How do I sound?
On why sonic emotion tracking misses the point, the new models of content moderation, clubhouses and the presses.
|Dave Beer||Feb 8|
Tia DeNora’s classic book Music in Everyday Life explores the close connection between music and emotions. In that book DeNora examines attachments to music and the role that songs play in routine life. The emotional response to music is key part of this, as is the way that music can be actively used to manage and negotiate feelings and sentiments.
I was reminded of DeNora’s work when I saw that Spotify are looking at ways to capture voice data and use this to try to identify moods that might then translate into song recommendations. This development brings together two trends in data: emotion tracking and sonic environment capture.
Sonic data has become a bit of a frontier in data gathering. Turning sound environments into data seems to be an area in which the limits of data usage are being pushed outwards. There are problems though. Sonic data seem to produce quite strong feelings - the idea of being listened-in-to can seem more invasive than other types of data tracking. The ambience of it seems to jar - it feels more like being spied upon perhaps.
Yet voice activated devices have grown in numbers drastically in the last three or four years or so. And here in the case of Spotify we see that future developments are likely to seek to gather more sonic data and to feed this into predictions, recommendations and targeting. The idea seems to be that listening-in to the environment and to voices will allow a more direct and anticipatory connection between emotions and playlists. It is a more embodied form of data. The tracking of mood would then become a more integrated part of our cultural landscapes.
One thing that Spotify would have realised if they had read DeNora’s book is that particular emotions do not necessarily lead to a predictable responses in music choice. People don’t necessarily seek to neatly match music to their emotions, nor do they reflect moods with the intended tonal sentiments of songs - they sometimes seek to use music to shift or shape those feelings.
Alongside this, Spotify are unlikely to be able to use the data they hold to understand the complex weave of music and biography that underpins the relations between music and emotions. DeNora identified how a significant part of emotional attachments to music is based upon the close links between songs and past moments and memories. It isn't simply about the properties of the music itself and how this might be considered to evoke certain emotions. In other words, the connections between music and emotion are not necessarily about the style or sound of a song, they are about how people have experienced those songs at different times (for a detailed exploration of this and other musical attachments, see Victor Avila Torres work on ‘How Music Matters’).
Looking not just at the broader responses to sonic data usage but also thinking about the type of active and individual attachments to music charted in DeNora’s work, would suggest that Spotify’s consideration of listening-in-to the streamer may clash with both feelings about data surveillance and with music's role in everyday life.
The prickly issue of content management continues to rumble on. Various models are emerging as social media platforms try to work out how to manage the content posted on their sites.
Here is a piece by Will Oremus that compares the different approaches toward content moderation that are currently being tried out by Facebook and Twitter. Highlighting the contrasting approaches, the piece summarises that:
‘One is punting moderation to its own private international court of justice; the other is hoping to crowdsource it to a cadre of volunteers’
Perhaps these contrasting steps will lead to improvements in this ongoing problem, but I can't help thinking that both seem unlikely to work. Oremus concludes that:
‘What they have in common is that they’re once again seeking legitimacy by offloading responsibility for their platforms’ content, rather than by internalizing and owning it.’
And so the issues persist. The questions about responsibility and the management of content will no doubt continue. Unless more directly addressed, content moderation problems might even be a constant presence within social media.
A new book from Katherine Hayles is always an important moment. Postprint: Books and Becoming Computational is out this month with Colombia University Press.
Nigel Thrift also has a book, Killer Cities, out with the TCS series at Sage.
Deborah Lupton and Clare Southerton have an article on Facebook’s thing-power in the Journal of Sociology.
Are you in the club…
If you are wondering what Clubhouse is, there is a piece describing it here. Effectively it’s a rapidly growing audio discussion focused social media app. In this article John Gapper also predicts that this new site will change quickly as its user base continues to rapidly expand. Taking this example, Gapper reflects on how social media tend to choose ‘growth’ over ‘order’. As Gapper puts it, ‘the valuation rises as the value is diluted’.
Here is an interesting article looking at WEB Dublin Bois' theory of time, progress and the future.
Substack, who are the service I use for this newsletter, have also been reflecting on the expansionist tendencies of social media this week. This has been provoked by big tech moving in on the market. Facebook and Twitter have both developed newsletter features and are seeking to integrate them into their platforms. Hamish McKenzie, one of Substack's cofounders, reflected on this move by the social media platforms in a recent post. It seems that who owns the means of newsletter printing and distribution is coming to matter.
This is the 51st issue of this newsletter. The previous issue on Fredric Jameson’s recent book about the work of Walter Benjamin was the 50th issue and brought with it a slight change of title for the newsletter. The title is still inspired by David Frisby’s book Fragments of Modernity, but I’ve shortened it down to ‘The Fragment’.