Drawing on a range of technology and evolutionary biology, scientists explore the development of pop music over the last 50 years. It’s the latest example of how the use of large datasets by scientists is likely to revolutionize our understanding of creativity and culture.
Analysis of cultural change and trends is always a hot topic of debate and when it comes to popular music, the UK has a long tradition of music magazines and journalists staking out critical positions around the value of different kinds of music – from Paul Morley in the 1970s, to Jon Savage in the 1980s to Simon Reynolds in the 1990s. But new technologies and Big Data are generating new disciplines such as Digital Humanities, so perhaps it should have been no surprise when computer scientists and evolutionary biologists – from Imperial College London and Queen Mary, University of London – turned their minds to considering the history of pop music over the last 50 years.
Published in the Royal Society Open Science journal, the title of the paper itself has something epic about it, like the name of a futuristic 1970s Progressive Rock album – The Evolution of Popular Music: USA 1960–2010. The paper looked at the evolution of music style through analyzing more than 17,000 songs on the US Billboard Hot 100 – before the age of downloads and MP3s the US Billboard Hot 100 was an accurate reflection of mainstream tastes.
Music remains diverse
‘We wanted to have a database of things which we knew people listened to,’ explains one of the study’s authors, Dr Matthias Mauch, from the school of Electronic Engineering and Computer Science at Queen Mary University of London. ‘So the Billboard Hot 100 seemed the most obvious choice.’ In an age when all this music from the past is readily available in digitized form makes, studies like this become possible.
And while a common assumption is often that music is becoming increasingly homogenous the study doesn’t reflect this. It is arguable that the stereotype often simply reflects a generational divide; younger people’s music often doesn’t speak to the immediate experience of an older generation, so ‘it all sounds the same’!
‘The most salient findings for me are that we measured the diversity,’ says Dr Mauch, ‘and we can see a definite drop in diversity only in the 80s.’ The area of study was a natural evolution for Dr Mauch himself – he had studied Maths in Germany, been an amateur musician in both classical and pop, and had a band in his hometown of Rostock. He became interested in informatics and so Queen Mary’s London was a natural destination, and he restarted his PhD there.
The project’s combination of evolution and informatics is what anchors the research interests of Dr Mauch and Senior Author of the study, evolutionary biologist Professor Armand Leroi, from the Department of Life Sciences at Imperial. ‘We have these complementary interests with him being interested in evolution and music and me a music processing guy interested in evolution actually even before I met him,’ explains Dr Mauch. ‘I was doing a post-doctorate in Japan and wondering what I could do when I got back to London and I saw this guy giving a talk on the evolution of music which was a preliminary study that he did at Imperial and I thought “wow!” But they do that without audio-analysis and I thought we can do that better!’
When Dr Mauch returned to London he worked at music recommendation website Last FM, working on audio processing, matching data from the billboard charts to the audio which laid some groundwork for the future project. The current method of analysis he explains, ‘mixes signal processing and techniques even from text mining, so we combined all those to get a grip on what the signal meant.’ There had been one study which had done something similar, but Dr Mauch believes ‘they had missed the opportunity to really make the reader understand what it is that changes. So we are focussing on timbre and harmony. There are genres that people tend to talk about and we also did a little exploration into that. We had data describing every song and this allowed us to cluster the songs, which turned out to deliver a very nice graph of how the charts were composed – the data categories that we found corresponded very nicely with the genre.’ By studying 100 songs a week over 50 years the charts effectively became the equivalent of a fossil record.
Dr Mauch believes that the reason for the lack of diversity in the 1980s sound was due the dominance of the percussive and the guitar. ‘Then of course we had these “revolutions” which is a slightly populist way of naming these kinds of things. They are what biologists would call “punctuations” of the slow rate of evolution.’ The three moments when this evolution accelerated were in 1964, 1983 and 1991. ‘We don’t have a fully causal analysis,’ says Dr Mauch, ‘but they coincided very nicely with firstly the British invasion – The Beatles and The Rolling Stones coming to the United States in 1964. 1983 saw the introduction of samplers, synthesisers and drum machines. Then the rap and hip-hop revolution in the 1990s when hip-hop suddenly became national music rather than being a regional phenomenon and it spread into the charts.’
The Beatles and The Stones
The hip-hop moment stood out as the study focussed on harmony, while hip-hop is driven by the spoken word and rhythm. The study discovered that while the British invasion didn’t spark the shift in 1964, The Beatles and The Rolling Stones accelerated and expanded a dynamic already taking place – these two bands had 66 Hot 100 hits between them during the period up to 1968. Also there was the impact on change of their musical influence as other musicians were inspired by their style and sound.
Though the study conducted by Dr Mauch and team confirms our intuitive understanding of the general history of pop, the kind of data scientists are increasingly able to draw on is likely to offer alternative readings of culture, or at least probe more deeply our accepted versions of cultural history. The paper for example raises the issue of the cultural dynamics of change in music, acknowledging it doesn’t address this. ‘For example,’ the authors write, ‘the rise of rap in the charts has been credited to the television show Yo, MTV Raps! first broadcast in 1988.’ But they suggest that there are approaches from evolutionary biology that in future can begin to test such popular wisdom. These are exciting times for both science and the humanities. As research such as this becomes more widespread, some studies will support whilst others will challenge our long-held beliefs around culture and creativity.