One of the gimmicks of the 1948 film Words and Music was the question of which is more important in a song—the words or the music.
The movie, a fictionalization of the lives of the legendary team of Broadway writers Richard Rogers (music) and Lorenz Hart (words), only addresses its titular conflict obliquely. The implicit answer, of course, is that you can’t have one without the other.
Or can you? While it’s difficult or perhaps impossible to scientifically determine whether words or music is more important in a song, Isabelle Peretz, Monique Radeau, and Martin Arguin may have done the next-best thing: they devised a study to try to determine whether language or music is more memorable.
In their study, Peretz et al. played recorded snippets of 48 familiar French folk songs to native French speakers. The snippets included two different recognizable parts of the songs: the beginning, and some other portion. (A corresponding example in English would be “happy birthday” and “to you.”) A preliminary experiment showed that listeners had difficulty identifying the songs from the middle portions, so the order of the snippets was reversed (so, for example, they would play “to you” followed by “happy birthday.”) Listeners were instructed to press a button as soon as they recognized the clip.
This experiment is an example of a priming study (we have discussed several of these before). The basic priming effect is simple: when a participant has been primed with a related word, they can complete a memory task more quickly. For example, if your task is to identify proper nouns, then you’ll respond more quickly to “The Beatles” if you’ve seen “The Rolling Stones” first.
What Peretz’s team wanted to know was whether music or language primed song recognition best. To do this, they created two versions of each snippet: one where the words of the song were spoken, and the second where the melody was sung—but instead of singing the words, the performer replaced each syllable with the sound “la”. The snippets were combined in four different ways: a spoken prime followed by a spoken target (e.g. “to you happy birthday”), a spoken prime followed by a sung target (“to you la la la la”), a sung prime followed by a sung target (“la la la la la la”), and a sung prime followed by a spoken target (“la la happy birthday”). The primes and targets were differentiated by how they were played: primes were played to just one ear of the listeners’ headphones, while targets were played in stereo.
Initially, Peretz et al. found that when listeners were primed in the same way they were tested (e.g. singing followed by singing, or speaking followed by speaking), they reacted more quickly. However, a vastly more significant effect was found for language targets: when the target was spoken, listeners recognized the song nearly twice as quickly as when it was sung.
While this may seem to be a clear victory for words over music, the team noted that spoken words tended to take much less time than the equivalent sung melody. Perhaps participants were merely getting their information more quickly.
So the team repeated the experiment, this time using a computer to speed up the singing to match the pace of spoken language without changing its pitch. While sung melodies were recognized more quickly than they had been previously, there was still a substantial advantage for spoken language.
But perhaps this is due to the more expressive nature of spoken language—while a sung note can only vary in pitch, each syllable of language conveys much more information. To compensate for this, the team added white noise to each spoken snippet, then compared the resulting recordings to the sped-up sung melodies until a pilot group recognized both types of recordings equally quickly. Then they repeated the original experiment again. This time, there was no significant difference between the priming effect of lyrics or melody. Words and music appear to contribute equally to the memorability of a song.
Peretz et al. suggest that this equal contribution of words and music to the priming effect may be due to the natural similarities of the rhythm of speech and song. It makes some sense—after all, lyricists select their words for their rhythmic qualities, and composers try to find music that appropriately complements the words. And try to imagine a musical written by just “Rodgers,” or only “Hart.” It doesn’t have the same ring as “Rodgers and Hart,” the team behind such unforgettable songs as “The Lady Is a Tramp,” “This Can’t Be Love,” and “Blue Moon.”
Peretz, I., Radeau, M., & Arguin, M. (2004). “Two-way interactions between music and language: Evidence from priming recognition of tune and lyrics in familiar songs.” Memory & Cognition, 32(1), 142-152.