28 August 2014

On Latour's Notion of the Digital

Bruno Latour at Digital Humanities 2014
Bruno Latour, professor at Sciences Po and director of the TARDE program (Theory of Actor-network and Research in Digital Environments), recently outlined his understanding of the digital in an interesting part of his plenary lecture at Digital Humanities 2014 conference. He was honest in accepting that his understanding may itself be a product of his own individuation and pre-digital training as a scholar which emphasised close-reading techniques and agonistic engagement around a shared text (Latour 2014). Nonetheless, in presenting his attempt to produce a system of what we might call augmented close-reading in the AIME system, he was also revealing about how the digital was being deployed methodologically and his notion of the digital's ontological constitution.[1]

Unsurprisingly, Latour's first move was to deny the specificity of the digital as a separate domain as such, highlighting both the materiality of the digital and its complex relationship with the analogue. He described both the analogue structures that underpin the digital processing that makes the digital possible at all (the materials, the specific electrical voltage structures and signalling mechanisms, the sheer matter of it all), but also the digital's relationship to a socio-technical environment. In other words, he swiftly moved away from what we might call the abstract materiality of the digital, its complex layering over an analogue carrier and instead reiterated the conditions under which the existing methodological approach of actor-network theory was justified – i.e. digital forms part of a network, is "physical" and material, requires a socio-techical environment to function, is a "complex function", and so on.

Slide drawn from Latour (2014)
It would be too strong, perhaps, to state that Latour denied the specificity of the digital as such, but rather through what we might unkindly call a sophisticated technique of bait and switch and the use of a convincingly deployed visualisation of what the digital "really" is, courtesy of an image drawn from Cantwell-Smith (2003) the digital as not-physical was considered to have been refuted. Indeed, this approach to the digital echoes his earlier statements from 1997 about the digital, such that Latour argues,[2]
I do not believe that computers are abstract... there is (either) 0 and (or) 1 has absolutely no connection with the abstractness. It is actually very concrete, never 0 and 1 (at the same time)... There is only transformation. Information as something which will be carried through space and time, without  deformation, is a complete myth. People who deal with the technology will actually use the practical notion of transformation. From the same bytes, in terms of 'abstract encoding', the output you get is entirely different, depending on  the medium  you use. Down with information (Lovink and Schultz 1997).
This is not a new position for Latour, indeed in earlier work he has stated "actually there is nothing entirely digital in digital computers either!" (original emphasis, Latour 2010a). Whilst this may well be Latour's polemical style getting rather out of hand, it does raise the question about what it is that is "digital" for Latour and therefore how this definition enables him to make such strong claims. One is tempted to suppose that it is the materiality of the 0 and 1s that Cantwell Smith's diagram points towards that enables Latour to dismiss out of hand the complex abstract digitality of the computer as an environment, which although not immaterial, still is located through a complex series of abstraction layers which actually do enable programmers to work and code in an abstract machine disconnected in a logical sense from the materiality of the underlying silicon. Indeed, without this abstraction within the space of digital computers there could be none of the complex computational systems and applications that are built today on abstraction layers. Here space is deployed both in a material sense as the shared memory abstracted across both memory chips and the hard disk (which itself may be memory chips) and as a metaphor for the way in which the space of computation is produced through complex system structures that enable programmers to work as programmers working within a notionally two-dimensional address space that is abstracted onto a multidimensional structure.

The Digital Iceberg (Berry 2014)
In any case, whilst our attention is distracted by his assertion, Latour moves to cement his switch by making the entirely reasonable claim that the digital lies within a socio-technical environment, and that the way to study the digital is therefore to identify what is observable of the digital. This he claims are "segments of trajectories through distributed sets of material practice only some of which are made visible through digital traces", thus he claims the digital is digital less as a domain and more as a set of practices. This approach to studying the digital is, of course, completely acceptable, providing one is cognisant of the way in which the digital in our post-digital world resembles the structure of an iceberg, with only a small part ever visible to everyday life – even to empirical researchers (see diagram above).  Otherwise, ethnographic approaches which a priori declare the abstractness of the digital as a research environment illegitimate, lose the very specificity of the digital that their well-meaning attempt to capture the materiality of the digital calls for. Indeed, the way in which the digital through complex processes of abstraction is then able to provide mediators to and interfaces over the material is one of the key research questions to be unpacked when attempting to get a handle on the increasing proliferation of the digital into "real" spaces. As such, ethnographic approaches will only ever be part of a set of research approaches for the study of the digital, rather than, as Latour claims, the only, or certainly most important research methodology.

This is significant because as the research agenda of the digital is heightened, in part due to financial pressures and research grants deployed to engage with digital systems, but also the now manifest presence of the digital in all aspects of life, and hence the deployment of methodological and theoretical positions on how such phenomena should be studied. Should one undertake digital humanities or computational social science? Digital sociology or some other approach such as actor-network theory? In his claim that "the more thinking and interpreting becomes traceable, the more humanities could merge with other disciplines" reveals the normative line of reasoning that (digital) humanities specificity as a research field could be usurped or supplemented by approaches that Latour himself thinks are better at capturing the digital (Latour 2014). Indeed, Latour claims in his book, Modes of Existence, that his project, AIME, "is part of the development of something known by the still- vague term 'digital humanities,' whose evolving style is beginning to supplement the more conventional styles of the social sciences and philosophy" (Latour 2013: xx).

To legitimate the claim of Latour's flavour of actor-network theory as a research approach to the digital, he refers to Boullier's (2014) work, Pour des sciences social de çéme génération, that there have been three ages of social context, with the latest emerging from the rise of digital technologies and the capture of digital traces they make possible. They are,
Age 1: Statistics and the idea of society 
Age 2: Polls and the idea of opinion 
Age 3: Digital traces and the idea of vibrations (quoted in Latour 2014).
Here, vibration follows from the work of Gabriel Tarde in 1903 who referred to the notion of "vibration" in connection to an empirical social science of data collection, arguing that,
If Statistics continues to progress as it has done for several years, if the in-formation which it gives us continues to gain in accuracy, in dispatch, in bulk, and in regularity, a time may come when upon the accomplishment of every social event a figure will at once issue forth automatically, so to speak, to takeits place on the statistical registers that will be continuously communicatedto the public and spread abroad pictorially by the daily press. Then, at every step, at every glance cast upon poster or newspaper, we shall be assailed, asit were, with statistical facts, with precise and condensed knowledge of allthe peculiarities of actual social conditions, of commercial gains or losses, of the rise or falling off of certain political parties, of the progress or decay of a certain doctrine, etc., in exactly the same way as we are assailed when weopen our eyes by the vibrations of the ether which tell us of the approach or withdrawal of such and such a so-called body and of many other things of a similar nature (Tarde 1962: 167–8).
This is the notion of vibration Latour deploys, although he prefers the notion of sublata (similar to capta, or captured data) rather than vibration. For Latour, the datascape is that which is captured by the digital and this digitality allows us to view a few segments, thus partially making visible the connections and communications of the social, understood as an actor-network. It is key here to note the focus on the visibility of the representation made possible by the digital, which becomes not a processual computational infrastructure but rather a set of inscriptions which can be collected by the keen-eyed ethnographer to help reassemble the complex socio-technical environments that the digital forms a part of. The digital is, then, a text within which are written the traces of complex social interactions between actants in a network, but only ever a repository of some of these traces.

Latour finishes his talk by reminding us that the "digital is not a domain, but a single entry into the materiality of interpreting complex data (sublata) within a collective of fellow co-inquirers". Reiterating his point about the downgraded status of the digital as a problematic within social research and its pacification through its articulation as an inscription technology (similar to books) rather than a machinery in and of itself, shows us again, I think, that Latour's understanding of the digital is correspondingly weak.

The use of the digital in such a desiccated form points to the limitations of Latour's ability to engage with the research programme of investigating the digital but also the way in which a theologically derived close-reading method derived from bookish practice may not be entirely appropriate for unpacking and "reading" computational media and software structures.[3] It is not that the digital does not leave traces, as patently it does, rather it is that these traces are encoded in such a form, at such quantities and high-resolutions of data compression that in many cases human attempts to read this information inscription directly are fruitless, and instead require the mediation of software, and hence a double-hermeneutic which places human researchers twice (or more) removed from the inscriptions they wish to examine and read.  This is not to deny the materiality of the digital, or of computation itself, but certainly makes the study of such matter and practices much more difficult than the claims to visibility that Latour presents. It also suggests that Latour's rejection of the abstraction in and of computation that electronic circuitry makes possible is highly problematic and ultimately flawed.


[1] Accepting the well-designed look of the website that contains the AIME project, there can be no disputing the fact that the user experience is shockingly bad. Not only is the layout of the web version of the book completely unintuitive but the process of finding information is clumsy and annoying to use. One can detect the faint glimmer of a network ontology guiding the design of the website, an ontology that has been forced onto the usage of the text rather than organically emerging from use, indeed the philosophical inquiry appears to have influenced the design in unproductive ways. Latour himself notes: "although I have learned from studying technological projects that innovating on all fronts at once is a recipe for failure, here we are determined to explore innovations in method, concept, style, and content simultaneously" (Latour 2013: xx). I have to say that unfortunately I do think that there is something rather odd about the interface that means that the recipe has been unsuccessful. In any case, it is faster and easier to negotiate the book via a PDF file than through the web interface, or certainly it is better to keep ready to hand the PDF or the paper copy when waiting for the website to slowly grind back into life. 
[2] See also, Latour stating: "the digital only adds a little speed to [connectivity]. But that is small compared to talks, prints or writing. The difficulty with computer development is to respect the little innovation there is, without making too much out of it. We add a little spirit to this thing when we use words like universal, unmediated or global. But if way say that, in order to make visible a collective of 5 to 10 billion people, in the long history of immutable mobiles, the byte conversion is adding a little speed, which favours certain connections more than others, than this seems a reasonable statement" (Lovink and Schultz 1997).
[3] The irony of Latour (2014) revealing the close reading practices of actor-network theory as a replacement for the close reading practices of the humanities/digital humanities is interesting (see Berry 2011). Particularly in relation to his continual reference to the question of distant reading within the digital humanities and his admission that actor-network theory offers little by way of distant reading methods. Latour (2010b) explains "under André Malet’s guidance, I discovered biblical exegesis, which had the effect of forcing me to renew my Catholic training, but, more importantly, which put me for the first time in contact with what came to be called a network of translations – something that was to have decisive influence on my thinking... Hence, my fascination for the literary aspects of science, for the visualizing tools, for the collective work of interpretation around barely distinguishable traces, for what I called inscriptions. Here too, exactly as in the work of biblical exegesis, truth could be obtained not by decreasing the number of intermediary steps, but by increasing the number of mediations" (Latour 2010b: 600-601, emphasis removed).


Berry, D. M. (2011) Understanding Digital Humanities, Basingstoke: Palgrave Macmillan.

Cantwell Smith, B. (2003). Digital Abstraction and Concrete Reality. In Impressiones, Calcografia Nacional, Madrid.

Latour, B. (2010a) The migration of the aura or how to explore the original through its fac similes, in Bartscherer, T. (ed.) Switching Codes, University of Chicago Press.

Latour, B. (2010b) Coming out as a philosopher, Social Studies of Science, 40(4) 599–608.

Latour, B (2013) An inquiry into modes of existence : an anthropology of the moderns, Harvard University Press.

Latour, B. (2014) Opening Plenary, Digital Humanities 2014 (DH2014), available from http://dh2014.org/videos/opening-night-bruno-latour/

Lovink, G. and Schultz, P. (1997) There is no information, only transformation: An Interview with Bruno Latour, available from http://thing.desk.nl/bilwet/Geert/Workspace/LATOUR.INT

Tarde, G. (1903/1962) The Laws of Imitation, New York, Henry Holt and Company

Disqus for Stunlaw: A critical review of politics, arts and technology