Where are musicians most creative?
In our recent paper “ Innovation in Creative Industries: Does (Related) Variety Matter for the Creativity of Urban Music Scenes?” together with Simone Strambach we looked at the relationship between the specialization of urban music scenes and innovation in music. New genres and new combinations of existing genres thrive best where there is a related variety of music genres. And the specialization of music scenes might not be such a good thing for their creativity.
As a musician, big music fan and geographer I was always fascinated by the origin stories of music genres and their relationship to places. Some cities stand out as birth places and centers for very distinct music styles, e.g. country in Nashville, post-punk/indie in Manchester or melodic death metal in Gothenburg. But what made these places the birth places of new music genres? Is it all random, or do the characteristics of music scenes at the time when music genres emerge play a role? If, and how do the surroundings of artists influence their creativity?
The histories of popular musicians and styles hinted at that some places seem to be more inspiring than others by providing the space in which music genres emerge because different styles mixed, or musicians were able to further develop existing styles, or benefitted from a variety of inspiring sources. There are a lot of narrations and qualitative case studies of specific music genres, but sadly, only a few quantitative studies that look at several places and several genres at once. Thus, we were motivated to find data that characterize music scenes and find some general relationships between characteristics of urban music scenes and the new genres that emerge from them.
Luckily, there is last.fm, a social music platform established in 2002, which provides user-generated tags and data on where musicians come from and which music genres they belong to. We used this to create a database of about 9,000 artists from 33 North American and European cities who were active at some point between 1970 and 2015. It allowed us to characterize music scenes, not only by some indicators of specialization and variety, but to visualize them as network-based music landscapes which we called “symbolic knowledge spaces” (see Figure 1).
In these networks, each node is a music genre that artists from this city at this point in time belong to. The links between two nodes indicate how often artists belong to both of them. The location of different music genres in the network indicates their relatedness — the closer they are to each other, the more similar they are. This way to visualize the activites present in a region is heavily inspired by product spaces (free link here) developed by Cesar Hidalgo et al. The colours of the nodes are derived from an own classification system that categorizes each genre to a “stream”. This type of visualization also shows how music scenes evolved over the years, as the figure 2 below demonstrates.
The link between specialization and innovation in music
We were most interested in which types of music scenes where more conducive to innovation in music. Is it the specialized ones, or those with a lot of variety? Or is it something in between?
For this, we derived sets of indicators for 1) innovation in music and 2) the characteristics of music scenes.
We understood innovation in music as the creation, combination and exploitation of so-called symbolic knowledge, indicated by the number of new genres created in a music scene, the number of pioneering artists, the number of artists making new combinations and the number of superstars in a music scene. Figure 3 shows the top and bottom five music scene for each indicator plus their definitions.
Furthermore, we measured the characteristics of each music scene every 5 years (starting 1970) by various indicators for specialization and (related) variety. We will spare you the math here (for all the formulas and regression tables, see the paper here), but Figure 4 shows the average values of the different measures over all time intervals (1970–74, 75–79, etc.). Following this, and this, we differentiated between variety that is unrelated, semirelated and related. Unrelated variety is the variety across the different music streams (see Figure 1). It’s high in rather unfocused music scenes with a high variety of very different music genres from different streams. Semirelated variety refers to the variety of music styles on the classification level below these streams (see Figure 5 for own classification system of music). Semirelated variety is high, e.g. when there are a lot of different electronic styles: For instance, there’s techno, house, dubstep, but also idm, industrial, ambient, etc. Finally, related variety refers to the variety of sub-genres within the music styles. This is high, for instance, where there are a lot of sub-genres of Metal, Techno, or Punk music.
What we found is, most importantly, that specialization was negatively related to the emergence of new music genres. While music fans, and more and more also city officials, love to associate cities with certain music scenes, it does not seem to foster innovation very much. Apparently, musicians in these cities fare best when they stick to the styles that are already present and and it may also prevent them from experimentation with new ideas. But at the same time, too much unrelated variety was also not linked to innovation in music. Instead, there is a sweet spot: Semirelated variety is linked positively to all indicators of innovation in music. In addition, the combination of music genres benefits from related variety.
This means that innovation in music benefits from music scenes in which there is a variety of music genres that are also related to each other. Because of the quantitative nature of this study, we cannot be exactly sure of the mechanisms behind this, but our results point to the combination of music genres as the driver of this relation. For new music genres to emerge, musicians need surroundings that provide them with a variety of inspirational sources that are also compatible with each other. Then, they can find ways to combine them in new exciting ways that work and capture new audiences. New, at first improbable combinations such as the one created by DEAFHEAVEN (according to last.fm users it’s a mix of shoegaze and black metal) may be more able to find a local following in music scenes with high semirelated variety than in specialized or diverse settings.
What does this mean for creative city policy?
This study also has implications for policy in the areas of creative/music city promotion. It calls attention to the role of the structure of music scenes for their further development. Not only the quantity of actors should be monitored but also to what genres/styles of music these belong and how these genres relate to each other. Policy actors concerned with the promotion of creative industries, especially the music industry, are well advised to pay attention to the evolution of the composition of music scenes. Most especially, an overspecialization has to be prevented, since this limits innovativeness and adaption to new trends, which are essential for the long-run success of creative cities. Thus, policy measures should counteract specialization tendencies by identifying scene-specific streams of music and coordinate efforts to enhance variety within these streams.
I hope you enjoyed this blog on our recent paper — please follow me here or on twitter (https://twitter.com/innogeo) for upcoming findings on the characteristics of music scenes and their innovativeness.
