The 20th century is the century of the industrialisation, the conservation and the transmission–that is, the selection–of memory. This industrialisation becomes concretized in the generalisation of the production of industrial temporal objects (phonograms, films, radio and television programs, etc)
[Bernard Stiegler, ‘The Time of Cinema. On the “New World” and “Cultural Exception”‘, p. 106.].
Following on from the post I made a few months ago, providing a version of a paper entitled ‘memory programmes‘ that I gave at the “Conditions of Mediation” conference at Birkbeck in June, I wanted to share some thoughts I have included in a recent submission around the theme of the ‘industrialisation of memory’, following Stiegler (as above), and the discourses of the ‘smart city’. I play around with the suggestion of the formation of a ‘municipal memory‘, which could be considered one of the cornerstones of what Nigel Thrift has diagnosed as “Lifeworld Inc.“.
Municipal memory
In the contemporary global North we increasingly retain information about ourselves and others using digital media. From supermarkets to banks and credit scoring agencies, much of our lives is being monitored and recorded through the proxy of our pecuniary activities. Networks of CCTV cameras, municipal transport ticketing and traffic management systems retain traces of our travel patterns. Furthermore we volunteer yet more information that is recorded by internet and media service providers, search engines and social media.
Thus many aspects of our lives are gathered and retained in databases and data stores variously owned by private and public organisations and institutions. This constitutes a growing system of memory–a memory of parts of life that might be otherwise forgotten or unthought. It also facilitates a kind of perpetual operation on memory, iteratively re-territorialising what is understood about our lives. The software programmes that drive digital media thus have significant agency in the various ways that we collectively communicate and remember. What is collected is not only retained but is also transmitted, folded into other data sets and utilised for civic and commercial purposes. In this section I want to focus on an ecosystem of mediated memory activities that is being constituted through networked digital technologies and the software that enables them to function.
We might think about this ecosystem of mediated memory activities in a number of ways. The capturing and collection of data to be retained constitute what Kitchin and Dodge have named ‘capta’, which are: ‘units that have been selected”¦ from the sum of all potential data”¦ [They are] what is selectively captured through measurement’ (Code/space, p. 261). We are variously enrolled, or enrol ourselves variously, into the systems in which ‘capta’ are constituted. Data are passively harvested and/or require an active contribution by us. For example, CCTV and traffic management systems can be seen as systems of passive collection because they do not require a form of action by those being surveilled (See Steve Graham’s “Software-sorted geographies“).
By contrast, social networking systems such as Facebook can, mostly, been seen as systems of contributive collection insofar as they require users to volunteer information to the system. There are of course varying degrees to which the systems for collecting and storing data about us for use in the future are passive or active within our everyday activities. ‘Capta’ collection activities feature both explicitly and implicitly in the activities of everyday life or perhaps what we might call attentive and inattentive collection of ‘capta’ by different kinds of mediated memory system.
The collection of ‘capta’ creates systems that blur the distinction between individual and collective. On the one hand, individual records of activity are kept, where they provide value, but on the other hand those individual records are always already a part of much larger aggregated systems that retain something like a collective memory of activity–the statistically derived average activities of people that can be variously grouped. For example, when we buy products through an online website that features a recommendation system, that system will use both the purchasing history of the individual and the average patterns of purchasing derived from all of the other consumers in the system in order to offer recommendations. The individual traces of activity, their retention outside of the more commonly discussed somatic memory, are, in this sense, never left alone, they are selected, aggregated and transformed to produce a kind of collective memory, something like a memory of habits.
In recent years there has been a growth in the popular discourse and business and government rhetoric concerning the governance and management of cities and conurbations through network technologies. With the decline in cost of the technological devices and thus the growth of installing networked ICTs into urban infrastructures and street furniture has come the proposition of the ‘smart city’.
The moniker ‘smart city’, as a form of discursive regime[1], freights a number of interconnected discourses that both justify and perpetuate particular understandings of what constitutes the ‘smart-ness’ of an urban environment, the socio-spatial conditions that can enable its development and who has the authority to name a city ‘smart’ . These discourses include: a commercial agenda for ‘Smart Cities’, aggressively marketed by corporations such as Cisco, IBM, Intel, Microsoft and Siemens, amongst others, to sell ‘intelligent’ automated systems to manage urban utilities and infrastructures [see Rob Kitchen’s ‘The Real-Time City? Big Data and Smart Urbanism‘]; the future oriented academic and industrial research agendas for the development of ambient intelligence and ubiquitous computing [see Anne Galloway’s ‘Intimations of everyday life‘; or my ‘Futures in the making‘], to enable ‘smarter’ ways of living ; and emerging systems of technocratic (urban) governance [see Steve Graham’s ‘Spaces of surveillant simulation‘].
The capability to create ‘smart cities’ has developed in parallel to, and somewhat in relation to, the development of technologies for the large scale collection, storage and processing of data [see Mike Batty’s ‘Smart Cities, Big Data‘]. Whereas, as Harvey Miller suggests [see ‘The data avalanche is here, shouldn’t we be digging?‘], much of what we know about cities has been derived from studies, set-up especially, that are characterised by data scarcity, the capacities of networked technologies embedded into the urban environment are marketed as a means of large-scale, ongoing data collection . Networked infrastructures are thus characterised as a means for ‘urban informatics’, a ‘smart’ and real-time system for governing cities. Alongside the civic and governmental applications for such technologies, a large number of commercial applications have been conceived for the monitoring of populations in order to target advertising. If the examples of loyalty cards and social media are largely concerned with the industrialisation of memory as a capture and retention of individual activities, then the emergence of city-scale ‘big data’ systems constitutes a different kind of industrial scale memory principally concerned with aggregated populations: something like a ‘municipal memory‘.
As Rob Kitchin notes [in ‘The Real-Time City? Big Data and Smart Urbanism‘], there has long been a production of large data sets for the purposes of governance as well as by companies for marketing, distribution and sales. However these data sets are not generate continuously, frequently rely upon sampling, and are often aggregated in broad geographical areas, and so function as coarse snaptshots . What constitutes contemporary ‘big’ data has not been clearly defined, but very broadly it can connote data sets that require bespoke systems for collection, processing and storage. Further, Kitchin has identified some of the definitions of ‘big data’ found in relevant literature as: very large in volume (petabytes of data); high in velocity (created in real-time); exhaustive in scope (striving to capture entire populations, or at least very large sample sizes); very detailed in resolution; and relational in nature, enabling the conjoining of different data sets. In addition to these factors, innovation in code and programming languages, and the ontologies or schema by which they are defined, have constituted ‘metadata’ formats–data that described data–to facilitate interoperability through translation and transformation between code types.
There are a number of modes of data collection that undergird the emergence of what is being called ‘smart’ city infrastructure, that illustrate the scales–of computation, data, labour, infrastructure and topography, amongst others–at which ‘urban informatics’ is required to operate. It is the combination of broadening of not only the resolution, scope and scales but also the speed of data collection that facilitates the institution of an emerging municipal memory. Kitchin articulates a typology of three sources for ‘big’ data that (can) drive ‘smart city’ systems: ‘automated’, where data is collected by software-driven systems–such as Automated Number Plate Recognition (ANPR); ‘directed’, where data is collected by systems that are directed by a skilled operator–such as CCTV; and ‘volunteered’, where data is provided by those previously surveilled, voluntarily, perhaps in exchange for value provided by a service based on those data–such as social media, e.g. Facebook, Foursquare and Twitter.
While systems for harvesting of capta that are ‘directed’ have a long history–we might for example include censuses in this category–the scale and sophistication with which contemporary automated and volunteered ‘big’ data collection is taking place is relatively novel. The big data sets have been made possible by the inclusion of techniques to embed individual and traceable identifiers on and into everything from public transit tickets (e.g. the London Oyster card) to cars (number plates), networked sensors distributed into the city to report back data (such as automatic meter reading) and a shift in culture that has brought a tolerance for making previously private activities conspicuous through personal technologies (such as location-based services like Foursquare). This combination of data collection systems has been envisioned as ‘constellations of instruments across many scales that are connected through multiple networks which provide continuous data regarding the movements of people and materials’ [quote from Batty et al. ‘Smart cities of the future‘] that, importantly here, is retained and stored constitutes a significant aspect of an industrialisation of memory.
The kinds of automated systems being built into the everyday urban environment are not limited to the provision of services or to the governance of public space but also include those specifically designed to put urban populations under surveillance for commercial purposes. There are systems where a form of surveillance can arise as a by-product: data are collected for the operation of a system that can then be repurposed by virtue of their metadata. For example, public transit systems, such as the London Underground, increasingly operate with individualised ‘smart’ cards, with a unique identifier encoded into them, for payment purposes, which also facilitates a form of tracking.
In addition to techniques for processing pre-existing data sets to infer further information, such as location, there are also technologies for which surveillance for marketing and sales purposes is the principal goal. A number of digital systems are being inserted into everyday spaces that attempt to calculate our bodily dispositions for commercial gain. Let’s explore two examples:
The company QuiVidi have created a system ‘automated audience measurement’ system, ‘VidiCube’, that purportedly can track the gaze of people near billboards using cameras to discern if and how long people look at adverts, and demographic distribution (specifically age and gender, with scope for further categorisations) of the ‘audience’. This technology is currently being implemented by the [Alan Sugar owned] British company, Amscreen, specialising in in-shop advertising screens and will be rolled out in forecourt and retail convenience stores, leisure centres and healthcare waiting areas.
The ‘Footpath‘ system created by the company Path Intelligence uses passive mobile phone signal detection to track movements through shopping centres in order to track consumer behaviour, which has been implemented in a number of shopping centres across the world.
Through the increasing number of systems that are being distributed over large topographical (urban) areas and potentially gathering and, importantly, storing very large quantities of capta, detailed representations of urban populations and spaces are made possible. Furthermore, the interoperability of data sets through sophisticated schemes of metadata makes ‘markets’ of data possible, and enables even greater detail. On the one hand, these ‘big’ data stores can be processed in order to feed back relevant information to a given population so that they can benefit, such as for planning travel. On the other hand, such detail also enables previously (relatively) anonymous spaces to be transformed into spaces of highly targeted advertising resulting from the combination of real-time data capture and statistical modelling of ‘big’ data sets.
What I have tentatively called ‘municipal memory’ then, is not only, or necessarily, large archives or databases of a huge numbers of individuals’ behaviours, being up-dated in real-time, but also, due to the scope, scale and interoperability involved, an emerging form of collective memory, in which the urban subject is simultaneously conceived of as both singular and multiple, individual and municipal.
Notes
1. In Discipline and Punish and later discussion of that text, Michel Foucault articulates ‘discursive regimes’ as disciplinary apparatuses for sustaining and reproducing, through processes of definition and exclusion, that facilitate and enact ‘procedures which allows the effects of power to circulate in a manner at once continuous, uninterupted, adapted and ‘individualised’ throughout the entire social body’ M. Foucault, ‘Truth and Power’, in Power/Knowledge. Selected Interviews and Other Writings, 1972-1977 trans. C. Gordon et al. (New York, Pantheon, 1980), p. 119. In relation to computation and software, see also: Kitchin and Dodge, Code/space, pp. 18-20.