Environment and Planning D: Society and Space 2007, volume 25, pages 1041 ^ 1060

DOI:10.1068/d77j

Subjecting cows to robots: farming technologies and the making

of animal subjects

Lewis Holloway
Department of Geography, University of Hull, Cottingham Road, Hull HU6 7RX, England;
e-mail: l.holloway@hull.ac.uk
Received 6 January 2006; in revised form 5 June 2006; published online 20 July 2007

Abstract. Recent representations of human ^ animal relationships in farming have tended to focus on
human experience, and to essentialise animal subjectivity in granting them a centred subjectivity akin
to that assumed to be possessed by humans. Instead, this paper develops an understanding of the
coproduction of domestic livestock animal subjectivities and the technologies used in farming domes-
tic livestock animals, based on an analysis of texts produced by agricultural scientists, farmers, and
equipment manufacturers relating to the effects of introducing new robotic milking technologies into
dairy farming. Drawing particularly on Foucault's conceptions of subjectivity and biopower, I explore
the emergence of particular forms of bovine subjectivity associated with robotic milking. Through an
analysis of a wide range of secondary sources, the paper shows that, although robotic technologies
have been presented as offering cows `freedom', better welfare, and a more `natural' experience, other
relations of domination come into effect in association with such technologies and their spatialities.
These are expressed through the manipulation of animal bodies and behaviours, in expectations
that cows move and act in particular ways, and through normalisation and individualisation processes.
I argue that nonhuman animal subjectivities in agriculture are thus heterogeneous, fluid, and con-
tingent on specific sets of relationships between animals, humans, and technologies and on specific
agricultural microgeographies. The paper ends by acknowledging that these relationships need further
empirical exploration in terms of both attempts to understand animals' changed experiences and ways
of being, and their ethical implications in particular situations.

Introduction: animal subjectivity and robot milking technologies

There has been a consistent, and generally accepted, call within the burgeoning `animal
geographies' literature that nonhuman animals should be accorded a status as subjects,
in opposition to tendencies to represent them in objectifying ways which suggest that
animals are simply `there' (Emel and Wolch, 1998; Emel et al, 2002). Yet, especially in
relation to farming, the question of what such a subjectivity might be is in most
instances underexamined. Instead, most writers have focused on the human dimension
of human ^ animal relationships (in exploring, for example, animals' symbolic and
economic value for people, or the human experience of particular farming situations)
and/or have tended to essentialise the subjectivity of farmed animals in ways which negate
the potential for them to become, to be coconstituted as they are entrained within various
and changing sets of socioeconomic, ecological, spatial, and technological relationships.
Other authors have attempted to do this in relation to `wild', `companion', and `working'
animals, for example (see Cox and Ashford, 1998; Haraway, 2003; Hinchliffe et al,
2005; Palmer, 2001; Whatmore, 2002; Whatmore and Thorne, 1998; 2000). Clearly,
trying to make sense of animal subjectivity is problematic, and might be approached
from a number of theoretical and empirical directions [see, for example, Pile and Thrift
(1995) and Pratt (2000) for reviews of approaches to human subjectivity]. However,
in this paper, I want to account for the emergence of particular forms of animal
subjectivity by focusing on one particular take on how subjectivity is produced in
and through the technologies and spaces involved in the ways specific farmed animals
are `kept'. I use as an example the changing technologies of dairy farming.
1042 L Holloway

The paper thus explores some of the complexities of what a domestic livestock
animal is in contemporary farming systems, especially those involving the engagement
of such animals, and their human associates, with robotic and information technol-
ogies. As Thrift (2005, page 201) argues, ``just as the materiality of technology has
become an insistent force in the world of animals, so the materiality of animals
has become an insistent force in the world of technology.'' In part, then, the paper
focuses on the material animal ^ technology relations evident in a particular farming
system, illustrating how, in material terms, dairy cows shape technology and, in
turn, what technology can do to dairy cows. In this sense, the paper is partly influ-
enced by science and technology studies accounts of the effects of hybrid relationships
between `living' and `nonliving'/technological actors and actants [see, for example Luke
(1997) and Michael (2000) and for a review of the coconstitution of users and technol-
ogies, see Oudshoorn and Pinch (2003)]. However, I also want to suggest that these
relationships effect emergent, and specific, bovine subjectivities. Going beyond animal
materiality in this way, the paper considers the constitution of animal subjectivities
in relation to technologies, spatial arrangements, and agricultural systems. (1)
In particular, I am interested in the ways in which particular bovine subjectivities
are produced in and through the specific set of technospatial configurations associated
with robotic milking technologies, or Automatic Milking Systems (AMS). These tech-
nologies milk cows automatically, without the immediate physical presence of humans,
contrasting with the proximity of human ^ animal relationships associated with con-
ventional milking parlours. Such technologies have implications for the design and use
of farm spaces and for the temporalities of dairy farming, as well as for dairy farm
human ^ animal relationships and for what cows are expected to do and be. As I argue,
it is these implications which afford a particular take on the animal subjectivities
associated with very specific contexts.
I begin by reviewing how farmed animals have been represented in different
literatures, before exploring some concepts inspired by Foucault's thinking on sub-
jectivity and power which lead to one possible way of thinking through animal
subjectivity. After this discussion, I return to cows and robotic milking, and begin
by outlining the recent emergence of robotic technologies for milking cows
and suggesting some of the immediate implications of these technologies for under-
standing animals and animal ^ human relationships. I then conduct an analysis of
secondary sources (including the proceedings of scientific conferences on robotic
milking, discussion of AMS in the farming press, and the promotional literature
of AMS manufacturers) to explore the constitution of bovine subjectivity in this
particular type of farming situation. This detailed textual analysis of scientific
and agricultural sources provides an important perspective on the ways in which,
through particular sets of human ^ animal ^ technology relationships, nonhuman
animal subjectivities are imagined, represented, and begin to be constituted.
However, it is a partial perspective, oriented towards a particular `agriculturalist'
set of human representations of animals and technologies. My conclusions are thus
in part concerned with identifying an empirical research agenda which would work
towards exploring other perspectives ö in particular, those of animals themselves and
those concerned with the situated ethics of particular human ^ animal ^ technology
relationships.

(1)
While this paper emphasises what might be referred to as `high-tech' livestock farming, it should
be pointed out that other `low-tech' farming technologies have had similar implications for animal
subjectivities, as Netz's (2004) study of barbed wire indicates.
Farming technologies and the making of animal subjects 1043

Representing farmed animals

Representations of the presence of farmed animals in recent accounts of the discourses
and practices of agriculture are increasingly numerous and protean. There are, too, rich
historical accounts of the (literal and metaphorical) social construction of animals
(eg Ritvo, 1987; Thomas, 1983). Summarising some of the approaches which have
been taken more recently, I here suggest three broad ways in which animals figure in
this work.
First, there is an important body of work which explores the objectification of
animals in a contemporary `animal ^ industrial complex' (Noske, 1997), focusing on
the rendering of animals as `things' associated with a rationalised, modernised, indus-
trialised food supply system (see, for example, Emel and Wolch, 1998; Franklin, 1999;
Wolch and Emel, 1995). Specific case studies explore this process in detail. For example,
Page (1997), Stibbe (2003), and Ufkes (1998) describe how pigs have been literally
restructured during efforts to produce lean bodies, and are discursively objectified
within agriculture; Boyd and Watts (1997) and Watts (2000) represent the bodies of
chickens in industrialised poultry production systems as sites for the accumulation
of capital; Stassart and Whatmore (2003) illustrate the enrolment of cattle into specific
marketing structures in ways which have effects on animal bodies; Grasseni (2005)
explores the combinations of knowledges which go into dairy cattle breeding; Holloway
(2005) points to the applications of `new' genetic knowledges to beef cattle breeding
in efforts to produce more marketable animal bodies; and many authors explore how
animals figure in the controversies and politics surrounding the unintended conse-
quences of `industrial' agriculture (eg Donaldson et al, 2002; Hinchliffe, 2001; Law,
2004; Woods, 1998). In this first type of account, then, animals figure mainly as
productive units, and although there may be a focus on the specific forms of human ^
animal relationship evident in the particular contexts, these are seemingly relationships
with nonhuman animal objects.
A second way in which farmed animals are represented focuses, in different ways,
on their symbolic value to people. Examples include the work by Evans and Yarwood
(1995) and Yarwood and Evans (1998) on `rare breeds' of livestock and on the `tradi-
tional' breeds associated with particular places, describing how they can be used to
signify locality, heritage, and quality in, for example, tourism or food marketing;
Quinn's (1993) and Smith's (1983) papers focusing on the symbolic value attached to
`pedigree' cattle in late-18th-century Britain and late-20th-century USA, respectively
(see also Walton, 1984; 1986; 1999); and Anderson's (2003) case study of the colonial
significance of cattle breeding and showing in Australia. In this literature, then,
animals are again broadly objectified, but in these cases they are things to which
symbolic significance is attached instead of, or as well as, being simply productive
units.
A third mode of representing farmed animals concentrates on the seemingly `close'
intersubjective human ^ animal relationships which are apparent in specific agricultural
circumstances. In this mode, farmed animals are understood to be something more than
either commodified things or the bearers of symbolic value. Nevertheless, work con-
ducted on specific human ^ animal relationships here has tended to be anthropocentric,
centring on the human experience of animals and on human practices even if they
attribute or imply an animal subjectivity. Gray (1998), for instance, illustrates how
particular knowledge and behaviour patterns exhibited by sheep are central to the
organisation of hill sheep farming, suggesting the significance of these animals
in embodying `consubstantial' relationships between people and farm holding.
Holloway (2001) and Wilkie (2005), on the other hand, focus on the `close' relation-
ships with animals experienced by different types of farmer, describing also how they
1044 L Holloway

enter into ethically problematic, seemingly intersubjective relationships with such

animals. Lastly, Convery et al's (2002; 2005) study of farmers' sense of bereavement
due to the slaughter of livestock in measures taken to contain foot and mouth disease
in the UK suggests that deeply felt and complex connections exist between many
farmers and their livestock. While valuable in making sense of the human experience
of human ^ animal relationships, these accounts are nevertheless constrained in their
approach to animal subjectivity, being more to do with defining the subjectivity and
emotional experience of the humans in their social and ethical relationships with
animals than with problematising the subjectivity of the animals themselves.
This constraint on understandings of animal subjectivity is also evident in the
recent concern with animal welfare in farming (eg Buller and Morris, 2003). Respond-
ing to wider debates in moral philosophy, society has increasingly regarded farmed
animals as having a status as morally considerable subjects as they are attributed with
characteristics such as sentience or sapience, meaning that they can, for example,
experience pain, stress, and boredom and can be attributed with rights which, for
some, give them a moral status equivalent to humans (see, for example, Midgley,
1983; Pluhar, 1995; Regan and Singer, 1976; see also the discussion in Emel and Wolch,
1998; Holloway, 2001; Philo and Wilbert, 2000). At issue here are two things. First,
many studies of farming practices assess welfare through the behaviour of the animals,
reducing their subjectivity to, first, their responses to the farming environment they
are in, and, second, a series of measurements which, in the last analysis, are mainly
related to the animals' agricultural productivity, assuming that, for instance, if animals
continue to grow or reproduce then welfare standards must be acceptable (see, for
example, Fraser, 2003; Webster, 2003). Second, representations of animals as morally
considerable, or as the bearers of `rights', risk attributing a particular, fixed subjectivity
to animals. That is, they essentialise what it is to be a subjectöaccepting a centred
subjective being rather than a continual process of becoming subject, and a hetero-
geneity of becoming which produces different subjectivities. Animal welfare discourse,
while undoubtedly important in thinking about the ethics of human ^ animal relations
in agriculture and in changing animal farming practices, is thus predicated on an
inherent, closed subjectivity, notwithstanding a history of domestication which makes
an understanding of these animals' existence outside of variable human ^ nonhuman
relationships impossible (Anderson, 1997).
What is so far lacking in these accounts is a more detailed consideration of what
farmed animal subjectivity might mean in particular contexts. This lack is mirrored to
a great degree in accounts of other kinds of nonhuman animal, although arguments
that `wild' or `companion' animals might `become' different in very different situations
have been explored by some [eg Whatmore and Thorne (1998; 2000), and Whatmore
(2002) on various `wild' species including leopards and elephants, and Haraway (2003),
and Palmer (2001), on dogs and cats, respectively]. In this context, Emel et al (2002),
reviewing `progress' in `animal geographies', suggest that the process of rethinking
human subjectivity which is, they argue, evident in the social sciences more widely,
should be extended to encompass reconceptualisations of animal subjectivity. Recon-
ceptualisations of human subjectivity have increasingly tended to be influenced
by poststructuralist and posthumanist perspectives which posit a decentring of sub-
jectivity, arguing that, instead of the existence of centred subjects being accepted,
subjectivities should be considered in a nonessentialist, emergent manner as frag-
mented, distributed, and becoming (eg Braun, 2004; Castree and Nash, 2006). This
perspective on decentring human subjects is taken up by Emel and Wolch (1998), who
argue that a parallel decentring of animal subjects permits a focus on the processes
which produce nonhuman subjectivities. The posthumanist discourse which thus
Farming technologies and the making of animal subjects 1045

works at decentring human subjectivity may, in this sense, become relevant to

conceptualisations of nonhuman subjectivity. Murdoch's (2004) argument that, ``where
humanism arrogantly places the human and human rationality at the centre of all
deliberations, posthumanism disaggregates, distributes, and dislodges the human sub-
ject. As Simon puts it, the subject `becomes local, fluid, contingent' (it is situated within
complex matrices of human and nonhuman relationships'' (pages 1356 ^ 1357, citing
Simon, 2003, page 4), would seem to encourage a case to be made for considerations of
nonhuman subjectivity alongside the nonessentialist human subjectivities which are
suggested.
This approach contrasts to an extent with those which have placed emphasis on
essences understood as inherent in animals and which seemingly give them a species-
specific and seemingly transhistorical subjectivity [see, for example, Jones (2003) and
Risan (2005) in relation to farmed animals]. Risan (2005), for instance, reports on
detailed observational research with dairy cows. He is critical of approaches such as
actor network theory which are agnostic in relation to the existence of distinctive
differences between `social' and `natural' entities, and argues from the perspective that
there are inherent, `natural' qualities of animal bodies to suggest that investigations
into, for example, technologies and animals must in essence be very different. Thus, for
example, a cow is defined as a `natural being' (page 787), which means that it must be
treated differently to a computer, a `cultural artefact' (page 787). Yet, this distinction is
problematic. The cow cannot be seen simply as `natural', even if we acknowledge that
her body, behaviour, and subjective experience are `bovine' in a way that retains some-
thing of a genealogical lineage that has become what we refer to as domestic cows.
The cow, first, emerges from histories of human intervention (for example, selective
breeding practices), and, second, exists in relation to the different and specific material
and social relationships cows are caught up in (ie different types of farming practice).
In this sense, the cow is a hybrid of the `natural' and the `social' (see Whatmore, 2000;
2002).
For Jones and Risan, the `levelling' or `symmetrical' strategies of writers such as
Callon (1986) act to negate the evident and real differences between entities and what
they can doötheir `affordances' or `capacities'. However, such strategies do not dis-
allow the recognition of the specific qualities which come to be possessed by entities.
The `symmetry' referred to relates more to nonacceptance of a priori categories than
to simply treating everything as if it were exactly the same. It does not mean that
we should not consider the specific human ^ nonhuman encounters involved in any
particular relationship, or that the ethics of such encounters and relations cannot be
discussed, but it is in this sense that the agency, subjectivity, and even bodily capacities
of an animal (for example) can be considered as the effects of sets of relationships
which have a history, rather than as essences simply `brought into' the establishment
of a relationship.
These emergent differences are clearly important and have effects. For example,
in designing milking equipment for cows, the size and conformation of their bodies,
along with their expected behaviours, need to be accounted for, and milking equip-
ment designed for cows will be different to that designed for other dairy animals such
as goats. But, those bodily and behavioural characteristics are not fixed: they have
varied in space and time (for example, as breeders have selected cows of a particular
size). The rest of the paper, then, following Braun's (2004, page 1354) comment that,
for posthumanists, ``the human has no essence, and never did'', draws on Foucault's
writing on subjectivity to construct a nonessentialist, `postbovine' understanding of
cattle subjectivities which suggests that there is no essential bovine nature, but that
instead contingent and fluid bovine subjectivities emerge in particular situations.
1046 L Holloway

Power, space, and subjectivity

Foucault's (1982) argument that his apparent focus on issues of power masked a real
emphasis on issues of subjectivity suggested that it was possible to understand how
subjects are `made' through three modalities which are actually processes of objectifica-
tion. In this sense, the individual becomes something which is subject to relations of
power, rather than something which simply possesses an inherent subjectivity. Foucault
argued, then, that subjects emerge as a result of the techniques of power which: first,
attempt to normalise, in accordance with scientific and other expert knowledges, what
is expected of subjects; then, second, focus on monitoring the extent to which individ-
uals adhere to or depart from what is `normal'; and, third, instil a continual, reflexive
self-discipline acting to enforce `normality'. As Emel and Wolch, (1998, page 18) thus
comment, ``Foucault, in particular, denied the transcendent or universal subject,
arguing that the subject is configured and reconfigured in the conjunction of discursive
and non-discursive practices.'' Foucault (eg 1975; 1979; 1982) emphasised that it
was within particular sorts of social institutions (such as schools, prisons, or hospitals),
with their specific microgeographies, that objectifying processes were at their most
intensive and could be most clearly identified, but that the creation of particular
subjectivities was also evident outside such institutions. In making this case, he argued
that objectifying processes, and the subjectification of individuals, was a process of
exercising power over bodies, so that embodied subjects found themselves more able or
less able to do particular things. These processes are necessarily context-specific,
relating to particular times and places, and associated with particular social structures,
technologies, knowledges, and so on.
Although evidently Foucault's work related to humans, some writers have begun to
suggest ways in which the ideas sketched above could apply to nonhuman animals, too
(eg Novek, 2005; Palmer, 2001; Williams, 1999), and in many instances it is possible
to make sense of human ^ animal relationships in contemporary agriculture in at least
the first two of Foucault's modalities, although the third is more problematic. For
example, technologies of monitoring are increasingly evident, and are related to tech-
niques of governing farmed animals as animals and groups. Expected performance
standards are developed (eg for milk volume, or weight gain) for particular populations
of animals, and animals are monitored to assess their individual performance against
these criteria. Palmer (2001) focuses on Foucault's conceptualisations of power as
something which circulates through, and constitutes, relationships between people. In
this understanding, power is not simply repressive, but is also productive with regard
to subjectivity. The creative construction of nonhuman subjectivity can be seen as an
effect of the power relations within which human and nonhuman animals are
enmeshed. Similarly, institutions, and their microgeographies, which structure the lives
(and deaths) of many animals in their associations with humans can be seen as key to
the emergence of contextually specific animal subjectivities, in the same way that
Foucault associated other institutions with human subjectivities. The agricultural archi-
tectures of farm buildings, etcöwhich can be seen as institutions which enclose farmed
animals and act on their bodies and behavioursöare powerful examples of this.
Going further, Palmer (2001) suggests that it is also possible to argue that animals
are able to interiorise aspects of those forces which act upon them, so that their being
as subjects is produced from those forces. As she argues, the fact ``[t]hat animal
behaviours can be affected by human actions; that animals interiorise elements of their
relationship with humans (and with one another) and that their relations to human
behaviours may be many and unpredictable are presupposed by commonplace inter-
actions which humans have with animals'' (page 349). This argument implies that it is
Farming technologies and the making of animal subjects 1047

at least possible to consider that Foucault's third mode of objectification which

produces subjects could apply to nonhuman animals.
Closely related to the comments above, Foucault (eg 1990 [1976]) uses the concept of
`biopower' to describe the entrainment of life itself into ``a densely-constituted field
of knowledge, power and technique'' (Best and Kellner, 1991, page 50). This biopower,
according to Foucault, has two dimensions. The first is concerned with the individual,
and is
``centred on the body as a machine: its disciplining, the optimisation of its capabilities,
the extortion of its forces, the parallel increase of its usefulness and its docility, its
integration into systems of efficient and economic controls'' (Foucault, 1990 [1976],
page 139).
The second concerns the regulation of populations rather than individuals, and
is referred to by Foucault as ``a biopolitics of the population'' (page 139). Foucault
describes how the emergence of biopower was characterised by ``an explosion of
numerous and diverse techniques for achieving the subjugation of bodies and the
control of populations, marking the beginning of an era of `biopower' '' (page 140).
This idea of biopower, in particular, is valuable in demonstrating how bodies and life
have been made understandable and rendered useful in the development of capitalism.
Foucault argues that an ascendant capitalist era required ``the controlled insertion
of bodies into the machinery of production and the adjustment of the phenomena of
population to economic processes'' (page 145).
Again, while Foucault is discussing biopower in relation to humans, there are
useful lessons here for an analysis of animals in capitalist agriculture. Clearly, through
the disciplines of agricultural science, the bodies and lives of farmed animals have been
manipulated in efforts to create productive, relatively docile entities. Crucially, the
disciplining of animals which this entails relies on the particular spatialities of farms
and farm buildings as they are designed to control animals' movements and activities.
Similarly, the monitoring and regulation of animal populations are consistently per-
formed, increasingly using computer databases and new genetic knowledges to record,
analyse, and work on aspects of animal productivity (Grasseni, 2005; Holloway, 2005;
Holloway and Morris, 2005). Yet, while these techniques might be seen as reproducing
animals simply as objects, biopower is also, for Foucault, productive of subjectivity.
Caught up in the relations of biopower, then, the internalisation of disciplinary
authority and particular knowledges about life means that individuals become subjects
through particular ways of understanding themselves and by behaving in particular
ways. If we take the example of cows and robotic milking, then, what farming does
to animal bodies, and what it makes them do with their bodies, is important in terms
of their subjectivities.
Rabinow and Rose (2003) acknowledge that little enough research has been
conducted on human subjects from the perspective of biopower as they define it;
``The three key elements that are brought together in the concept of biopowerö
knowledge of vital life processes, power relations that take humans as living beings
as their object, and the modes of subjectification through which subjects work
on themselves qua living beingsöas well as their multiple combinations remain
to be charted'' (page 24).
Applying this perspective to nonhuman animals is perhaps even more problematic,
with theoretical uncertainties surrounding animal intentionality, agency, and reflexivity
making it difficult to explore the element of how they might `work on themselves', even
if the elements of knowledge of key life processes, and power relations focusing on
living beings, are manifest in contemporary agriculture. However, returning to Palmer's
argument (2001), the power relations in which animals are embroiled frequently involve
1048 L Holloway

assumptions that there are emergent characteristics which can be understood as akin
to (but different from) human intentionality, agency, and reflexivity, and thus cannot
be reduced simply to capacities over mere things. What is of interest in this paper is,
in part, the ways in which such characteristics are anticipated and represented in
scientific and agricultural sources discussing AMS. So, considering how they might
be understood in animals, and how they might be related to particular agricultural
technologies, systems, and organisations of space, I turn in the following section to
examine the case of AMS.
After an introduction to AMS, I analyse how particular bovine subjectivities are
constituted through the ways in which dairy cows are represented in a range of texts by
focusing, first, on notions of cows' freedom and choice which are integral to AMS but
which become complicated by continued human control over individuals' movement
and the social relationships necessary to herd membership. Second, I focus on the
microgeographical structures and informational precision technologies of farms using
AMS, which act to implement individualisation and normalisation strategies. I show
how the implementation of AMS produces and expects particular bovine subjectivities.

Robotic milking and the production of bovine subjects

AMS technology: automating the dairy farm
Robotic or automatic milking parlours are a recent phenomenon on dairy farms, with
the first commercial installation in the Netherlands in 1992 (de Koning and
Rodenburg, 2004; Lind et al, 2000; Mottram and Masson, 2001). Their use is currently
limited to a relatively small number of farms, mainly in Western Europe and
North America, although the rate of uptake of the technology is rapid. Accordingly,
AMS manufacturers represent their technology as part of a continuing process of
agricultural `modernisation', emphasising its purported increases in productivity and
efficiency (Fullwood, 2005; Gascoigne Melotte, 2005; Lely, 2005; no date). De Koning
and Rodenburg (2004) estimate that there were around 2200 installations globally in
2004örising from an estimated 500 in 2000, and 1000 in 2001 (Mottram and Masson,
2001). These authors expect the number of installations to increase as the price of the
equipment falls, as the technology is argued to improve milk yields and animal welfare
(Lind et al, 2000). At present, however, only a very small proportion of farms use
AMS. For instance, it is estimated that there were 50 AMS in the UK in 2003 (Buss,
2004a), out of a total of around 23 600 dairy farms (DEFRA, 2004): that is, only 0.2%
of all dairy farms used AMS in that year. Despite this relative lack of significance
in terms of overall UK or global dairy farming, it is precisely because the emergence
of AMS (as a particular moment of innovation associated with a restructuring of
human ^ animal ^ technological relationships) has produced a great volume of debate
amongst people involved in developing and using a technology which purports to
dramatically change the lives of farmers and dairy cows that this example provides
interesting insights.
Robotic milking differs from conventional milking parlours in a number of key
ways, with important effects on the people and cows involved. First, rather than people
milking the cows altogether two or three times daily, robotic milking machines milk
cows individually, at any time, without direct human involvement or presence. Thus,
there may be only one milking unit per group of 60 or 70 cows, in the place of the
conventional `herringbone', `series', or `rotary' parlours in which cows are milked in
groups. Second, the robot can recognise each cow (by corresponding with collars or
tags), supply them with the correct portion of concentrate feed, attach the milking
cluster to the udder (using imaging technology and a `memory' of the placing of a cow's
teats), conduct the milking process, and release the cow from the unit, again without
Farming technologies and the making of animal subjects 1049

immediate human involvement. Third, as part of the milking process, the robot is also
involved in managing the health and productivity of the cows, including cleaning the udder
and monitoring each cow's milk quality and quantity individually. Maltz (2000, page 132)
refers to this as the application of precision technology to agriculture, suggesting that
``Automatic Milking Systems ... enable the management of every cow individually, thus
opening a new horizon for dairy management.'' While previous milking technologies can
also record individual milking `performance', Maltz suggests that management response
can also be conducted individually rather than at the herd scale using AMS.
For those farmers adopting this technology, robotic milking implies significant
changes to farms and farming practices. First, the spatial organisation of the farm
is changed. On most conventional UK dairy farms, cattle are milked as a herd, then
walk out to a field to graze, being `fetched' later in the day for the second milking.
They may actually thus cover an extensive area over the course of a grazing season,
although they are likely to be confined indoors for at least part of the winter. Robot-
ically milked cows move individually between feeding areas and the milking unit. Yet,
because it seems that being milked does not have high motivational attraction for
cows, there is also a need to keep them close enough to the milking unit to attract
them into it regularly (Lind et al, 2000). Many robotically milked herds thus experience
a regime of `zero grazing', being kept in buildings all year, with fodder daily cut from
the fields and brought to them. Second, and relatedly, farm temporality is radically
altered, shifting from the twice-daily rhythm of milking the cows as a herd to an
almost continuous flow of cows being milked individually. Together, these spatial and
temporal shifts imply a change in how cows move through space and time: in robotic
milking systems cows potentially have more autonomy of movement in both these
dimensions and can move as individuals as well as in a herd. Third, robotic milking
implies an important shift in human ^ animal relationships on dairy farms. Since the
actual milking is conducted by the robot, there is greatly reduced direct, physical
contact between humans and cows in this system, and a sense in which the human ^
cow relationship is mediated by the robotic technology. As a result, the human role is
redefined in robotic milking systems, with an emphasis placed on the stockperson's
need to enhance his or her skills in closely observing his or her cows at other times and
places, as there is not the usual opportunity to do this at milking time (Owen, 2003),
and to become proficient in the use of the large volumes of computerised information
which the robotic milking system collects and stores (Knight, 2001). While this
paper focuses on animal subjectivity, human subjectivity, in the sense of what it is to
be a stockperson, is similarly coproduced in association with changed agricultural
technologies (see, for example, Seabrook, 1992).
Autonomy and expectation
``The cows are actually feeding themselves outside the buildings at the moment, but
when they feel like it, and they think they want to milk themselves, they'll literally
get up, and go to the machine.''
Suffolk dairy farmer speaking on BBC Radio 4
The Food Programme 28 March 2004 (author's transcription)
Key to the constitution of bovine subjectivity in relation to AMS is the idea that cows
gain individual freedom. This is strongly asserted by AMS manufacturers, who argue
that this freedom has productivity and welfare benefits as cows are offered the ability to
follow what are presented as more `natural' routines of feeding, resting, and milking,
compared with those on `conventional' farms. For example, Gascoigne Melotte (2005)
suggests that, in its Zenith AMS, ``each cow defines her own rhythm of resting, milking
and eating, and these matters have a positive influence on the cow's life.'' Lely (2005)
1050 L Holloway

promotes the productivity of its Astronaut AMS by explaining that ``The reason
[for yield increases] is that the cow visits the robot, from its [sic] own free will, as
often as she likes. Milking more than twice a day comes closer to the cow's natural
needs in terms of feeding her calf. In addition, it enables the animal to get rid of undue
udder pressure on her own accord.'' This emphasis on freedom and choice, used by
manufacturers to promote AMS technology, is reiterated in the texts of agricultural
scientists. Owen (2003, page 15), for example, reporting on an experimental AMS in
Wales, writes that ``The cow can choose when she is milked ... . Cows visit the single
robotic stall voluntarily at any time, for food and for milking.'' Similarly, Wiktorsson
and SÖrensen (2004, page 371) argue that, ``compared with conventional dairy cow
management, cows in automatic milking systems are provided with more freedom to
choose their daily activities and rhythms.'' There is agreement that AMS produces cows
whose subjectivity is characterised by the effects of freedom and autonomy which, first,
are produced by their relationships with particular technologies and management systems,
and, second, differentiate them from cows in `conventional' systems, whose `freedom'
to follow `natural' behaviours is withheld. However, such freedom and autonomy are
complicated in at least two respects: first, they are associated with expectations regarding
cows' behaviour, and the monitoring and regulation of that behaviour; and, second,
individual cows' freedom and autonomyöfor example, in relation to making a choice
about when to be milked ö is complicated by the ways individuals are enmeshed in
social (herd) relationships. Each is explored in detail below. Both have implications
for the microgeographies of dairy farms and farm buildings, and thus for the
spatialities producing and being produced by bovine subjectivity in AMS.
First, then, cows are required to be able to learn quickly how to use AMS (de Koning
and Rodenburg, 2004; Owen, 2003) and to choose to regularly attend the robot to be
milked. In this sense, although cows milked using AMS are granted freedom of choice, it is
necessary that they make the right choices and exercise their freedom appropriately. Lind
et al (2000) thus suggest that human control over milking frequency and access to food
needs to be retained to some extent so that milking productivity is maintained. As
mentioned above, for many cows, being milked is insufficiently attractive to be worth a
long walk, and various management and technological processes have been devised in
order to ensure that cows do `choose' to visit the robot (eg Spo«rndly et al, 2004). As Millar
(2000a; 2000b) thus suggests, the autonomy granted to cows under AMS is rhetorical
rather than `real'. Further, cows which do not `fit' the AMS mode of farming (for example,
they do not learn how to use the system, or attend for milking too infrequently) are
removed (Owen, 2003). This sense of expectation, and the implication that cows choosing
`badly' will face controls of one sort or another, begins to emphasise that what is not
established by using AMS is simply the releasing of an inherent subjectivity which was
suppressed in `conventional' milking systems.
In this sense, Knight's (2001, page 53) comment is particularly significant. He
writes, ``To date, AMS has been viewed almost as a romantic technology, enabling
the cow to choose when and how often she wishes to be milked. The obvious flaw
with this approach is that if she chooses not to be milked there is little one can do to
force her, and it has quickly become apparent that motivation has to be encouraged.''
Although recognising the `romance' of the emancipated bovine subject, Knight under-
mines that romance with reference to the circumscription of cows' liberty, which is
often done by using various mechanisms to restrict their access to food until they have
been robotically milked. Owen thus describes how ``Some installations operate a one-
way system, forcing cows to visit the AMS to get feed or water on the other side''
(2003, page 16), and de Koning and Rodenburg (2004) emphasise the importance
of appropriate in-barn architectures to creating appropriate voluntary behaviour:
Farming technologies and the making of animal subjects 1051

``Since these systems depend on voluntary attendance, a well laid-out freestall barn is
essential to success ... . The routing in the barn should be according to the Eating ^
Lying ^ Milking principle'' (page 31). A focus of AMS research has thus been on the
design of the barn layout, involving the placing of the robot in relation to sources of
food and water, and the cows' movement and use of space. The designed microgeog-
raphy of robotic milking barns is a prominent theme in agricultural science discussion
of AMS, with several articles (eg Halachmi et al, 2000; Ketelaar-de Lauwere and
Ipema 2000) including plans for `model' layouts. Such models organise the internal
space of the AMS barn in such a way as to effect the particular behavioural routines
suggested above. The cows' experience of movement, rhythm, and routine in AMS
barns will differ significantly from those in `conventional' milking systems. According
to Sharp, discussing an AMS installation in Aberdeenshire, Scotland, ``the building
layout, areas and cow movement within the building differ dramatically because of the
presence of robotic milkers. An individual-box robotic milking unit works on a one-
way system with a series of non-return gates between the lying, milking and feeding
areas. This means the cows have to go through the milking unit to get to the feed
stance and ensures they visit the milking unit on a regular basis'' (2003, page 88).
Architecture is thus used to structure cows' behaviour and coerce them into moving
in particular ways and according to particular rhythms. Paradoxically, then, the sub-
jective autonomy effected by the AMS requires a barn with an internal ordering
structure which controls that autonomy differently öby controlling the movements of
individual animals, and the flow and rhythm of the herd (which are thereby more struc-
tured and controlled than in a `conventional' system). The barn becomes more complex
as the robotic technology requires a means of controlling the autonomy which the cows
are attributed with in AMS systems. In effect, the barn is an important part of the
technology, imposing particular forms of discipline on bovine subjects and representing
an expression of biopower in the alignment of technology and spatial organisation with
cows' bodies (and the human need to achieve particular effects from those bodies) and
subjectivities. This arrangement is thus a coproduction of technology, spatial organisation,
body, and subjectivity. Rather than technology and layout simply drawing on or out pre-
existing animal capacities, there is an entrainment of technology, layout, body, and
subjectivity, with each involved in the coproduction of the others: technology and layout
are affected by cows' bodies and behaviours, and bodies and behaviours are affected by the
technology and the layout.
The structure and discipline associated with barn design are particularly important
as AMS is also associated with an increasing movement towards keeping dairy herds
indoors all year, rather than, as is the case in many areas, allowing them to graze in
fields for at least part of the year. Cows grazing in fields some way from their robotic
milkers have the capacity to resist their configuring as parts of productive technological
assemblages by refusing to voluntarily get up and walk from field to machines to be
milked. `Zero grazing' has been a common response. Yet, this in turn leads to problems:
walking is `good' for cows, maintaining healthy feet and bone structure. Cows grazing
in fields also contribute to a positive image of dairy farming, something increasingly
of concern to the farming community (Buss, 2004b; Ketelaar-de Lauwere and Ipema,
2000; Mathijs, 2000). Mathijs, in particular, is concerned with the effects of zero grazing
on public perceptions of dairy farming. As he writes, ``Generally, the introduction of
robotic milking will move the cow from the meadows to the barn'' (2000, page 246),
and he cites as evidence of public concern an article from the Dutch newspaper NRC
Handelsblad (van der Schans and van der Weijden, 2000) which urges that ``Koe moet
in de wei blijven'' (Cows must remain in the field). The issue here seems to be that
zero grazing changes the whole being and image of the cow, from being a `visible',
1052 L Holloway

grazing animal to being a housed one, invisible to the public. Aside from the effects
on public perceptions of dairy farming, this, coupled with the structuring of cows'
behaviour achieved through the microgeography of barn layouts, implies that bovine
subjectivities are particular to the combination of AMS and zero grazing.
The second complication is related to the herd context, which is key to under-
standing the constitution of animal subjectivity under different farming systems. As
well as the production of individual subjectivity, herd dynamics are effected by the
farming system. For example, the hierarchies within herds, with cows having `rank'
within their groups, are worked out in different ways according to the technologies and
spaces the herd is associated with. Hurnik (1992), for example, points out that cows
prefer to do things as a group. Arguing that AMS, which requires cows to be milked
one-by-one, can frustrate this desire to do things together, Hurnik suggests there can
be negative effects on the tolerance displayed by higher ranking cows towards lower
ranking ones as they effectively compete for access to the AMS (see Millar, 2000a).
Wiktorsson and SÖrensen's ethological study (2004) further illustrates this effect,
indicating that lower ranked cows have less choice than more highly ranked ones about
when to approach the robot to be milked, tending to either have to queue for long
periods or use `unsociable' times of the day or night. As they suggest, this might have
further effects on the microspatial organisation of the milking system. For example,
``it is desirable that the waiting area is constructed with a back door so cows can leave
the waiting area when it is full or when for any other reason they feel threatened.
Observation of a robotically-milked herd certainly shows that some cows are pushed
aside by others in the queue to be milked. A low-ranking cow might be trapped for
hours in a closed waiting area'' (page 374). Wiktorsson and SÖrensen are clearly
concerned about social differentiation within the herd, and show how that differen-
tiation has different effects in AMS than in `conventional' systemsöfor example, in
access to feed and milking, and in the differential experience and use of space of
individual cows. Thus, ``Instead of spending time in the eating area or in the alleys,
these [low-ranked] cows spent more time standing in the cubicles, suggesting that it
was a more relaxing environment. The low-ranked cows were found to be closer to
the milking station, especially during resting, indicating a need for them to monitor the
milking queue'' (page 376). Similarly, Ketelaar-de Lauwere et al (1996, page 199) argue
that ``It is concluded that the introduction of fully automated milking systems will
trigger effects of social dominance, especially concerning the timing of visits to the
AMS and the feeding gate, and the waiting of low-ranking cows in front of the AMS.''
Yet, this set of effects on the herd can be represented differently, particularly
by manufacturers wanting to portray AMS as a positive intervention in cows'
intersubjective experiences. Fullwood (2005) suggests, for example, that, under the
protection of the individual milking stall of an AMS, ``Bully cows can no longer
intimidate the quieter animals''. Lely (2005) positions the robot as a beneficent part
of the herd's environment, while recognising the social hierarchy and differentiation of
experience that occur within the herd:
``Scientific research has proved that separation of a cow from the herd is one of the most
stressful experiences any individual cow can encounter. The stall of the Lely milking
robot is designed and positioned in such a way that it forms an integrated part of the
herd habitat, and the cow is not separated from the herd during milking. It is also a
known fact that each cow has its own order of ranking within the herd. The lower the
ranking of the individual cow, the more stress she experiences when being confined to a
restricted space with the complete herd. Looking at the operation of the Astronaut
Robotic Milking System we see that each cow selects its own most convenient times
for milking, in line with the animal's ranking in the herd.''
Farming technologies and the making of animal subjects 1053

This discussion of the representation of enactment of bovine subjectivity in AMS

suggests a set of relationships between people, technologies, and cows (as a group and
as individuals) which is `biopowerful' in working on bodies and behaviours, and which
produces or effects particular animal subjectivities in conditions which can be repre-
sented as negative or positive in terms of cows' experience. In the next section I look in
more detail at the monitoring and regulation of life (as body and subjectivity) in AMS,
focusing, in particular, on the individualisation and normalisation processes associated
with AMS.
AMS institutions: biopower and subjectivity
It is clear from the agricultural scientific literature on robotic milking that cow suit-
ability in relation to AMS is an important concern. The focus of biopower on life itself
(Rabinow and Rose, 2003)öin terms of animals' bodily conformation and, increas-
ingly, their genetic `potential'öis thus apparent in restructuring dairy farming around
robotic milking technologies. While in the first instance AMS robots have been
designed around the bodies and behaviours of cows, the establishment of AMS on
farms, in turn, requires that cows become suited to AMS. Comments such as the
following support this process, and draw attention to the combination of bodily
character and behaviour upon which individual cows may be assessed as suitable or
unsuitable. Owen, for instance, writes that ``It is normally expected that approximately
5% ^ 10% of any herd are not suitable for AMS for various reasons i.e. temperament and
udder conformation'' (2003, page 16). More expansively, de Koning and Rodenburg
(2004) argue that, ``In terms of the impact on cows, the AM-system is not suitable for
all cows. Poor udder shape and teat position may make attachment difficult and some
cows may not be trainable to attend for milking voluntarily'' (page 29); successful
automatic milking requires ``Healthy cows with good feet and `aggressive' eating behav-
iour'' (page 31). They describe `problem cows' who need to be made to attend the AMS,
and whose milk yields may decrease in the system, unlike those who respond `well' and
whose yields increase. Artmann and Bohlsen (2000) put it in terms of cow `appro-
priateness', stating that ``All cows not suited to the automatic milking processes should
be replaced'' (page 223). In response to this requirement that animals `fit' the AMS,
rather than simply (de)selecting cows according to their (un)suitability, attention is
being paid to breeding future generations of cows which are suited to robotic milking.
Adjusting the genotype of animals through selective breeding and genetic appraisal is
thus an important strategy of an agricultural biopower, seeking to produce cows which
conform to technological requirements. Although not unproblematic for agricultural
science, both conformation and behaviour are seen as genetically malleable; Gravert,
for example, argues that ``Changing udder shape will not be a major problem, but
selection of cows reacting positively to the new system will take several generations''
(1992, page 395). For AMS to be effective, then, a process of normalisation is neces-
sary, with cows individually assessable against common criteria, and with strategies for
producing suitable future cows being implemented.
In the previous section, robotic milking systems were shown to be akin in some
ways to the institutions described by Foucault and to be associated with microgeog-
raphies and rules producing and expecting particular sorts of cow subjectivity in the
name of efficiency, rationality, and modernisation. The normalisation and individuali-
sation strategies associated with such institutions, which Foucault applied in his studies
of, for example, criminality or madness (1979; 2001 [1961]), are further evident in the
ways additional technologies associated with AMS make each cow both `a subject of '
and `subject to' institutional discipline. The AMS manufacturer Lely (no date), for
example, in common with other manufacturers, emphasises how its system ``gives an
1054 L Holloway

accurate up-to-the-minute knowledge of what is happening to the herd and to every

individual cow at all times'' (page 3). The agricultural scientific literature defines this in
terms of a shift, enabled by the introduction of AMS, from herd management to
monitoring and management at the level of the individual cow (Maltz, 2000). Highly
suggestive of the biopowerful implications of this move, it is stressed that maximal
animal `performance' will be achievable through the implementation of such new
management possibilities. For example, Knight (2001) writes that ``The technologies
used gather very large amounts of data as they operate, data that in the future will be
fed into integrated management systems which will ensure optimal performance from
individual animals'' (page 53). Thus, for example, AMS combined with automatic
weighing during milking might be responsible for obtaining ideal bodily conformation
from animals: ``if one adds body weight recording [food intake] can be adjusted to
achieve ideal body condition'' (page 53).
Mottram and Masson (2001) further explore the implications of this. They report,
for example, on the use of sensors within AMS which analyse the milk of individual
cows, and which detect ``deviations from optimal performance'' (page 77), thus allowing
management intervention. They further write that,
``The traditional method of managing cows is to use open loop control models (for
example to plan a ration scheme for an entire year or lactation) and to use manage-
ment by exception (i.e. the cow that performed below the target level) as the means
of adapting the model. The availability of smart systems will allow closed loop
control systems to be designed and implemented that will be based on models
that allow subtle deviations from predicted values of sub-clinical conditions to be
detected before homeostasis is asserted. These systems offer the ability of allowing
the cow to reach her genetic potential'' (page 82).
The suggestion here is that veterinary interventions could be made, on the basis of
suboptimal performance, in relation to medical or physical conditions which are not
evident from normal management and observation practices. Again emphasising the
possibility of realising a notion of genetic potential, the focus here is on interventions
in life (in terms of bodies, genes, and health/illness) which aim to achieve particular
effects in terms of bodily performance. In these terms, and following the understanding
of biopower described by Rabinow and Rose (2003), bovine subjectivities are produced
through these technological interventions in the sense that what a cow is, or has to
become, emerges from strategies, technologies, and knowledges for working on the
bodies and behaviours of living organisms.

Conclusions
I have argued that livestock animal subjectivities are effects of specific agricultural
systems, focusing on the recent development of a particular technological intervention
in the practices and knowledges of dairy farming and adopting a Foucauldian under-
standing of the production of subjectivities and of a biopower associated with strategies
of working on or with life. Processes of subjectification have been identified which:
first, are associated with the expectations placed on cows, via an idea of freedom
and choice, to establish particular relationships with a new technology; second, are
related to the particular spatialities and disciplines linked to the technology; third,
involve processes of normalisation and individualisation in pursuit of the production
of suitable bodies and behaviours; and, fourth, can be understood under the rubric of
biopower as acting on, producing, and controlling life.
The approach taken in the paper counters assumptions that, in terms of the ethics
of human ^ nonhuman animal relationships, simply acknowledging an animal subjec-
tivity potentially produces, or at least allows an argument for, the better treatment of
Farming technologies and the making of animal subjects 1055

animals in agricultural production. Instead, it examines how subjectivity is constituted

in particular circumstances, and is associated with concurrent processes of `becoming-
subject-to'. As a result, the paper also counters suggestions that allowing animals to be
`subjects' rather than `things' presupposes less intensive, smaller scale agricultural
systems. These arguments suggest that industrial agricultural systems deny animals
their subjectivity, and that, as technological, rational, capitalist aspects of farming
increase, animals are simply objectified. Instead, smaller scale or `alternative' agricul-
tural systems are more likely to promote an awareness of animals as sentient, sapient
individualsöthat is, subjects. However, in the context of the intensive, technological
systems presented in the case study of AMS, a subjectivity still emerges. In the case of
AMS, this subjectivity is associated with a particular spatiality and set of technological
relations, within a highly technological system designed to rationalise farming. It is
thus not possible to simply contrast the `domination' of animals in industrial farming
with a `freedom' granted by other modes of farming, or to simply say that (for example)
AMS are ethically `better' or `worse' for cows than other dairying systems. Instead,
particular farming systems produce varying and related effects of freedom and domi-
nation according to their use of particular technologies, spatialities, knowledges, and
so on.
Bovine subjectivity is thus contingent and heterogeneous. Two key lines of
differentiation are evident. First, there is a contemporaneous differentiation in
subjectification ö in terms of behaviours, experiences, expectations, bodily and behav-
ioural disciplines, and effects of biopowerö between cows in robotic milking systems
and those in `conventional' milking systems. Similar differentiations might be made
between dairy cows and other cows kept as suckler cows in beef production, `house
cows' on hobby farms, etc. Secondly, there is a historical differentiation: bovine
subjectivity has a history rather than an essence, and bovine being and bodily
capacities are relational in terms of the different technologies, economies, and social
relations (with humans and with other cows) that cows are associated with. Thus,
for example, the emergent subjectivity of robotically milked cows in the late 20th and
early 21st centuries differs from that of hand-milked cows in the 19th century, and so
on (and is clearly different from that of the `wild' progenitors of `domestic' livestock).
Cows have thus experienced different ways of being, have been and are differently
monitored, circumscribed, and related to by people, and have experienced different
spatialities, ways of moving, rhythms, routines, and herd ^ individual relationships.
In parallel with this, for those people involved in dairy farming, human subjectivity
is similarly contingent on the varying social and technological relations of different
farming systems, with implications for the ethics of human ^ animal relations. The
shift in AMS towards more technologically mediated relationships with animals has
effects on the subjectivities of both cows and humans, and on the relationships
between them. The different representations of cows in AMS (as having choice
or freedom, as needing to fulfil their genetic potential, or as being bodily and/or
behaviourally suitable or unsuitable, etc) also raise problematic issues about how
livestock animals are understood and related to in particular agricultural contexts.
More widely, questions are raised by this specific case study about the extent to
which particular technological interventions in the lives of domestic animals are
regarded, by groups including farmers and consumers, as acceptable in the ways in
which they affect animal bodies, behaviours, and experience. The ethics of how
technological `advances' are deployed in farming, and how they might restructure
human ^ animal relations, are of increasing social concern, as seen, for example, in
a recent report by the UK's Farm Animal Welfare Council on the implementation
of new livestock breeding technologies (FAWC, 2004) and in relation to developing
1056 L Holloway

ways of `rationally' evaluating and remaking animal bodies (Holloway, 2005; Stassart
and Whatmore, 2003). As such, the situated ethical responses of those directly or
indirectly involved in the application of such technologies warrant further research,
complementing the existing more abstract techniques developed for assessing the
ethical implications of particular technologies for livestock animals (eg Millar, 2000a;
2000b).
This paper has analysed secondary sources associated with ongoing debates in
farming and agricultural science concerning AMS. As such, the ways in which
humans involved in agricultural science and practice are implicated in the represen-
tation and constitution of animal subjectivities has been examined in relation to this
technology: the paper has focused explicitly on such human perspectives. This is
only one of a number of perspectives which could be developed, however, and
primary research with people and animals on farms and research institutions would
develop what is presented here and, in particular, move towards embracing the non-
human more fully in research strategies. Empirical research would further extend
understandings of human ^ animal ^ technology relationships, and gain a sense of
the coconstitution of human and animal subjectivities in differently technologically
mediated relationships. In particular, systematic observational research with cows in
AMS (and elsewhere) would afford a perspective on the coconstitution of technol-
ogies and their `users' (both cows and humans), their implications for each other, and
their adaptations to each other (Oudshoorn and Pinch, 2003), exploring, in partic-
ular, the relationships between this technology and cows' sensual experiences and
their individual and social behaviours, routines, and agencies. In this, a perspective
would be gained not only on how cows are themselves shaped, in their bodies,
behaviours, and experiences, by particular technologies, but also how the technol-
ogies themselves recall animal bodies and technologists' understandings of how those
bodies work and how livestock animals `live'. An approach involving the close follow-
ing of individual animals, observation of herd dynamics and bovine interactions,
along with systematic observation of cow ^ human ^ technology relations would allow
exploration, first, of how the engagement of cows with particular technological
apparatuses erases particular aspects of their being while also affording other
trajectories of bovine becoming, and, second, of the complex, relational ways in
which livestock animals, humans, and particular technologies (which need not be as
`high-tech' as AMS) are tangled together and emerge together, both in the design
and circulation of technologies (and their associated knowledges) and through their
everyday use in particular agricultural situations.
Nevertheless, analysis of the secondary sources used in this paper proves useful in
assessing the ongoing production of livestock animals as subjects and as subjectified
in agriculture, allowing a focus on how an emergent subjectivity is constituted by
a focus on the representation, manipulation, and control of life. The bringing to bear
of particular technologies and knowledges on animal bodies and behaviours is, as
a result, shown to be implicated in how farmed animals are subjects of and subject
to farming systems. Although this does imply that it is not possible to suggest ways of
farming which simply grant animals `freedom' to express `natural' behaviours, it does
mean that a critical analysis can explore in more complex ways the effects of power
within particular farming practices.
Acknowledgements. I would like to thank two anonymous referees, along with Phil Dunham,
Sally Eden, Russell Hitchings, and Carol Morris, for constructive comments on earlier versions
of this paper.
Farming technologies and the making of animal subjects 1057

References
Anderson K, 1997, ``A walk on the wild side: a critical geography of domestication'' Progress in
Human Geography 21 463 ^ 485
Anderson K, 2003, ``White natures: Sydney's Royal Agricultural Show in post-humanist perspective''
Transactions of the Institute of British Geographers, New Series 28 422 ^ 441
Artmann R, Bohlsen E, 2000, ``Results from the implementation of automatic milking system
multi-box facilities'', in Robotic Milking: Proceedings of the International Symposium Held in
Lelystad, The Netherlands, 17 ^ 19 August 2000 Eds H Hogeveen, A Meijering (Wageningen
Press, Wageningen) pp 221 ^ 231
Best S, Kellner D, 1991Postmodern Theory: Critical Interrogations (Macmillan, Basingstoke, Hants)
Boyd W, Watts M, 1997, ``Agro-industrial just-in-time: the chicken industry and postwar American
capitalism'', in Globalising Food: Agrarian Questions and Global Restructuring Eds D Goodman,
M Watts (Routledge, London) pp 192 ^ 225
Braun B, 2004, ``Modalities of posthumanism'' Environment and Planning A 36 1352 ^ 1355
Buller H, Morris C, 2003, ``Farm animal welfare: a new repertoire of nature ^ society relations
or modernism re-embedded?'' Sociologia Ruralis 43 216 ^ 237
Buss J, 2004a, ``EU-wide study gives automatic milking systems the all-clear'' Farmers Weekly
9 April, page 36
Buss J, 2004b, ``Grazing cows will return to the robot at milking time'' Farmers Weekly 9 April,
pp 36 ^ 37
Callon M, 1986, ``Some elements of a sociology of translation: domestication of the scallops and
the fishermen of St. Brieuc Bay'', in Power, Action and Belief Ed. J Law (Routledge and Kegan
Paul, London) pp 196 ^ 233
Castree N, Nash C, 2006, ``Editorial: posthuman geographies'' Social and Cultural Geography 7
501 ^ 504
Convery I, Bailey C, Mort M, Baxter J, 2002, ``Farmers and the farmed'', paper presented at the
Emotional Geographies conference, Lancaster University, 23 ^ 25 September
Convery I, Bailey C, Mort M, Baxter J, 2005, ``Death in the wrong place? Emotional geographies
of the UK 2001 foot and mouth disease epidemic'' Journal of Rural Studies 21 99 ^ 109
Cox G, Ashford T, 1998, ``Riddle me this: the craft and concept of the animal mind'' Science,
Technology and Human Values 23 425 ^ 438
DEFRA, 2004 Agriculture in the United Kingdom 2003 Department for Environment, Food and
Rural Affairs; Scottish Executive Environment and Rural Affairs Department; Department of
Agriculture and Rural Development (Northern Ireland) and Welsh Assembly Government,
The Department for Environment, Planning and the Countryside (The Stationery Office,
London)
de Koning K, Rodenburg J, 2004, ``Automatic milking: the state of the art'', in Automatic Milking:
A Better Understanding Eds A Meijering, C Hogeveen, C deKonig (Academic Publishers,
Wageningen) pp 27 ^ 37
Donaldson A, Lowe P, Ward N, 2002, ``Virus ^ crisis ^ institutional change: the foot and mouth
actor network and the governance of rural affairs in the UK'' Sociologia Ruralis 42 201 ^ 214
Emel J, Wolch J, 1998, ``Witnessing the animal moment'', in Animal Geographies: Place, Politics and
Identity in the Nature ^ Culture Borderlands Eds J Wolch, J Emel (Verso, London) pp 1 ^ 24
Emel J, Wilbert C, Wolch J, 2002, ``Animal geographies'' Society and Animals 10 407 ^ 412
Evans N, Yarwood R, 1995, ``Livestock and landscape'' Landscape Research 20 141 ^ 146
FAWC, 2004 Report on the Welfare Implications of Animal Breeding and Breeding Technologies in
Commercial Agriculture (Farm Animal Welfare Council, London)
Foucault M, 1975 The Birth of the Clinic (Vintage Books, New York)
Foucault M, 1979 Discipline and Punish (Penguin Books, Harmondsworth, Middx)
Foucault M, 1982, ``The subject and power'', in Michel Foucault: Beyond Structuration and
Hermeneutics Eds H Dreyfus, P Rabinow (Harvester Press, Hemel Hempstead, Herts)
pp 208 ^ 226
Foucault M,1990 The History of Sexuality.Volume1: An Introduction (Penguin Books, Harmondsworth,
Middx); first published in 1976
Foucault M, 2001 Madness and Civilisation (Routledge, London); first published in 1961
Franklin A, 1999 Animals and Modern Culture: A Sociology of Human ^ Animal Relations in
Modernity (Sage, London)
Fraser D, 2003, ``Assessing animal welfare at the farm and group level: the interplay of science
and values'' Animal Welfare 12 433 ^ 443
1058 L Holloway

Fullwood, 2005, ``A total concept from Fullwood. Merlin. The fully automated milking unit'',
http://www.fullwood.com/merlinbrochure/introduction.htm
Gascoigne Mellotte, 2005, ``GM Zenith milking system'', http://gascoigne-mellotte.com)
Grasseni C, 2005,``Designer cows: the practice of cattle breeding between skill and standardisation''
Society and Animals 13 33 ^ 50
Gravert H, 1992, ``Automatic milking and animal breeding'', in Prospects for Automatic Milking:
Proceedings of the International Symposium on Prospects for Automatic Milking, Wageningen,
Netherlands, 23 ^ 25 November 1992 Eds A Ipema, A Lippus, J Metz, W Rossing, European
Association for Animal Production, Rome, pp 395 ^ 400
Gray J, 1998, ``Family farms in the Scottish Borders: a practical definition by hill sheep farmers''
Journal of Rural Studies 14 341 ^ 356
Halachmi I, Metz J, Speelman L, 2000, ``Designing the optimal robotic milking barn: theory
and practice'', in Robotic Milking: Proceedings of the International Symposium, Lelystad,
The Netherlands, 17 ^ 19 August 2000 Eds H Hogeveen, A Meijering (Wageningen Press,
Wageningen) pp 159 ^ 166
Haraway D, 2003 The Companion Species Manifesto: Dogs, People and Significant Otherness
(Prickly Paradigm Press, Chicago, IL)
Hinchliffe S, 2001, ``Indeterminacy in-decisions: science, policy and politics in the BSE crisis''
Transactions of the Institute of British Geographers, New Series 26 182 ^ 204
Hinchliffe S, Kearnes M, Degen M, Whatmore S, 2005, ``Urban wild things: a cosmopolitical
experiment'' Environment and Planning D: Society and Space 23 643 ^ 658
Holloway L, 2001, ``Pets and protein: placing domestic livestock on hobby-farms in England and
Wales'' Journal of Rural Studies 17 293 ^ 307
Holloway L, 2005, ``Aesthetics, genetics, and evaluating animal bodies: locating and displacing
cattle on show and in figures'' Environment and Planning D: Society and Space 23 883 ^ 902
Holloway L, Morris C, 2005, ``Animal bodies and genes: contested expertise in the fields of breeding
for beef and disease eradication'', paper presented at the RGS-IBC Annual International
Conference, London, 31 August ^ 2 September 2005; copy available from the author
Hurnik J, 1992, ``Ethology and technology: the role of ethology in automation of animal production
processes'', in Prospects for Automatic Milking: Proceedings of the International Symposium on
Prospects for Automatic Milking,Wageningen, Netherlands, 23 ^ 25 November 1992 Eds A Ipema,
A Lippus , J Metz, W Rossing, European Association for Animal Production, Rome,
pp 401 ^ 408
Jones O, 2003, `` `The restraint of beasts': rurality, animality, actor network theory and dwelling'',
in Country Visions Ed. P Cloke (Pearson Education, Harlow, Essex) pp 283 ^ 307
Ketelaar-de Lauwere C, Ipema A, 2000, ``Cow behaviour and welfare if grazing is combined with
voluntary automatic milking'', in Robotic Milking: Proceedings of the International Symposium,
Lelystad, The Netherlands, 17 ^ 19 August 2000 Eds H Hogeveen, A Meijering (Wageningen
Press, Wageningen) pp 289 ^ 296
Ketelaar-de Lauwere C, Devir S, Metz J, 1996, ``The influence of social hierarchy on the time
budget of cows and their visits to an automatic milking system'' Applied Animal Behaviour
Science 49 199 ^ 211
Knight C, 2001, ``A vision of dairy production in the UK'', in Integrated Management Systems for
Livestock Eds C Wathes, A Frost, F Gordon, J Wood, (British Society of Animal Science,
Edinburgh) pp 51 ^ 57
Law J, 2004, ``Disaster in agriculture: or foot and mouth mobilities'', Centre for Science Studies,
Lancaster University, Lancaster, http://www.comp.lancs.ac.uk/sociology/papers/law-disaster-
mobilities-foot-and-mouth.pdf
Lely, 2005, ``LELY Astronaut robotic milking system'', http://www.lely.ca/main.htm
Lely, no date, ``We offer you the best, naturally'', advertising brochure, Lely (UK) Ltd, St Neots,
Cambs
Lind O, Ipema A, de Koning C, Mottram T, Hermann H-J, 2000, ``Automatic milking: reality,
challenges and opportunities'', in Robotic Milking: Proceedings of the International Symposium,
Lelystad, The Netherlands, 17 ^ 19 August 2000 Eds H Hogeveen, A Meijering (Wageningen
Press, Wageningen) pp 19 ^ 31
Luke T, 1997, ``At the end of Nature: cyborgs, `humachines', and environments in postmodernity''
Environment and Planning A 29 1367 ^ 1380
Farming technologies and the making of animal subjects 1059

Maltz E, 2000, ``Precision agriculture in dairying: individual management by automatic milking

systems'', in Robotic Milking: Proceedings of the International Symposium, Lelystad, The
Netherlands, 17 ^ 19 August 2000 Eds H Hogeveen, A Meijering (Wageningen Press,Wageningen)
pp 132 ^ 142
Mathijs E, 2000, ``Public acceptance of robotic milking'', in Robotic Milking: Proceedings of the
International Symposium, Lelystad, The Netherlands, 17 ^ 19 August 2000 Eds H Hogeveen,
A Meijering (Wageningen Press, Wageningen) pp 243 ^ 247
Michael M, 2000 Reconnecting Culture, Technology and Nature: From Society to Heterogeneity
(Routledge, London)
Midgley M, 1983 Animals and Why They Matter: A Journey Around the Species Barrier (Penguin
Books, Harmondsworth, Middx)
Millar K, 2000a, ``The role of bioethical analysis in assessing automatic milking systems: examples
of animal issues'', in Robotic Milking: Proceedings of the International Symposium, Lelystad, The
Netherlands, 17 ^ 19 August 2000 Eds H Hogeveen, A Meijering (Wageningen Press,Wageningen)
pp 248 ^ 258
Millar K, 2000b, ``Respect for animal autonomy in bioethical analysis: the case of automatic
milking systems'' Journal of Agricultural and Environmental Ethics 12 41 ^ 50
Mottram T, Masson L, 2001, ``Dumb animals and smart machines: the implications of modern
milking systems for integrated management of dairy cows'', in Integrated Management Systems
for Livestock Eds C Wathes, A Frost, F Gordon, J Wood, (British Society of Animal Science,
Edinburgh) pp 77 ^ 84
Murdoch J, 2004, ``Humanising posthumanism'' Environment and Planning A 36 1356 ^ 1359
Netz R, 2004 Barbed Wire: An Ecology of Modernity (Wesleyan University Press, Middletown, CT)
Noske B, 1997 Beyond Boundaries: Humans and Animals (Black Rose Books, Montreal)
Novek J, 2005, ``Pigs and people: sociological perspectives on the discipline of nonhuman animals
in intensive confinement'' Society and Animals 13 221 ^ 244
Oudshoorn N, Pinch T, 2003,``Introduction: how users and non-users matter'', in How Users Matter:
The Co-construction of Users and Technologies Eds N Oudshoorn, T Pinch (MIT Press,
Cambridge, MA) pp 1 ^ 25
Owen J, 2003, ``Evaluating robotic milking at Gelli Aur College'' State Veterinary Journal 13 15 ^ 18
Page B, 1997, ``Restructuring pork production, remaking rural Iowa'', in Globalising Food: Agrarian
Questions and Global Restructuring Eds D Goodman, M Watts (Routledge, London) pp 133 ^ 157
Palmer C, 2001, `` `Taming the wild profusion of existing things'? A study of Foucault, power and
human/animal relationships'' Environmental Ethics 23 339 ^ 358
Philo C,Wilbert C, 2000,``Animal spaces, beastly places: an introduction'', in Animal Spaces, Beastly
Places: New Geographies of Human ^ Animal Relations Eds C Philo, C Wilbert (Routledge,
London) pp 1 ^ 34
Pile S,Thrift N,1995,``Introduction'', in Mapping the Subject: Geographies of Cultural Transformation
Eds S Pile, N Thrift (Routledge, London) pp 241 ^ 266
Pluhar E, 1995 Beyond Prejudice: The Moral Significance of Human and Nonhuman Animals
(Duke University Press, Durham, NC)
Pratt G, 2000, ``Geographies of subject formation'', in The Dictionary of Human Geography
4th edition, Eds R Johnston, D Gregory, G Pratt, M Watts (Blackwell, Oxford) pp 802 ^ 804
Quinn M, 1993, ``Corpulent cattle and milk machines: nature, art and the ideal type'' Society and
Animals 1 145 ^ 158
Rabinow P, Rose N, 2003, ``Thoughts on the concept of biopower today'', http://www.molsci.org/
files/Rose Rabinow Biopower Today.pdf
Regan T, Singer P, 1976 Animal Rights and Human Obligations (Prentice-Hall, Englewood Cliffs,
NJ)
Risan L, 2005, ``The boundary of animality'' Environment and Planning D: Society and Space 23
787 ^ 793
Ritvo H, 1987 The Animal Estate: The English and Other Creatures in the Victorian Age
(Harvard University Press, Cambridge, MA)
Seabrook M, 1992, ``The perception by stockpersons of the effect on their esteem, self concept
and satisfaction of the incorporation of automatic milking into their herds'', in Prospects for
Automatic Milking: Proceedings of the International Symposium on Prospects for Automatic
Milking,Wageningen, The Netherlands, 23 ^ 25 November 1992 Eds A Ipema, A Lippus, J Metz,
W Rossing, European Association for Animal Production, Rome, pp 409 ^ 413
Sharp M, 2003, ``Choosing the layout that fits the robots'' Farmers Weekly 28 March, pp 87 ^ 88
Simon B, 2003, ``Introduction'' Cultural Critique 53 3 ^ 8
1060 L Holloway

Smith C, 1983, ``A case study of structuration: the pure-bred beef business'' Journal for Theory of
Social Behaviour 13 3 ^ 18
Spo«rndly E, Krohn C, van Dooren H, Wiktorsson H, 2004, ``Automatic milking and grazing'',
in Automatic Milking: A Better Understanding Eds A Meijering, C Hogeveen, C deKonig
(Academic Publishers, Wageningen) pp 263 ^ 272
Stassart P, Whatmore S, 2003, ``Metabolising risk: food scares and the un/re-making of Belgian
beef '' Environment and Planning A 35 449 ^ 462
Stibbe A, 2003, ``As charming as a pig: the discursive construction of the relationship between
pigs and humans'' Society and Animals 11 375 ^ 392
Thomas K, 1983, Man and the Natural World: Changing Attitudes in England 1500 ^ 1800 (Allen
Lane, London)
Thrift N, 2005 Knowing Capitalism (Sage, London)
Ufkes F, 1998,``Building a better pig: fat profits in lean meat'', in Animal Geographies: Place, Politics
and Identity at the Nature ^ Culture Borderlands Eds J Wolch, J Emel (Verso, London)
pp 241 ^ 258
van der Schans F, van der Weijden W, 2000, ``Koe moet in de wei blijven'' NRC Handelsblad
18 April, http://www.clm.nl/artikelen/koeindewei.html
Walton J, 1984, ``The diffusion of the improved Shorthorn breed of cattle in Britain during the
18th and 19th centuries'' Transactions of the Institute of British Geographers. New Series 9
22 ^ 36
Walton J, 1986, ``Pedigree and the national cattle herd circa 1750 ^ 1950'' Agricultural History
Review 34 149 ^ 170
Walton J, 1999, ``Pedigree and productivity in the British and North American cattle kingdoms
before 1930'' Journal of Historical Geography 25 441 ^ 462
Watts M, 2000, ``Afterword: enclosure'', in Animal Spaces, Beastly Places: New Geographies of
Human ^ Animal Relations Eds C Philo, C Wilbert (Routledge, London) pp 292 ^ 304
Webster A, 2003,``Assessment of animal welfare at farm and group level: introduction and overview''
Animal Welfare 12 429 ^ 431
Whatmore S, 2000, ``Heterogeneous geographies: reimagining the spaces of N/nature'', in Cultural
Turns/Geographical Turns Eds I Cook, D Crouch, S Naylor, J Ryan (Pearson Education, Harlow,
Essex) pp 265 ^ 272
Whatmore S, 2002 Hybrid Geographies: Natures Cultures Spaces (Sage, London)
Whatmore S, Thorne L, 1998,``Wild(er)ness: reconfiguring the geographies of wildlife'' Transactions
of the Institute of British Geographers, New Series 23 435 ^ 454
Whatmore S, Thorne L, 2000, ``Elephants on the move: spatial formations of wildlife exchange''
Environment and Planning D: Society and Space 18 185 ^ 203
Wiktorsson H, SÖrensen J, 2004, ``Implications of automatic milking on animal welfare'', in Automatic
Milking: A Better Understanding Eds A Meijering, C Hogeveen, C deKonig (Academic
Publishers, Wageningen) pp 371 ^ 381
Wilkie R, 2005, ``Sentient commodities and productive paradoxes: the ambiguous nature of
human ^ livestock relations in Northeast Scotland'' Journal of Rural Studies 21 213 ^ 230
Williams A, 1999, Review of Animal Geographies: Place Politics and Identity in the Nature ^ Culture
Borderlands Eds J Wolch, J Emel Environment and Planning D: Society and Space 17 621 ^ 624
Wolch J, Emel J, 1995, ``Bringing the animals back in'' Environment and Planning D: Society and
Space 13 632 ^ 636
Woods M, 1998, ``Mad cows and hounded deer: political representations of animals in the British
countryside'' Environment and Planning A 30 1219 ^ 1234
Yarwood R, Evans N, 1998, ``New places for `Old Spots': the changing geographies of domestic
livestock animals'' Society and Animals 6 137 ^ 165

ß 2007 a Pion publication printed in Great Britain

Conditions of use. This article may be downloaded from the E&P website for personal research
by members of subscribing organisations. This PDF may not be placed on any website (or other
online distribution system) without permission of the publisher.