The Life of an Image (on Tumblr)
Despite shifts towards an increasingly visual web through the development of highly visual content creation platforms such as Tumblr, Pinterest and Flickr, images are still a difficult kind of digital object to track and trace. We were fascinated by this problem, and wanted to see if we could follow an image through the web. As Tumblr had not yet been scraped using Digital Methods tools, we decided to take up the challenge.
As well as looking at typical Tumblrs which aggregate content according to users' tastes, we also took inspiration from an artwork called Hypergeography
, in which artist Joe Hamilton approaches Tumblr as a (manipulable) medium, rather than a platform to be taken at face value. He was interested in notions of environment (natural, built, networked), and mined Tumblr for images evoking this theme. He collaged other users’ Tumblr images to create a montage of 100 posts that link together horizontally and vertically. The viewer can click on any square in the montage to view the 'source' Tumblrs - e.g Tumblrs from which images have been taken and collaged to form a new image.
While Hamilton's intervention is made by coding, looping, collaging etc, we wanted to take a different approach using the digital methods principle of following the digital object.
Instead of intervening in the platform (or medium), we proposed to follow the medium itself with what it enables and what it constrains.
The project was also inspired and informed by Thrift's concept of 'theatricality' and Marres notion of 'issuefied' objects. In "The Environmental Teapot and Other Loaded Household Objects: Re-connecting the Politics of Technology, Issues and Things" Noortje Marres distinguishes between two types of objects: scripted (Akrich 1992) and issuefied (becoming charged with issues). She also sees digital devices as being capable of 'loading' objects with issues. Embracing this, we were aiming to see whether the opposite is also possible: can an already “issuefied” object become "de-issuefied" by the digital device, in our case Tumblr?
In their article entitled “Materials and Devices of the Public: An Introduction” Noortje Marres and Javier Lezaun (2011) say that the “design of a particular environment for the use and handling of objects is a central aspect of the configuration of a particular public” (p.496). In an attempt to articulate this function of design the authors introduce three capabilities of environmental design which they coin as a “normative valuation” of material publics. These capabilities are described in terms of how they enable degrees of (p.500-503):
1. dynamics of engagement and separation
2. effort and comfort
3. experimentation and diffussion
This characterization of “environments” provides a language to describe platforms' potential ability to handle objects and enable publics around them.
The research aims to critique Tumblr using digital methods as opposed to an interventionist artistic approach (as per Joe Hamilton).
1) How does an image circulate through Tumblr - how does it shift contextually on its journey?
2) If we take an already politicised image can we trace its de-politicisation on Tumblr? Does Tumblr dumb the image?
The Tumblr platform was chosen because it is a highly visual content creation platform which forms economies of taste and community.
Image 1 selection criteria
Sufficient number of reblogs (1000)
Image is issufied
Image 2 Selection criteria
Sufficient number of reblogs (1000)
Image is neutral
We decided to follow 2 photos: one 'issuefied' and one 'neutral'. Our issufied image was the official 'Kony2012' poster taken from the Forgotten Children Tumbler page
and our 'neutral' image, was of a mineral chosen from the Tumblr page of Mineralia
. The neutrality of the mineral was heightened, as we had been presented with images and accounts of conflict minerals in presentations earlier in the week. Our mineral was unconflicted - a low friction mineral.
We built two devices on the top of Tumblr using its API.
1. Retrieve re-blog list: enter URL of the image's page into the rebloggers/likers list tool
2. Copy list of reblogged/liked usernames into first field in the tumblr scraper tool (ensure there are no blanks before or after the username)
3. Go to the actual image - copy image url and enter into second field in the tumblr scraper tool
4. Database generated which gives us profile name, profile description, image caption, timestamp, tags
5. Go to the image source in archive and take screenshot. Screenshot protocol