[I wrote a recent Muse for Nature News on an interesting study of the emotional qualities of major and minor keys. Here it is (pre-edited). I should say that I could do no more here than hint at the problems I had with the Bowling et al. paper. It is very stimulating – I’d not seen a claim of this sort made before – but ultimately I find it unconvincing. Their procedure is pretty hard to follow, but I think I’ve got it right in the end. I find it very odd that they are apparently digging out some ‘implied fundamental’ for all the tonic intervals they consider, more or less regardless of whether there is any evidence that such a thing is heard (in the absence of the tonic actually being simultaneously played!). And as I say, the formant ratios for both types of speech are dominated by major intervals, but simply less so for ‘subdued’ speech – that’s to say, this speech doesn’t seem to have a ‘minor’ feel to it (if such a thing is meaningful anyway), but just less strongly major. So the issue is very much open. But in any event, empirical evidence surely shows us that music using modes close to the Western diatonic minor needn’t be sad at all in other cultures.]
Spinal Tap’s Nigel Tufnell famously declared that D minor is “the saddest of keys”. But is music in a minor key inevitably sad?
Why does Handel’s Water Music and the Beatles’ ‘There Comes The Sun’ sound happy, while Albinoni’s Adagio and ‘Eleanor Rigby’ sound sad? The first two are in major keys, the second two in minor keys. But are the emotional associations of major and minor intrinsic to the notes themselves, or culturally imposed? Many music psychologists suspect the latter, but a new study suggests there’s something fundamentally similar about major and minor keys and the properties of typically happy and sad speech, respectively.
Neuroscientists Daniel Bowling and colleagues at Duke University in Durham, North Carolina, say in a paper in the Journal of the Acoustical Society of America that the sound spectra – the profiles of different acoustic frequencies – in major-key music are close to those in excited speech, while the spectra of minor-key music are more similar to subdued speech . They compared the frequency ratios of the most prominent acoustic peaks in speech (called formants) with those in Western classical music and Finnish folk songs.
The acoustic characteristics of happy, excited speech, which is relatively fast and loud, are common in most cultures, while sadness elicits slower, quieter vocalizations. We have a natural tendency to project such physiognomic associations onto non-sentient objects: a drooping willow is seen as ‘weeping’. There’s good reason to believe that music mimics some of these universal emotional behaviours, supplying a universal vocabulary that permits listeners sometimes to deduce the intended emotion in unfamiliar music. For example, Western listeners were able to judge fairly reliably whether pieces of Kyrghistani, Hindustani and Navajo Native American music were meant to be joyous or sad [2,3], while the Mafa people of Cameroon who had never heard Western music could guess more often than chance whether extracts were intended to be happy, sad or ‘fearful’ . Here it seems that tempo was the main clue.
Of course, it’s simplistic to suppose that all music is ‘happy’ or ‘sad’, or that all ‘happy’ music is equally and identically ‘happy’, as opposed to joyous, blissful, contented and so forth. But these crude universal indicators of emotion do seem to work across borders.
Is mode (major/minor) another of them? The idea that the minor key, and in particular the musical interval between the first and third note of the scale (a so-called minor third) is intrinsically more anguished than the major (where the major third seems naturally ‘bright’ and optimistic) is so deeply ingrained in Western listeners that many have deemed this to be a ‘natural’ principle of music. This notion was influentially argued by musicologist Deryck Cooke in his 1959 book The Language of Music.
Cooke pointed out that musicians throughout the ages have used minor keys for vocal music with an explicitly sad content, and major keys for happy lyrics. But he failed to acknowledge that this might simply be conventional rather than innate. And when faced with the fact that some cultures, such as Spanish and Slavic, use minor keys for happy music, he offered the patronizing suggestion that such rustic people were inured to a hard life and didn’t expect to be happy.
No such chauvinism afflicts the latest work of Bowling and colleagues. But their conclusions are still open to question. For one thing, they don’t establish that people actually hear in music the characteristic spectral signatures that they identify. Also, they assume that the ratios of frequencies sounded simultaneously in speech (what in music are called harmonic intervals) can be compared with the ratios of frequencies sounded sequentially in music (melodic intervals). And most troublingly, major-type frequency ratios dominate the spectra of both excited and subdued speech, but merely less so in the latter case.
In any event, this work still faces the problem that some cultures (including Europe before the Renaissance, not to mention the ancient Greeks) don’t link minor keys to sadness. Western listeners sometimes misjudge the emotional quality of Javanese music that uses a scale with similarities to the minor mode yet is deemed ‘happy’ by the musicians. So even if a fundamental ‘sadness’ is present in the minor mode, it seems likely to be weak and easily over-written by acculturation. It’s possible even in the Western idiom to write ‘happy’ minor-key music (for example, van Morrison’s ‘Moondance’) or ‘sad’ major-key music (Billie Holiday’s ‘No Good Man’).
So let’s not conclude too soon that minor keys give everyone the blues.
1. Bowling, D. L., Gill, K., Choi, J. D., Prinz, J. & Purves, D. J. Acoust. Soc. Am. 127, 491-503 (2010).
2. Balkwill, L. L. & Thompson, W. F. Music Perception 17, 43-64 (1999).
3. Juslin, P. N. & Kaukka, P. Psychological Bulletin 129, 770-814 (2003).
4. Fritz, T. et al., Curr. Biol. 19, 1-4 (2009).