognitive measures, strongly supports the idea that listeners develop constantly changing expectations while listening to music, and these give rise to waves of tension and relaxation. At present, however, it appears that these moment-to-moment responses do not map in a simple way onto the traditional emotional states studied within psychology. Although the emotions evoked by music may be different or more complex than real-life emotions, they are fundamental to the musical experience.