Big Data and Governance

Maude Bonenfant, Marc Mnard,

Andr Mondoux and Maxime Ouellet
Abstract
For several years now, the media, the business world, and information technology
(IT) have often used "Big Data" to describe a new society-wide dynamic It is
characteri!ed not only by the "roduction of massive amounts of data, but also#
and es"ecially#by the huge "otential benefits that new statistical data-analysis
tools would confer The "roliferation of data is so e$tensive that data ca"ture and
analysis are increasingly "resented as e$ceeding human reach, thus necessitating
the use of tools and IT methods for inter"retation Data e$traction and analysis are
defined as "data mining," without "resenting data "roduction, access, and analysis
as socially constructed (eg, with ideological, "olitical, economic dimensions)
Instead, they constitute a means for deriving "natural" information (the %eal) In
this light, Big Data may be seen as a techni&ue that dis"enses with symbolic
mediation and thus lies outside the field of "olitico-ideological debate 'ur
"resentation deals with the "otential conse&uences of models based on Big Data
Big Data modelling relies, in "art, on the "roduction, diffusion, and use of
individual-s"ecific data (eg, recommender systems, data-collection-based
"rofiling) It is "rimarily used in conte$ts where future actions must be forecast
from "resent behaviours (et insofar as this )ind of modelling infers social
re"resentations from claims of direct access to the %eal (ie, the non-ideological
and non-"olitical), one must wonder whether it is not hel"ing create a governance
dynamic that lin)s individuals not to a set of views (eg, a "olitical choice) but
rather#because necessary and inevitable#to the %eal itself In other words, Big
Data could be a way of (re)"roducing the social by "reventing the sudden
emergence of the "olitical#that is, a means for assuring control by arranging for
social re"resentations of the %eal
Key Words: Big Data, governance, technology, social media, "olitical economy,
ideology
*****
This wor)ing "a"er comes out of a recently created research grou" that focuses
on information and surveillance in daily life
*
The grou" has begun a research
Big Data and Governance
__________________________________________________________________
"ro+ect on Big Data and governance Its first "hase ("roblematisation) is what we
shall be "resenting at this conference
The emergence of social media has brought with it a significant increase in the
"roduction, circulation and consum"tion of data ,lthough this wealth of
commercial and financial data is not itself new, the "roliferation of -social- data
(eg, values, o"inions, +udgements) does constitute something different Ta)ing
both this "henomenon and the law of large numbers into account, some believe that
such massive &uantities of data -have something to say- and "resent a coherent
whole in the image of the social issues themselves
The necessity of dealing with the %eal has given rise to two main "hiloso"hical
"ositions The hermetic tradition "ostulates that the %eal is not in itself accessible,
that it always re"resented (ie, built on a foundation of symbolic mediation) In
contrast, the "ositivist tradition asserts that e$"eriential data ma)e it "ossible to
)now the %eal directly Big Data, as a collection of com"uterised techni&ues for
analysing data, o"erates on the "remise that these "ractices ma)e it "ossible to
e$"ose the %eal com"letely In this light, however, the "roliferation of -social- data
"oses two ma+or challenges. *) how to ma)e sense out of data that in essence are
"olitically and ideologically loaded because socially generated, and /) how to
ma)e these data meaningful given that they are "roduced on such a massive scale
that they are considered -too big to )now-
0athematical information theory allows res"onding to the first challenge First,
the theory ma)es it "ossible to neutralise "olitical and ideological loads by
transforming the symbolic (ie, information in its "rimary sense of informing 1 a
re"resentational "rocess) into data, that is, an 2ob+ectivised3 and 2real3 form
4econd, by its very transformation into &uantifiable data, information becomes
sub+ect to collection, storage, and management The second challenge, how vast
&uantities of data might be harnessed, is in turn met by the very means of "resent-
day data "roduction. data "rocessing as tool for automating calculation
This is not without im"lications for the dynamics of "ower relationshi"s 5ith
the symbolic and its "olitical and ideological loads being, as it were, neutralised,
the legitimacy of "ower actually rests less on the "olitical than on the "rocess
relationshi"s of control and management that result from Technic
/
To "ara"hrase
4aint-4imon, who used to refer to the transition from governing "eo"le to
administering things, we hy"othesise that we are about to move from governing
"eo"le to administering "eo"le who have become things (ie, the -data double- as
the digital version of an individual6 7yon, /889)
1. Big Data
Big Data is the ability to lin) several databases and, using advanced correlation
methodologies, identify "atterns of information that would otherwise remain
invisible (Bollier, /8*8)
/
M. Bonenfant, M. Mnard, A. Mondoux and M. Ouellet
__________________________________________________________________
From the first su"ercom"uters in the "ost-war "eriod to business intelligence
systems designed to analyse commercial databases, the idea of analysing data
using data "rocessing is not itself new Big Data continues this lineage by ta)ing it
in new directions 5here Big Data differs is, first, its e$"licit design to incor"orate
all data (a "oint we shall return to) since its intention is to add traditional business
data to its com"ilations of data from social media In so doing, Big Data throws its
hat into the -social- arena, with a s"ecial debt to -sentiment analysis-, which uses
com"uter-based linguistic and literary analysis techni&ues to identify and e$tract
sub+ective data (:ang and 7ee, /88;) 4econd, Big Data is "art of a "roactive
a""roach motivated by the e$"licit "ur"ose of finding correlations that suggest
future trends (eg, credit card fraud, targeted advertising, su""ly and demand,
swings in financial mar)ets) Third and last, though once the e$clusive domain of
business, Big Data sets itself a"art by lending itself to the "ur"oses of the 4tate
(eg, surveillance, national defence, territorial integrity)
4ustained by data from the social media, Big Data claims in fact to have
"ertinent a""lications directed at the social commons, thus hel"ing to transform the
vocation of the body "olitic into governance. the management of o"erating
"rocedures for a society that has been converted into data <ot only is Big Data
used more and more often in, for e$am"le, "olitical cam"aigns (eg, surveys of
voter intentions) and "ublic-interest decision-ma)ing (eg, health, education), but
it may also move directly from the descri"tive to the "rescri"tive when its tools for
forecasting and ris) management are a""lied to the field of -security-
,s a model, Big Data has two main features, both arising from its
e"istemological "osition, as noted above Because Big Data bills itself as neutral
with a direct line to the %eal, it aims to re"resent the totality (ie, an omni"resent
%eal) and vaunts the immediacy (real time) of its )nowledge
Big Data intends to incor"orate and analyse all data, whether traditional data
structured in databases, metadata "roduced by analysing these data, or unstructured
data from social media (eg, email, te$t messages, "osts on Faceboo) and Twitter)
The notion of the %eal is what sets the whole dynamic in motion, as attests the
well-)nown ,merican initiative Total Information ,wareness TI, sought to
integrate information technologies in order to aggregate data, analy!e lin)s and
develo" descri"tive and "rescri"tive models by using data mining It would then
a""ly these models to other data sources to identify terrorists and terrorist grou"s
Because it was controversial, the "rogram-s funding was cut in /88= 7ater,
however, it was revived than)s to a secret "rogram of the <ational 4ecurity
,gency that some describe as a veritable modern-day 0anhattan :ro+ect This
"rogram is thought to have established listening "osts throughout the country that
interce"t and analyse billions of emails and tele"hone calls to and from the >nited
4tates
=

:resented as according with the %eal, Big Data lends itself well to data mining,
as though data were a natural resource or raw material (ie, without symbolic
=
Big Data and Governance
__________________________________________________________________
mediation) e$tracted from the %eal If it is true that, from a ?eideggerian "oint of
view, Technic ta)es "art in the unveiling of the %eal, the act of desymbolising rests
on what ?eidegger called the anthro"ological-instrumental re"resentation of
Technic and consists in seeing this latter solely as a neutral tool sub+ect to the will
of the user 5hen measured against Big Data, all instances of technical data
"roduction are viewed as -natural-, that is, as corres"onding to a %eal that "resents
itself as unveiled in and of itself by Technic
,lthough the act of unveiling the world may also signify its com"letion, the
world is not a common s"ace to be a""ro"riated (oikoumene). it is not a to-be6 it
sim"ly i It is no longer the fruit of "ra$is as dialectical relationshi" between
sub+ect and world6 it no longer lays down a tem"oral and dialectical "rocess of
"sychic and collective individuation (4imondon, /889) This is why Big Data is, at
heart, a dynamic whose essence is mostly driven by instantaneity, the time
described as real Indeed, Big Data insists on obtaining the very latet data The
Big Data ideal is thus to be continually fed by the no!, that is, to be incor"orated
into the very flows of "roduction, diffusion and consum"tion of data. -This
lifestream 1 a heterogeneous, content-searchable, real-time messaging stream 1
arrived in the form of blog "osts and %44 feeds, Twitter and other chatstreams, and
Faceboo) walls and timelines-
@

That this real-time dynamic characteri!es Big Data is no accident In fact, it is
in the nature of real time that meaning may a""ear finali!ed and e$em"t from
diachronic symbolic mediation It is in the nature of real time that "reventive action
(as in ris) management) may have ma$imum effectiveness as it moves from
uncertainty to ris)s that have "robabilities, which then become &uantifiable
Finally, it is only in real time that the "rimacy of "rocessual relationshi"s 1 the
wor)ings of which themselves become the teleology of the "rocess 1 may foster a
sustained ecli"se of symbolic mediation In this res"ect the Big Data "henomenon
conforms to <egri-s characterisation.
0ediation is absorbed within the "roductive machine This is a form
of legitimation that rests on nothing outside itself and is re"ro"osed
ceaselessly by develo"ing its own language of self-validation
A

?ere we are well into ?eideggerian territory. ?umans use Technic while
believing themselves its absolute masters (the anthro"ological-instrumental
re"resentation), which misleads them into a state of vulnerability since, as
?eidegger "ointed out, modern Technic also embodies the "ower of "ower ,s a
conse&uence, the sub+ectBDasein confuses its own destiny with that of Technic In
the Big Data universe, symbolic "roduction (eg, of ideas, feelings, values)
actually becomes a flow of data "roduced by Deleu!ian -dividuals-, which
?or)heimer had indeed antici"ated.
@
M. Bonenfant, M. Mnard, A. Mondoux and M. Ouellet
__________________________________________________________________
,dvances in technical facilities for enlightenment is accom"anied by a
"rocess of dehumani!ation Thus "rogress threatens to nullify the very
goal it is su""osed to reali!e. the idea of man
C
2. A non-ideological ideology
Drounded as it is in immediacy and totality, Big Data has assumed a role in
(re)"roducing the 4ystem-5orld (0ondou$, 7acroi$, /88E). it ma)es the
ontological claims of embodying the world itself sim"ly by the -mediation- (defined
as neutral act) of Technic :erfunctory a""roaches to the %eal are not new, as can
be seen in traditional em"irical-"ositivist e"istemological "ositions (et such
"ositions were once rooted in more com"rehensive ideological systems (eg,
"rogress, civilisation) <ow, by contrast, they are self-referential. access to the
%eal has become the goal itself6 the means have become the ends The emergence
of an ideology (a "osition ado"ted to deal with the %eal) "resenting itself as non-
ideological (ie, mere technical mediation) followed, an ideology that 7abelle,
building on the wor) of 7efort, describes as -invisible- There
the %eal no longer refers to anything but itself and is no longer measured
against a higher standard from which it derives meaning6 as a rule, it
s"ea)s for, about and by itself
9

Big Data thus is ta)ing "art in the "resent-day dynamics that obscure the
domains of "olitics and ideology and their ca"acities as meaning-generating,
mediating agents lin)ed to the %eal
(et, and this is a criticism that might be levelled at ?eidegger (4tiegler, /88C),
one must not overloo) the social commons as an essential dimension of being-in-
the-world, nor that readiness-to-hand also im"lies a "ra$is as well as collective
suitability ,s 7efebvre "ointed out, the (su""osed) end of history nevertheless
remains a historical "henomenon 4o, too, the claim to being non-ideological
(Technic) remains an ideology The lineage of the e"istemological "osition of Big
Data has come down to us via, among other things, the ideology of cybernetics
Flaborated around the central idea of homeostasis, cybernetics "resu""oses what
4imondon called the -right sha"e-, that is, the sha"e at the outset, the sha"e
homeostasis must re"roduce By dis"ensing with the actual genesis of -right sha"e-,
the before"and, the moment when all -right sha"es- are "ossible and thus could
have cleared the way for a "olitical-ideological "luralism a""ro"riate to the
symbolic, cybernetic e"istemology manages to raise "rocessual relationshi"s to the
ran) of endB"ur"ose <ot only is the %eal no longer sub+ect to ("olitical-
ideological) symbolic mediation, but the "olitical and the ideological are also set
u" as hindrances to fully o"erational "rocessual relationshi"s (Technic) that lie at
the root of the %eal
A
Big Data and Governance
__________________________________________________________________
?ere we again encounter the ideology of neo-liberalism. once freed from the
bonds of the "olitical (individual freedom having become the only real dimension),
the mar)et will naturally o"erate at full ca"acity 4et forward as the world itself,
neo-liberal ideology becomes totali!ing. the world sto"s being an oikoumene (a
world to be collectively "roduced) in order to embody the %eal itself In this sense,
as Foessel maintains, we are living in the "ost-end-of-the-world.
The end of the world has become the hori!on of the "resent because
technical ob+ects, instead of humans, now define what should be done
The image of the end ratifies the considerable contraction of s"ace for
o"tions that arose after the a""earance of a universe of technical
instruments that insinuated themselves between humans and nature
,ccording to this "oint of view, the end of the world does not designate
so much an event to come that must be delayed as a way of describing
what the "resent loo)s li)e It has already ta)en "lace 5e are living after
the world has ended
;

In accordance with communicative ca"italism (Dean, /88E), Big Data ta)es
"art in "roducing totality by transforming the world into data that may be held
simultaneously and that is storable, manageable and as a result -commodifiable- By
advocating ris) management by antici"ating trends, Big Data brings bac) the idea
of the %eal as mac"ina, an entity having as its goal to de"loy ma$imum and
o"timal "roduction :roducing in this sense is sub+ect to the im"eratives of a
ca"italism that now reifies itself as a category of the %eal (the economy) Thus, as
Gamieson observed, it is easier to imagine the end of the world than the end of
ca"italism
,s discussed above, Big Data is "art of a dynamic in which time accelerates,
thus s"reading the idea that the %eal is essentially a #roce, and to access it one
must sei!e it while it is evolving Hounter to what one might thin), this "rimacy of
real time is not a "roduct of our technological times 0ar$ long ago noted this
"ro"ensity as one of ca"italism-s significant characteristics.
Ha"ital by its nature drives beyond every s"atial barrier Thus the creation
of the "hysical conditions of e$change 1 of the means of communication
and trans"ort 1 the annihilation of s"ace by time 1 becomes an
e$traordinary necessity for it Ha"ital thus sets about "roducing the
means of communication and trans"ortation
E

The more ca"ital is develo"ed, the vaster the mar)et in which it circulates6
but the greater the s"atial tra+ectory of its circulation, the more it will tend
to e$"and mar)et s"ace and thusI with timeI to destroy s"ace
*8

C
M. Bonenfant, M. Mnard, A. Mondoux and M. Ouellet
__________________________________________________________________
3. Big Data and governance
Fver since Foucault (Foucault, *E9A), the idea that socio-historical orders are
based on "ower relationshi"s in which sub+ects disci"line themselves (as in a
"ano"ticon) is more or less ta)en for granted (et o"inions diverge as to whether
we are in a disci"line- or control-oriented society Han one still really s"ea) of
disci"line when this order, to borrow 7efort-s e$"ression (7efort, *E;C), can no
longer withstand the -trial of its institution-, that is, when the order is no longer
re"resented as "olitical-ideologicalJ ,re we still disci"line-driven when the real-
time im"erative, what Fdwards called cybernetic behaviourism (Fdwards, *EEC),
wor)s overtly to transform adherence to ethics into an im"ulse-based dynamicJ
Hloser to home, %osa has found the same thing.
0odern societies are regulated, coordinated and dominated by a strict and
rigorous time system that is not structured in ethical terms 0odern
sub+ects may therefore be described as being only minimally restricted by
rules and ethical "enalties, and therefore as being free, whereas they are
ruled, dominated and re"ressed by a time system that is largely invisible,
de"olitici!ed, undis"uted, under-theorised and inarticulate
**

Hontrary to the em"owerment heralded by the social media that feed Big Data,
a form of coercive "ower is actually at wor). -The time most valued is the
"er"etual "resent, which serves as norm and has the distinctive feature of being
coercively im"erative for its agents-
*/
This dynamic is strengthened all the more by its obvious and -natural- harmony
with the %eal In other words, integration with the world now occurs because it is
im"ossible to be outside the world F$teriority undermines any ontological
"retension to being the %eal or the world itself This is why any sort of otherness
ultimately not only "ortends the return of the "olitical and the ideological, but it
also assumes an a""earance of Fvil (Dean, /88E) since it calls into &uestion the
-world- itself It is thus necessary to )ee" the sub+ect continuously inside the
-world- The Big Data dynamic advances this cause by re&uiring the ceaseless
"roduction of individuals who are continuously -connected- (,',H. ,lways 'n,
,lways Honnected) via devices that feature the automatic storage of information
trac)ing usage Big Data may thus be seen as governance by management of
individual user information
,s a technical tool that may be considered at once an instance of both ontology
and mediation, the Big Data "henomenon becomes a self-actuali!ing dynamic. ,ll
that is necessary is to "ro+ect itself as %eal in order to model behaviours that as a
result find confirmation for themselves in Big Data-s ontological status This self-
referentiality is what allows Big Data to (a""ear to) do without symbolic mediation
and to be at once a "rocess of both "roduction and control ,s such, the tools for
"roducing the social (Faceboo) describes itself as a -social utility-) are the same
9
Big Data and Governance
__________________________________________________________________
ones that also de"loy as dynamics of surveillance and control, rendered ordinary by
being seen as the very means for +oining and "artici"ating in the social commons
5ith Big Data we have moved from controlling "roduction to "roducing control
Notes
Bibliograpy
Bamford, Games -The <4, Is Building the HountryKs Biggest 4"y Henter 5atch
5hat (ou 4ay- In $ired, 5iredcom, 0arch *A, /8*/ ,ccessed February =,
/8*=
Dean, Godi Democracy and Ot"er %eoliberal &antaie <ew (or). Du)e
>niversity :ress Boo)s, /88E
Fdwards, :aul < '"e (loed $orld Hambridge. 0IT :ress, *EEC
Fischbach, Franc) )a #rivation de monde. 'em#, e#ace et ca#ital :aris.
7ibrairie "hiloso"hi&ue G Lrin, /8**
Foessel, 0ichaMl A#r* la fin du monde. (riti+ue de la raion a#ocaly#ti+ue
:aris. 4euil, /8*/
Foucault, 0ichel ,urveiller et #unir :aris. Dallimard, *E9A
Delernter, David -The Fnd of the 5eb, 4earch, and Hom"uter as 5e Nnow It- In
$ired, 5iredcom, February *, /8*= ,ccessed February =, /8*=
?or)heimer, 0a$ and ,dorno Theodor 5 )a dialecti+ue de la raion :aris.
Dallimard, *E9@
7abelle, Dilles OFssai sur lKidPologie dominante de la dPmocratie avancPe
(ou . Q &uoi sert le diffPrentialisme J)K In -luralime et dmocratie. .ntre culture,
droit et #oliti+ue, edited by Libert 4tP"hane Libert, *8=-*/C 0ontrPal. RuPbec
,mPri&ue, /889
;
M. Bonenfant, M. Mnard, A. Mondoux and M. Ouellet
__________________________________________________________________
7efort, Hlaude .ai ur le #oliti+ue :aris. :oints, *E;C
7yon, David ,urveillance ,tudie/ An Overvie! <ew (or). :olity, /889
0ar$, Narl Grundrie. &ondement de la criti+ue de l0conomie #oliti+ue, vol. 1
:aris. Sditions ,nthro"os, *EC;
111, -F$change of labour for labour rests on the wor)er-s "ro"ertylessness-,
Grundrie, <oteboo) L, *;A;, A/A
htt".BBwwwmar$istsorgBarchiveBmar$Bwor)sB*;A9BgrundrisseBch*8htmT"A/A
,ccessed ; 0arch /8*=
0ondou$, ,ndrP and 7acroi$, Gean-Duy OFausse et vraie diversitP culturelle dans
le systUme-mondeK In David contre Goliat"/ )a (onvention ur la #rotection et la
#romotion de la diverit de ex#reion culturelle de l02%.,(O, edited by
ThPorVt, (ves, *A9-*98 0ontrPal. Sditions ?urtubise ?0?, /88E
<egri, ,ntonio and ?ardt, 0ichael .m#ire :aris. *8B*;, /888
:ang, B and 7ee, 7 OFoundations and TrendsK In 3nformation 4etrieval, Lol /,
<os *1/, *1*=A, /88;
%osa, ?artmut Alination et acclration. 5er une t"orie criti+ue de la
modernit tardive :aris. 7a DPcouverte, /8*/
4imondon, Dilbert )0individuation #yc"i+ue et collective, :aris. ,ubier, /889
4tiegler, Bernard O7e thPWtre de lKindividuation DP"hasage et rPsolution che!
4imondon et ?eideggerK In 'ec"ni+ue, monde, individuation, edited by Laysse,
Gean-0arie, C*-;= ?ildesheim. Deorg 'lms Lerlag, /88C
E
*
D%I4R (Drou"e de recherche sur lKinformation et la surveillance au &uotidienB%esearch Drou" on Information and
4urveillance in Daily 7ife), "art of the D%IHI4 research centre of the >niversity of RuPbec at 0ontrPal (>R,0)
/
French scholars use the word -techni&ue- (singular) to refer to what Fnglish-s"ea)ing scholars call -technics- ("lural), ie,
the common denominator of technologies 5e have chosen to use the French term s"elled as a singular Fnglish noun
because, in our view, it is more suited to describe the -technical essence- behind technologies
=
Games Bamford, -The <4, Is Building the HountryKs Biggest 4"y Henter 5atch 5hat (ou 4ay- in ($ired, 5iredcom,
0arch *A, /8*/ ,ccessed February =, /8*=)
@
David Delernter, -The Fnd of the 5eb, 4earch, and Hom"uter as 5e Nnow It- in $ired, (5iredcom, February *, /8*=
,ccessed February =, /8*=)
A
,ntonio <egri and 0ichael ?ardt, .m#ire (:aris. *8B*;, /888), AE ,uthorsK translation
C
0a$ ?or)heimer and Theodore ,dorno, )a dialecti+ue de la raion (:aris. Dallimard, *E9@), = ,uthorsK translation
9
Dilles 7abelle, OFssai sur lKidPologie dominante de la dPmocratie avancPe
(ou . Q &uoi sert le diffPrentialisme J)K (In -luralime et dmocratie. .ntre culture, droit et #oliti+ue 0ontrPal. RuPbec
,mPri&ue, /889), *8E ,uthorsK translation
;
0ichaMl Foessel, A#r* la fin du monde. (riti+ue de la raion a#ocaly#ti+ue (:aris. 4euil, /8*/), */ ,uthorsK translation
E
Narl 0ar$, Grundrie. &ondement de la criti+ue de l0conomie #oliti+ue, vol. 1 (:aris. Sditions ,nthro"os, *EC;), AE
,uthorsK translation
*8
0ar$, Grundrie, =@ ,uthorsK translation
**
?ermut %osa, Alination et acclration. 5er une t"orie criti+ue de la modernit tardive (:aris. 7a DPcouverte, /8*/),
; ,uthorsK translation
*/
Franc) Fischbach, )a #rivation de monde. 'em#, e#ace et ca#ital (:aris. 7ibrairie "hiloso"hi&ue G Lrin, /8**), ;A
,uthorsK translation