Almost everyone can recognize a familiar melody like Happy Birthday, God Save the Queen, or Yesterday, even if the tune is played instrumentally with no words. Listeners can also recognize these tunes when they are played very quickly in a high pitch register on a piccolo, or very slowly in a low pitch register on a tuba. In other words, a melody is defined based on the relations between consecutive notes in terms of their pitch and duration, such that the actual key, tempo, and musical instrument (timbre) are irrelevant to a tune’s identity. In order for this to be possible, a listener’s mental representation of a familiar melody must include the relational information that defines the tune.
Initially, the basic research conducted in my laboratory asked whether some musical relations are naturally easier than others to process and remember (Schellenberg & Trainor, 1996; Schellenberg & Trehub, 1994a, 1994b, 1996, 1998), and whether other universal principles govern the structure of melodies (Schellenberg, 1996, 1997; Schellenberg, Adachi, Purdy, & McKinnon, 2002). We are now investigating whether aspects of a particular musical performance—its key, tempo, or timbre—are incorporated into listeners’ mental representations for music, and how such representations change over development.
Some of the more interesting findings show that undergraduates with no music lessons can recognize familiar recordings from excerpts that are only 1/10th of a second long (Schellenberg, Iverson, & McKinnon, 1999), and that they can identify whether a familiar recording is presented in the correct key (Schellenberg & Trehub, 2003). Even infants and children can tell whether a familiar recording is presented in the proper key (Volkova, Trehub, & Schellenberg, 2006; Schellenberg & Trehub, 2008; Trehub, Schellenberg, & Nakata, 2008). As children get older, they focus less on key and more on the pitch relations that define music (Stalinski & Schellenberg, 2010).
When listeners are exposed to melodies they have never heard before, they are better at recognizing them subsequently if they are reheard in the original key, tempo, or timbre (Schellenberg & Habashi, 2015; Schellenberg, Stalinski, & Marks, 2014), which means that these features must be incorporated into their memories. If you wait to test recognition until a full week after listeners first hear them, however, it no longer matters whether the melodies are heard in the original key or tempo—these features fade from memory over time. Playing the melodies in the same timbre still has a benefit, which indicates that memory for timbre is longer lasting than memory for key or tempo.
As it turns out, all timbres are not created equal. Adults’ and children’s memory for melodies is better if they are sung rather than performed on a musical instrument, even when the singer doesn’t sing any words, just la, la, la (Weiss, Schellenberg, Trehub, & Dawber, 2015; Weiss, Trehub, & Schellenberg, 2012; Weiss, Vanzella, Schellenberg, & Trehub, 2015). The human voice appears to be a special auditory stimulus, one that captures a listener’s attention such that individuals with perfect (absolute) pitch find it particularly difficult to name notes when they are sung rather than presented instrumentally (Vanzella & Schellenberg, 2010).
See the Publications page to view (and see complete citations for) the journal articles referenced above.