To what extent is music like language? Previously, I’ve reviewed how music and language share semantic characteristics, at least insofar as similar scalp electrical activity follows incongruent musical passages as follows incongruent words. But is it also possible that music has grammar, just like language?
In the context of language, ungrammatical words are often accompanied by a sharp rise in scalp electrical activity, around 600 milliseconds after the grammatical violation. This is known as the P600, or sometimes as the syntactic positive shift (SPS). Because this scalp electrical component shows sensitivity to grammatical violations, some suggest that it indexes the function of a grammar-specific module. However, the P600 can also be observed in response to violations of number and gender agreement, phrase structure, verb tense, and even case, suggesting the response is much more general than would be expected from a correlate of a “grammar module” (if one were to exist).
In their 1998 JoCN paper, Patel et al. show that the P600 can distinguish between chord changes that are appropriate to western music versus those that violate more typical chord progressions. To demonstrate this, the authors compared the ERP responses of subjects to phrases like “an old idea” when embedded in three different types of sentences: those that were grammatically OK (“Some of the senators had promoted an old idea of justice”) grammatically poor (“Some of the senators endorsed promoted an old idea of justice.”) and those that were flat-out ungrammatical (“Some of the senators endorsed the promoted an old idea of justice.”)
They then did the same thing, except with music: the authors constructed musical passages based on the “circle of fifths” and inserted notes that were either concordant or discordant with those passages. The inserted notes were never “wrong” in and of themselves (i.e., they were not mistuned or unusual); they simply did not seem to match the harmonic progressions of the musical passage.
The recordings of scalp electrical activity in response to these stimuli revealed that ungrammatical phrases and discordant notes alike were associated with a P600 component, when compared with grammatical phrases and concordant notes respectively. The positive shifts of scalp electrical activity in response to structurally/grammatical incongruous stimuli were statistically indistinguishable from one another, at least in terms of their amplitudes and scalp distributions. So, music seems to have a form of implicit rule structure very much like grammar in language, to the extent that similar neural generators may be involved in processing anomalies across those modalities!
Patel et al note that this similarity is even more striking considering that P600s were evoked from musical passages by manipulating the identity of the target, whereas they were evoked from ungrammatical phrases by manipulating the stimuli preceeding the target.
The authors conclude that the P600 may reflect “knowledge-based structural integration” rather than the activity of a grammar-specific cognitive module. I’m not entirely sure what this means, and they don’t fully explain, but it is important to note that the P600 may be a very late version of the P300 response. The P300 is thought to represent attentional reorienting or the updating of working memory as a result of context violations. Unfortunately, they do not discuss why the P600 differs from the N400 response, also shown to be sensitive to words that are incongruent with the preceeding context (whether verbal or musical), but only when the violations have to do with mismatches in meaning.