The Ethics of Touch and the Importance of Nonhuman Relationships in Animal Agriculture

Animal agriculture predominantly involves farming social animals. At the same time, the nature of agriculture requires severely disrupting, eliminating, and controlling the relationships that matter to those animals, resulting in harm and unhappiness for them. These disruptions harm animals, both physically and psychologically. Stressed animals are also bad for farmers because stressed animals are less safe to handle, produce less, get sick more, and produce poorer quality meat. As a result, considerable efforts have gone into developing stress-reduction methods. Many of these attempt to replicate behaviours or physiological responses that develop or constitute bonding between animals. In other words, humans try to mitigate or ameliorate the damage done by preventing and undermining intraspecies relationships. In doing so, the wrong of relational harms is compounded by an instrumentalisation of trust and care. The techniques used are emblematic of the welfarist approach to animal ethics. Using the example of gentle touching in the farming of cows for beef and dairy, the paper highlights two types of wrong. First, a wrong done in the form of relational harms, and second, a wrong done by instrumentalising relationships of care and trust. Relational harms are done to nonhuman animals, whilst instrumentalisation of care and trust indicates an insensitivity to morally salient features of the situation and a potential character flaw in the agents that carry it out.


Introduction
An important and under-discussed ethical dimension of animal agriculture is that it predominantly involves the farming of social animals. The more intensive the farming, the more it requires animals to be reared closely to one another and in greater numbers. By its very nature then, the more intensive the farming, the more social the nonhuman animals farmed need to be. At the same time, it is an unavoidable part of animal agriculture that it requires disrupting, controlling, and eliminating many of the relationships that matter most to those animals. In this paper, I argue that these features of agriculture give rise to what I term relational harms. These relational harms directly affect nonhuman animals as well as indirectly harming farmers. Unhappy animals are more difficult and unsafe to handle, produce less milk, are more prone to illness, grow more slowly, are smaller, and produce poorer quality meat (Waiblinger et al., 2006, pp. 186-187). As a result, a great deal of effort has been expended on improving and discovering new means of stress-reduction. Many of these have involved trying to find ways to mitigate or ameliorate the damage done by disrupting the intraspecies relationships of animals in human care. One way to achieve this has been by replicating behaviours or physiological responses involved in bonding between animals, and which the animals themselves use to reduce stress. These methods have received little ethical attention, having generally been viewed as positive welfare-improving interventions. As I will show, this view misses some important morally relevant features of the practices. One of these involves a lack of consideration of the underlying systems and structures that create a need for the interventions in the first place. Another is that the methods used have wrongly instrumentalised relationships of care and trust between humans and nonhumans.
The starting point of my argument is that nonhuman animals have morally relevant interests, and that these interests are sufficiently strong as to ground rights for nonhuman animals and duties in humans. The interests that matter most are those non-contingent basic interests necessary for any kind of decent life. Key amongst those, are the interest in not suffering, and in continued existence. 1 Additionally, as will become clear in the course of the paper, social animals also possess strong interests in forming and maintaining good quality relationships. Good quality relationships are those that satisfy the need for sociability in non-harmful ways, thus excluding abusive and exploitative relationships. The importance of these interests to social nonhuman animals grounds duties of non-interference in humans.
Because they are by nature social, the kinds of animals commonly farmed need good quality relationships to live decent lives. We know from decades of research on social animals that depriving them of mutual contact causes serious harm. 2 1 For influential defences and discussions of the interest-theory of animal rights see (Cavalieri, 2001;Cochrane, 2013;Feinberg, 1986;Pepper, 2018) 2 We know this because social animals have been extensively used as models for humans in experiments designed to understand the harm of isolation and separation (see Lieberman, 2013, Chap. 3). In the case of farm animals, studies have shown increased levels of stress in cattle separated from familiar conspecifics or placed in unwanted proximity with others (Boissy and Le Neindre, 1997;Boyland et al., 2016;Val-Laillet et al., 2009;Weary and Chua, 2000).
Similarly, they suffer reductions in wellbeing when the kinds of contact they have are poor quality, such as when they are placed in overcrowded conditions, placed with unfamiliar conspecifics, and placed in unnatural social groupings by age or sex. Both the disruption of desired relationships between nonhuman animals and the enforcement of undesired ones set-back nonhuman animal interests. Thus, animal agriculture, especially intensive farming, by nature does relational harms to nonhuman animals.
Unlike rights against suffering and being killed, relational rights for nonhuman animals have received little attention. One important area that I will focus on is how the structure of modern animal agriculture threatens the conditions necessary for nonhuman relational rights to be realised. Although my overarching approach is rights-based, a concern for the protection of interests is present across ethical theory. This means that many of my arguments can also be justified by appeal to alternative ethical theories.
Whilst rights act as way of representing and realising basic requirements of justice, they do not describe the whole of the moral landscape. Besides respecting the rights of others, our character, motivations, and intentions also matter morally, and they matter separately from their role in the fulfilment of duties correlative to rights. For example, the person who does not steal from others out of concern for the good of those others ought to be thought of differently than the person who does not steal merely because they are frightened of being caught and punished. We ought to think differently of the person who acts rightly because they have good underlying intentions and motives and the person who does the right thing but whose intentions and/or motivations are bad. Well-motivated agents are likely to go beyond the minimal requirements of justice, showing care and attention towards others rather than merely seeking to avoiding harming them. For this reason, my argument is also concerned with uncovering and evaluating the intentions and motivations of agents involved in the practices I describe. This latter element means that even if a rightsbased framework for governing our moral relationships with nonhuman animals is rejected, there remain strong reasons for re-evaluating the ethical status of those practices.
This framework for evaluating farming practices is at odds with the welfarist approach frequently taken in the farm animal welfare literature. In the welfarist paradigm, acts are often evaluated in more consequentialist terms involving comparative assessments of harms done by different practices (c.f. Garner, 1993, pp. 16-20). Consequentialist approaches evaluate the moral status of acts by reference to their consequences; acts are good insofar as they produce more goodness in the world and bad insofar as they do not. There are three significant problems with this means of evaluating practices. The first is that the counterfactual harm comparison often assumes the necessity of animal agriculture. Since this premiss is false, the argument is rendered unsound. The second problem is that the interests of nonhuman animals are often given too little weight, allowing quite serious nonhuman interests to frequently be sacrificed for relatively trivial interests of humans. However, the difference in relative importance of interests is seldom justified. As others have shown, this premiss is highly questionable (Singer, 1974). Both of these issues can be addressed under a consequentialist framework, and this would require very 12 Page 4 of 20 significant changes to farming. However, the third problem is with consequentialism itself rather than the way that arguments for welfare improvement are typically constructed. As has been frequently been argued, consequentialist approaches place too much emphasis on the overall goodness produced and not enough on moral status of individual beings. As a result, very serious harms to individuals are permitted, even required, if they are likely to increase aggregate levels of good in the world. Such a view is incompatible with the thought that individual beings matter for their own sake and so ought not be used as a mere means to satisfy the preferences or increase the happiness of others. Some of these concerns will be discussed in more detail in "Motive and Intent" and "The Welfare Paradigm" sections.
Given the importance of social interactions to animals themselves, exploring the ethical implications of farming in terms of those relationships is important. Recent literature has tended to be more focussed on physical harms done to individual creatures, on the use of nonhuman animals as a form of exploitation, or on more abstract consideration of freedom. Even care ethicists, who rightly draw attention to the ethical significance of relationships, have tended to pay more attention to the humananimal bond than the relationships nonhuman animals have with one another. In her valuable discussion of the incarceration and confinement of nonhuman animals, Lori Gruen draws attention to the fact that captivity takes away freedom to move, privacy, choice over what and when to eat, and the ability to choose one's associations (Gruen, 2018). Gruen discusses these features of captivity in the context of humans, and explores dignity and choice in nonhuman contexts, but she does not devote much space to nonhuman association. It is this gap that I seek to address here.
My contribution to the debate around farm animal welfare is twofold. First, I show that the relationships of farmed animals are important to them and demonstrate how farming practices interfere with those relationships. The harms arising out of these interferences, I argue, are an inescapable product of systems of domination present in farming. I conclude that animal agriculture, by necessity, involves wrongs done to nonhuman animals themselves in the form of relational harms. These harms violate a right held by nonhuman animals to enjoy associations that matter to them. Second, I explore the ethics of welfarist approaches, developed to lessen the impact of these harms. These interventions do not merely involve a wrong done to an animal but, as I will show, are also symptomatic of a moral blind spot in the agents who use them. Instead of than being conceived of primarily as welfare improvements, I will argue that they represent an ethically questionable instrumentalisation of care and the social needs of animals. My purpose throughout will be to draw attention to ethically salient but hitherto underappreciated features of common practices in order to provoke reflection and reassessment.
In order to illustrate these two points, I focus predominantly on the situation of cows, who are farmed for both beef and dairy. I do this because of the ubiquity of cattle farming, and because they are a species often mistakenly regarded as lacking in individuality or personality. Despite the focus on cows, the arguments I make also apply to other farmed social animals. Whilst many techniques are used to address the welfare harms done to cows, my main concern is with what are termed gentle touching methods. 3 In gentle touching, touch is used as a means to secure particular ends: habituating animals to the presence of humans and reducing stress. The handler, stocksperson, or researcher applies touch as a therapeutic or training technique, often in highly systematised ways. Gentle touching has been extensively tested on cows, but has also been tried on pigs, deer, horses, and dogs. These techniques are particularly interesting as an example because the tactile nature of them suggests caring and intimacy in the human animal relationship. These features are at odds with the systems they occur within invite ethical evaluation. I argue that, in the contexts it is used, gentle touching ought to be thought of as an ethically rich crossspecies encounter, and that conceiving it merely in instrumental welfarist terms is a mistake. Using touch as a means to secure instrumental benefits fails to properly attend to its moral dimensions, or to acknowledge the systematic wrongs that make touch a useful tool in welfarist approaches to animal husbandry. Whilst there are contexts where gentle touching is morally praiseworthy, it unlikely that animal agriculture is one of them.

The Sociability of Cows
Cows are by nature highly social animals, and this sociability manifests in both free-roaming and farmed settings. In their herds, cows keep close contact with one another and maintain their relationships using a variety of behaviours including touch, vocalisation, body language, and olfactory signals. Individual cows form particular associations, including both non-kin friendships and long-term mother-calf bonds (Marino & Allen, 2017;Reinhardt & Reinhardt, 1981;Tresoldi et al., 2015;von Keyserlingk & Weary, 2007).
Many of the relationships that are important to cows, and the means by which they are maintained, are disrupted by farming practices. Whether and when they have children is determined by the farmer. Commonly, dairy calves are separated from their mothers within twenty-four hours of birth and then reared and weaned in same sex groups of calves. On dairy farms, calves are typically housed individually. At various points in their lives cows are separated and kept either in individual pens or small group stalls. Members of the herd will be sold, kept apart for a variety of management and veterinary procedures, and slaughtered. Existing social groups are also disrupted and controlled by being merged into other groups, and having unfamiliar individuals introduced to them. Individuals who cease to be productive are taken from the herd and killed. Preferences that cows commonly exhibit, such as for the company of their own kin, for grooming partners, and for grazing companions, are ignored and overridden at the convenience of the farmer, and often to the detriment of the animals.
The result of these disruptions is increased stress. Since cows are by nature social, it is unsurprising that they exhibit classic behavioural and physiological signs of stress when isolated from their herd: vocalisations, avoidance behaviour, antagonism, elevated heart rate, and increased cortisol levels. Separated from friends and family, cows become more fearful, reacting more strongly to unusual noises and other stimuli (Boissy & Le Neindre, 1997). If their herdmates are stressed, then they suffer from reduced appetite and increased wariness (Boissy et al., 1998). Mixing with unfamiliar conspecifics elicits antagonism and changes in appetite (Gutmann et al., 2015, p. 2). Isolated calves are less active and grow more slowly (Weary & Chua, 2000;Bouissou et al., 2001, pp. 130-135;Wagner et al., 2015;Marino & Allen, 2017, pp. 486-487). Of all their relationships, the bond between dams and their calves is particularly important to cows. Mother-child relationships persist for many years, with cows preferring contact with their children, whether sons or daughters, as both grazing and licking partners (Reinhardt & Reinhardt, 1981, p. 144). Intensive dairy production effectively eliminates all maternal behaviour and prevents these relationships. In beef farming, cows are also hindered from forming maternal bonds by batch breeding and close housing. Batch breeding involves dividing cows into breeding groups so that they calve at the same time, thus ensuring maximal efficiency in reproduction. The use of artificial insemination to facilitate this is widespread. Calving periods are dictated by availability of feed, labour costs, and prices on meat and dairy markets. Close housing is a cattle management system where cattle are confined for long periods indoors in close proximity to one another. These features make it more difficult for a mother to spend the time she normally would bonding with her calf at slight distance from the herd (von Keyserlingk & Weary, 2007, p. 108). In both dairy and beef farming, calves are often forced into early weaning though the use of spiked nose rings or spiked head harnesses. These devices are designed to make suckling painful to the mother so that her maternal instincts are overcome, and she pushes her calf away. Separation and isolation of calves from their mothers has long term negative physical and psychological consequences for them (Bouissou et al., 2001, pp. 130-132), including a reduced capacity to form and maintain their own social relationships (Wagner et al., 2015;Monsó et al., 2018, pp. 302-303). In short, disruption of cows' relationships and separation from their friends and family reduces their wellbeing and their capacity for future wellbeing.
The negative consequences that come from disrupting their relationships are bad for both cows and for farmers and handlers. Cows suffer a reduction in wellbeing, and farmers suffer both increased personal risks from more aggressive and skittish animals, and instrumental disbenefits: reduced production, increased costs, poorer quality products. Unfortunately, many of these disruptive practices are an intrinsic and inevitable feature of both beef and dairy farming. After all, what could be more disruptive of a relationship than the abrupt separation and killing of some of those involved in it-a practice intrinsic to beef farming? Other practices cannot be avoided without radical change to farming. As a result, considerable effort has been expended on discovering means to mitigate and ameliorate some of the relational harms done by humans to cows. Of particular interest here are efforts designed to replicate and systematise cows' own intraspecies behaviours in human animal interactions on the farm. The use of gentle touching is an increasingly significant area of research in this area.
Cows are highly sensitive to touch, and this often makes them fearful when touched by other animals such as their human handlers (Marino & Allen, 2017, p. 476). Touch, particularly in the form of licking, is an important means by which bovine social bonds are established and maintained (Tresoldi et al., 2015), serving also to calm and reassure. When they lick each other in this context, cows focus upon one another's head, neck and shoulder areas, parts of the body that are particularly sensitive. These lickings are sought out from preferred partners, suggesting that it forms part of relationships of friendship (Val-Laillet et al., 2009). Licking is solicited though behavioural signals; a lowering of head and neck and gentle pushing under a herdmate's head or neck (Bouissou et al., 2001, pp. 122-123). Experiments have shown that licking forms as much as 10% of interactions between cows, with this figure increasing drastically when they are forced into more crowded and stressful situations where they must compete for food, water, and space (Tresoldi et al., 2015). Licking forms a long-term and important element of mother-calf relationships (Reinhardt & Reinhardt, 1981, p. 140) and is key to establishing initial mother-child bonds. After giving birth, cows spend several hours intensively licking their children. If they are prevented from doing this, then there is a far greater risk of a calf being rejected by its mother (von Keyserlingk & Weary, 2007, pp. 107-108). As we have seen, these are perhaps the most important relationships to cows.

Gentle Touching
Since cows are calmed and reassured by licking, it is not surprising that efforts have been made to calm and reassure them by replicating those behaviours. Numerous studies have investigated the effects of touch on animals reared for slaughter, including cows, pigs, and deer. Researchers have been interested to discover if tactile interactions can lower stress and reduce fearful reactions, resulting in easier handling and more tender flesh. Experiments have been carried out on both infants and adult animals to determine if gentle touching produces persistent benefits (Schmied et al., 2008a(Schmied et al., , b, 2010Probst et al., 2012Probst et al., , 2013Tallet et al., 2014;Lürzel et al., 2015Lürzel et al., , 2016Wang et al., 2016;Lange et al., 2020;Okabe et al., 2020). The intended outcome of much of this research is to improve the welfare of farm animals and their handlers, whether that be as a goal sought for the sake of the animal itself or as a means to increasing production and meat quality. Because of this, and because these methods are often contrasted with techniques such as slapping and use of prods and goads, it is easy to see gentle touching techniques as in purely positive terms. Later, I will show why we ought to question this initial assumption.
Researchers interested in reducing and repairing the relational harms done by farming practices have attempted to replicate the effects of social licking in cows via gentle touching by humans. From observed behaviour, these researchers hypothesised that stroking cows using similar motions and on the same body regions as are concentrated upon in social licking might have physiological and behavioural benefits. Numerous experiments to test this have been carried out (Waiblinger et al., 2004;Schmied et al., 2008a, b;Probst et al., 2012Probst et al., , 2013Lürzel et al., 2015;Lange et al., 2020). These experiments found that stroking a cow's withers and ventral 12 Page 8 of 20 neck, regions most commonly licked, resulted in greatest stress reduction measured in terms of decreased heart rate, reduced levels of stress hormones, and relaxed behaviour. Many have concluded that 'cows may in part perceive human stroking of body regions often licked similarly to social licking' (Schmied et al., 2008a, p. 25). In other words, stroking these parts of the body mimics cows' consoling and bonding behaviour. Effects of repeated short periods of stroking over a few weeks resulted in cows becoming habituated to humans and more willing to approach them (Schmied et al., 2008b, p. 604), effects which last for several weeks (Schmied et al., 2008a, p. 602).
Whilst the studies referenced so far generally refer to welfare benefits for the animals themselves, it is never the sole benefit sought and is most commonly pursued as a means to benefit farmers. Reducing fear and increasing approach reactions makes it easier to control the movements of animals being farmed (Schmied et al., 2008a). Habituation also makes handling easier, (ibid. p.604) reducing behaviour such as stepping and kicking during milking, rectal palpitation (for pregnancy testing), and artificial insemination. Reducing prevalence of these behaviours lowers the risk of injury to stockspersons and veterinarians. Stress reduction achieved through simulated licking can result in increased daily weight gain in calves and thus higher milk yield in later life (Lürzel et al., 2015, p. 15). Researchers have also stressed the benefits in marketing terms as a response to consumer demand for higher-welfare animal products (Probst et al., 2013, p. 47;Lürzel et al., 2015, p. 9), and as a means of improving the public perception of farmers (Waiblinger et al., 2006, p. 188). Besides dairy cows, gentle touching has also been directed at making the transport and slaughter process of beef cattle less stressful. This allows for safer travel to and through the abattoir, with fewer injuries to both cattle and handers. It also reduces the need for forceful methods such as hitting, prodding, and goading, which can result in bruising and parts of carcasses that cannot be sold for human consumption. Finally, it lowers the levels of stress hormones in cattle, increasing the tenderness and thus value of their flesh. Pre-slaughter, stroking leads to quickergrowing, healthier, and larger animals, which means more and better-quality meat (Hemsworth et al., 2011;Probst et al., 2012Probst et al., , 2013. To summarise, agricultural scientists have sought to mimic the behaviours that create and maintain positive intraspecies relationships (Schmied et al., 2008a, p.596). Their purpose in doing so has been to gain the physical and behavioural benefits either at the cost of the relationships themselves or without the need to support or maintain those relationships. Gentle touching is seen as a means of counteracting the unwanted effects of fear and violence (Waiblinger et al., 2006, pp. 186-187).
Besides the harms typically highlighted by animal rights theorists, two features of these approaches to animal welfare are ethically troubling. The first is that the need for them reveals the extent to which the relationships of nonhuman animals are controlled and sacrificed for human convenience and gain. Underpinning the ordering of animal lives is an attitude that values their relationships primarily in terms of physiological terms connected with economic productivity. This attitude displays an insensitivity to something that is morally valuable in their affective positive relationships. Second, approaches to repairing or reducing the relational harms done, such as gentle touching, often involve deception and the fostering of trust as a means to create and exploit vulnerability. The use of trust in this way suggests an untrustworthy character and a willingness to use the vulnerability of innocents for gain. Let us consider these two features in turn, starting with the control farming practices exert over nonhuman relationships.

Domination and Relational Harms
Farming practices structure every aspect of an animal's life. What is more, the animals farmed by humans have been selected and bred partly in order to increase their vulnerability. Breeds are chosen not just for productivity, but also for docility and sociability (Haskell et al., 2014). Animal agriculture amounts to a system of complete domination over beings created in order to be subject to it. This control has extended so far that farmers have even found ways to increase milk yields by controlling cows' perception of reality, using VR headsets to make them think they are in a summer field (BBC News, 2019). This system of domination involves the exercise of power over the vulnerable in order to do them serious harm. In the context of farming, humans exercise almost unconstrained power over nonhuman animals, able to regulate and control every aspect of their existence including the moment of their birth and death and even their perception of the world around them. Farming not only narrows down the range of possibilities for action open to nonhuman animals, it also constitutes a system that they have minimal capacity to resist and none to reverse (c.f. Palmer, 2017, pp. 351-352). In contrast, nonhuman animals have no power over the lives of the humans that farm them. The power of humans over the animals in their care is not only possessed, it is also exercised. Indeed, its exercise is what characterises animal agriculture: the breeding and captivity of nonhuman animals in order to consume them and the products of their bodies. Thus, the form of domination at work is in large part structural; it is an intrinsic part of the practice of farming and so involvement in its systems requires agents within it to carry out acts of domination. Furthermore, the vulnerability and powerlessness of nonhuman animals subject to domination comes as a result of processes of domestication, selection, and breeding partly done in order to create and increase the possibility of domination. Palmer describes this feature of domestication as 'physiologically internalizing dominance' (Palmer, 2017, p. 355). From a rights-based perspective, that these power-relationships exist and are used to do serious, systematic, and sustained harm is an injustice. 4 4 Some theories about the wrongness of domination argue that the mere possibility of the arbitrary exercise of power is part of what makes it wrong. The normative force of this claim comes from the value of freedom. As Isaiah Berlin famously argued, slavery is not merely objectionable because it prevents people from doing what they want, it is also wrong because those enslaved are aware that they are constrained whatever they may wish to do (see Berlin, 1969, p. xliii, footnote). This element of the wrongness of domination is less pressing in nonhuman cases because nonhuman animals lack the level of autonomy and self-consciousness relevant to this conception of freedom (c.f. Cochrane, 2009). Nevertheless, the capacity to act freely in pursuit of their desires and to exercise control over their attachments remains important for the good of nonhuman animals.
The ordering of nonhuman animal lives is not just objectionable because its purpose is to harm and exploit their bodies for human gain. It is also objectionable because it robs animals of the ability to form and maintain relationships that matter to them and that are necessary for them to live good lives. As we have seen, cows develop preferences for other cows in their herds. These preferences are for both their own kin and for non-kin partners, and they manifest in different kinds of shared behaviours. Cows seek out and enjoy spending time grooming one another and engaging in other activities such as grazing with their chosen partners. We have also seen that they become unhappy if prevented from spending time in these shared activities or if isolated. We also know that cows try to reassure each other when stressed, and that they suffer from loneliness and grief (Weary & Chua, 2000;McGrath et al., 2013;King, 2014, p. 38). Whilst there has been consideration given to the way in which farming limits the freedom of nonhuman animals to pursue their desires, this has largely been framed in terms of their freedom of movement, freedom from harm, or about whether they possess autonomy (See, for example, Cochrane, 2009;DeGrazia, 2011;Hadley, 2013;Giroux, 2016). Little space by contrast has been devoted to the ways in which farming controls the emotional lives and associative freedoms of nonhuman animals. Surprisingly, even care ethics has passed little comment on the relationships between nonhuman animals, instead concentrating on what humans owe to them in virtue of morally relevant entanglements. 5 The evidence I have presented so far shows that farmed nonhuman animals have a strong interest in their relationships with one another, strong enough that they are seriously harmed by our controlling of their lives. This interest is sufficient to underpin a prima facie duty of non-interference and a correlative right for them to enjoy their associations. Even if their associations are not chosen autonomously, it is clear that the enjoyment of them matters a great deal to the animals themselves. These associations are valuable in much the same way as unchosen friendships and loves in humans are. This means that when animals are made to suffer through having their relationships interfered with, the wrong done to them is not wholly captured by reference to the suffering alone, it must also include consideration of restrictions to their freedom to associate.
Besides limiting their freedom to control their associations, practices that aim to repair the damage from doing so also carry ethical costs. Let us turn now to consideration of the role of deception and the use of trust in those practices, using gentle touching to illustrate the issues. 5 One interesting exception to this is Nel Noddings. Noddings argues that the relationships between animals who care for one another gives us reason to care for them. She uses this to show that we owe something to cetaceans. However, Noddings also claims that because cows do not care for one another that we therefore have lesser reason to care for them. Since Noddings' claim about cows is false, by the logic of her own argument we must therefore care for cows to a much greater degree than she allows. See Noddings (1991, p. 419)

The Ethics of Touch
In the literature, touch is discussed in highly structured, controlled, and formalised ways, with the aim being to secure defined benefits through replicable techniques. Outcomes of different methods of touching are measured and ranked in order to efficiently maximise benefits. The following example is paradigmatic: These special forms of touches are defined in velocity, direction of movements and intensity of contact (Tellington-Jones, 1999). To apply the treatment, the experimenter placed both hands close to the withers. The left hand rested, whilst the right hand moved in clock-wise circles with one-and-a-quarter turns. Each of these movements lasted for 2 s and was carried out on different parts of the body. Stroking and touching were conducted with the grain of the coat. During treatment, both hands were kept in touch with the animal's coat. The experimenter approached the animal from the side by speaking to them in a gentle and standardised manner. Then, the experimenter crouched down to eye level to draw slightly closer to the calf and extended one arm slowly (10 cm/s) in the direction of the calf's shoulder, then began contact (applying hands onto the coat) at the calf's withers and the ventral part of the neck at the experimenter's side (Schmied et al., 2008b). If it was not possible to touch the calf, the experimenter stood still for 10 min as close as possible to it with the goal that the calf would not move. (Probst et al., 2012, p. 44) What is troubling about such a mechanistic approach is that it suggests an inattentiveness to the moral dimensions of touch. In their recent excellent paper on intraspecies touch, Susana Monsó and Birte Wrage, show how touch can communicate moral emotions, be a moral practice in itself, and can signal normative capacities (Monsó and Wrage, 2021). A good example of this lies in the role played by touch in consoling behaviour, where animals calm and reassure the distressed though tactile contact. The fact that cows appear to experience emotional contagion (Marino & Allen, 2017, p. 483;and c.f. Monsó and Wrage, 2021, p. 4), becoming distressed when their herdmates are frightened or in pain, suggests empathic capacity (Monsó et al., 2018b, p. 284), which helps explain why they engage in licking behaviours in response to encountering distress. Psychological capacities like empathy underpin caring relationships, making it possible for animals to be motivated by care for others rather than merely by a desire for social contact (Monsó and Andrews, forthcoming).
More importantly, touch might also serve as a means by which nonhuman animals non-linguistically signal and learn trustworthiness. As Monsó and Wrage explain, touch is an important sense for navigating vulnerability and threat. Animals use touch to learn about the world outside the body and keep themselves safe from danger. When an animal moves close enough to touch another, they not only make it possible to harm that other, but they also open themselves up to being harmed in return. The uncertainty in these situations is often resolved through the use of touch. By purposefully allowing others to touch them, an animal may demonstrate trust, and by exercising restraint they show trustworthiness. Monsó and Wrage illustrate this restraint using the example of the toleration adult animals show towards 12 Page 12 of 20 juveniles engaged in rough play. Enduring discomfort rather than exploiting vulnerability suggests a moral capacity for tolerance (Monsó and Wrage, 2021, pp. 15-16). Young animals feel safe acting roughly because they have learned to trust their fellows. When rough play exceeds certain bounds, animals respond in ways that set boundaries and develop social norms (Monsó and Wrage, 2021, p. 18). The sorts of behaviours at play in boundary-setting and the development of norms of reciprocity are present in a range of nonhuman animals (Monsó and Andrews, forthcoming).
Elsewhere, and drawing from Annette Baier's idea of infant trust (Baier, 1986) and the work of Anita Silvers and Leslie Pickering Francis (Silvers & Francis, 2005), I have argued for a relatively thin conception of trust in nonhuman animals. Under this conception, the ability to trust requires only a capacity to be responsive to others, to be optimistic about another's behaviour, and to be able to behave in ways that gain and maintain positive relationships. This view of trusting differs from mere reliance because the beliefs involved can have affective elements such as love, affection, or friendship (Cooke, 2019, pp. 189-191). The required optimism about the behaviour of others is not one that requires a rich conception of rational agency and a complex theory of mind. Rather, trust can be grounded in unchosen beliefs and emotional connections. Trusting someone is not something that occurs as a result of a choice, but rather as a product of our relationships, our interactions, and our familiarity (Gambetta, 1990, p. 230). These beliefs are ones with important unchosen affective components. Nonhuman animals, much like young children, reliably behave in ways that suggest they have the sorts of beliefs necessary for this thin conception of trust. Indeed, the domestication and handling procedures necessary for many farming practices depend upon their having those sorts of beliefs (c.f. Midgley, 1983, pp. 58-59). Habituation attempts are often not merely about giving nonhuman animals reliable expectations about their handlers, but also about forming emotional bonds. As a result of interactions like gentle touching, nonhuman animals may come to form beliefs about the sincerity and goodwill of their handlers.
The behaviour of cows in response to gentle touching suggests at least this kind of trust in their handlers. However, Monsó and Wrage's argument indicates that nonhuman animals might be capable of an even richer notion of trust. Some accounts of trust, such as that of Carolyn McLeod, root trusting relationships in shared normative commitments. McLeod argues that for her to trust in someone not to harm her she must believe that that the other person shares a belief in the wrongness of harming and that they have sufficient integrity to act on that belief (McLeod, 2000). Whilst this vision of trust appears to demand a certain level of rationality that appearance might be mistaken. The capacity to act with restraint, to experience and express moral emotions, and to engage in moral practices, does not appear to require propositional judgements. This means that even under more demanding conceptions of trust than I previously endorsed, nonhuman animals can be said to enter into trusting relationships. It should be noted that highlighting the possession of moral emotions by nonhuman animals does not amount to a claim that they are moral agents. Rather, what it does is paint a richer picture of the affective dimensions of nonhuman animal relationships and so give reason to accord those relationships more weight in our moral deliberations.
When researchers or stockspersons habituate cows to handling though gentle touching, they are doing more than merely stimulating nerves to gain a biological response, they are fostering a relationship of trust. When handlers simulate gentle licking by stroking, they signal to a cow that they can be trusted not to exploit the vulnerability of their closeness. As gentle touch is repeated, cows learn that their handlers can be relied upon not to exploit that vulnerability into the future. Very often, touch is used to habituate calves to human contact by simulating the bonding practices of their mothers. In this way, trust is fostered by exploiting the mother-child bond and the cow's deep-seated need for sociability.
When an animal trusts, they become vulnerable. Vulnerability is intrinsic to trust, as is a sense of optimism or confidence that the object of trust means no harm. That is why when they trust their handlers cows become easier to handle. Habituation and calming makes transport to slaughter and through the abattoir easier. Indeed, this is seen as a key reason for doing it. Ultimately, through the exploitation of trust, harms become easier to perpetrate (Cooke, 2019). Touch is thus turned from expression of love, care, trust, and friendship into a systematised tool for exploitation. In this way, nonhuman animals are robbed of their capacity to resist through an act of deception and become willing victims. 6

Caring
One way to think about what is at stake here is to consider what an ideal relationship between animal and handler would look like, and the principles that would govern it. Since touch acts as a bridge to establish affective connections, creating a relationship between handler and animal, a natural way to think about this is in terms of care. Joan Tronto defines care as the activities done to look after the world so that we can live our best possible lives (Tronto, 1995, p. 142). Tronto argues that the process of care can be analysed in terms of virtues, each mapping on to a particular phase of care. A good giver and receiver of care will exhibit virtues of attentiveness, responsibility, competence, and responsiveness (ibid). Furthermore, these virtues will also manifest in political and social institutions where care takes place. In the case of radically vulnerable and dependent nonhuman animals, good care has to be attentive to what it is like for the animal themselves, and to providing for their needs because they are their needs (c.f. Tronto, 1995, p. 146). The virtue of attentiveness governs how we care about others. To truly care about another being we must prioritise or centre their needs in some way. A truly caring agent does not sacrifice the wellbeing of the object of their care for economic gain, nor do they deceive them in order to enable harm to be done to them. As Tronto observes, in order to be done well, caring must begin with a proper account of those needs. For farmed animals, this must include the full range of properly realised relational goods. Gentle touching should thus be regarded not as mere management technique but as a moral encounter, one that carries far greater responsibilities than is commonly thought. Instrumentalising the relationships between friends and family, and mimicking the behaviours used to foster and maintain those relationship for that purpose, fails to express virtues of care, as well as wronging the animals themselves. Similarly, the structures that justify techniques like gentle touching demonstrate a lack of care about, and a failure to take care of, nonhumans animals owed care.

Motive and Intent
One potential response to my argument so far might be to try to show that the motivations of those who try to reduce the impact of welfare harms is good, and that this ought to temper our judgements of them. It seems likely that as least some of those who try to reduce welfare harms, particularly, for example, veterinarians, and potentially some stockpersons too, will be primarily motivated by consideration for the nonhuman animals. It is also probably that at least some of those involved in gentle touching research act out of genuine consideration for the animals themselves. A charitable reading of the literature on the subject might interpret the continual highlighting of instrumental benefits to farmers as representing an attempt to persuade those who would otherwise be unmoved by animal suffering to nevertheless reduce the harms they do.
Another possible response might be to point to the fact that animal agriculture is likely to be a feature of our world for some considerable time to come. Given the world as it is, and the reality that animals will continue to be harmed in serious and systematic ways for the foreseeable future, perhaps we ought to regard methods such as gentle touching in a purely positive light. Surely, acts that reduce the amount of harm that would otherwise have been caused are good, particularly if carried out with this intention? Along similar lines, we will probably want to distinguish those who act to ameliorate or mitigate harms from those who are responsible for them. Amongst that group, at least some who are involved in practices designed to reduce harms may see their actions as a means of resisting wrongs present in animal agriculture.
To evaluate these sorts of cases, we can draw upon the lessons from Rob Garner's discussion of achieving justice for nonhuman animals. In A Theory of Justice for Animals, Garner uses a Rawlsian framework to show that actions taken to secure a more just state of affairs for nonhuman animals need to meet three conditions. First, even if they secure improvements, those improvements must not at the same time cut-off further progress. Second, the steps taken towards a better world must themselves be morally permissible. Third, attempts made to secure justice must be politically possible. These conditions circumscribe a range of morally permissible means of making progress from non-ideal towards ideal circumstances.
Garner argues that alleviating suffering without giving much weight to life or liberty is an acceptable step because it ends the worst harms (Garner, 2013, p. 135). This, Garner argues, offers an acceptable compromise between existing norms around animal treatment and the demands of animal rights advocates. As I have shown, it is not as straightforward to separate suffering, life, and liberty as Garner assumes. This means that the conditions for permissible steps are more stringent than he realises. Even without that being the case, Garner's argument suggests that using animals to benefit humans in ways that cause suffering, even if that suffering is reduced somehow, is impermissible. Whilst acts taken to reduce harms are politically possible, they do not eliminate suffering. Worse, they create a framework in which a certain level of suffering is normalised and rendered legitimate, hindering further steps towards the ideal (Garner, 2013, p. 91). Thus, even if we conclude that techniques like gentle touching are sometimes permissible, agents who involve themselves in structures of farming still need demonstrate that they recognise its wrongness. Silence and inaction regarding the structural and individual wrongs risks tacit endorsement and perpetuation of harm and wrongdoing. Additionally, whilst we can distinguish those who harm from those who act to reduce it, direct harms, along with the enabling of them, remain impermissible. Someone who reassures an animal through gentle touch as they guide them into a transport or chute for slaughter is complicit in serious harm and thus blameworthy both for the harm and for the use of deception and betrayal to secure it. Whilst may be to make a comparative judgement against the harms caused by using more violent methods, the fact that one method is worse than another does not thereby make its alternative good.

The Welfare Paradigm
Given the possibility of mixed motivations and contextual considerations, we have reason not to rush to judgement in all cases. Nevertheless, it remains important to draw attention to morally salient features of farming and welfare practices that have hitherto been overlooked. In order to do so, it is helpful to understand some of the processes that lead even well-motivated agents to fail to care in the right sorts of ways. Part of the explanation for this lies in the dominance of welfarist approaches to animal ethics.
Welfarist approaches represent the prevailing orthodoxy regarding our treatment of nonhuman animals, and they dominate approaches to farm animal wellbeing. Welfarism is the ethical position that nonhuman animals are owed something as a matter of morality, and that therefore their interest carry some weight in moral deliberation. Furthermore, their interests count for the sake of the animal itself. Kicking a cow is at least partly wrong because the cow is hurt by the kick, not just because the cow is someone's property or because the kick is expressive of a cruel character. At the same time, however, those interests are not considered sufficiently weighty as to grant strong moral protections such as rights, and count for far less than human interests. This means that animal interests do not defeat or exclude other considerations for action. Rather, under the welfare ethic, their interests can be weighed against other sorts of considerations, both moral and non-moral: economic cost, cultural significance, convenience, efficiency, etc. Whether the benefit to humans is a permissible one is rarely considered, and if it is then the standard for permissibility is set low enough that even the trivial and easily substitutable interests of humans can trump the most serious interests of nonhuman animals. So long as sufficient benefits can be derived from harming, then harm is permitted. Welfarism rules out gratuitous and obvious cruelty but 12 Page 16 of 20 allows serious suffering if benefits to humans can be shown. Under a welfarist ethic, whether nonhuman interests rule out harming them depends upon balancing considerations that are heavily contextual, with the costs and benefits given their value according to local social norms. Quite what counts as a sufficient benefit to override a strong interest is somewhat mysterious, often varying according the role an animal has been categorised into. For example, the interest a dog has in not suffering is weighted differently if the dog has the status of a pet or a service animal than if they are a laboratory subject, wild, feral, or a food animal (Garner, 2013, Chap. 5). Welfarist arguments for improving the lives of nonhuman animals involve arguing that we ought to give greater weight to animal interests -for the sake of those animals-and lesser weight to competing considerations. The sacrifice of animal interests for the sake of producing some good, however, remains always on the table. Gentle touching is one welfarist practice among many, all designed to raise animal welfare standards and make the lives of animals used by humans less unbearable.
In contrast to welfarist approaches, animal rights positions rule-out sacrificing the vital interests of nonhuman animals without much stronger moral justification. From that perspective, it can seem puzzling that merely reducing the suffering associated with a deliberate and unnecessary harm can be regarded as a welfare improvement. What makes welfarist approaches seem positive is their reliance upon a false counterfactual condition. Reductions in harms are regarded as improvements in welfare because the set of alternative possibilities they are evaluated against does not include doing no harm in the first place. Instead, the range of possible harms to be ranked often stops at the point where not harming might prevent humans from gaining benefits they desire. In this way, a wrongful harm is turned from a blameworthy act into a praiseworthy one.
In nearly all of the literature on the use of touch in farm animal management, it is cast as a means of securing improved welfare for nonhuman animals. However, describing reductions in deliberate harms as benefits to those harmed is misleading. We would not suggest that punching someone and then offering them a soothing hug ought to be thought of as increasing their wellbeing. An evaluation of the hug needs to be considered in terms of its context, history, and the structures it occurs within. Welfarist approaches, like gentle touching, require us to set aside that context and evaluate only the consequences of an act in isolation. In the case of gentle touching, the pleasure gained by eating animal products trumps the interest those animals have in living and not suffering. However, for that suffering to be justified then it must be minimised as far as is practicably possible given constraints such as cost to the farmer, time, effort, etc. Hence, researchers have been keen to highlight their awareness of the time and labour costs of gentle touching and demonstrate its positive economic value in those terms (Probst et al., 2012, p. 47;Lürzel et al., 2015, p. 15). Strictly speaking then, approaches like gentle touching should thought of less of as securing welfare improvements and more as a form of partial welfare repair. At the heart of welfarist approaches is an ethical ambiguity. This ambiguity enables welfarist approaches to be critical of suffering whilst at the same time endorsing the systems that require it. Similarly, it enables the use of intimate, caring encounters to enable and ease acts of serious harm.
Elsewhere, I have argued that the cultural and historical contexts in which practices of animal exploitation occur can undermine an agent's ability to discern or give genuine consideration to the full range of moral options open to them (Cooke, 2017). The strength with which animal agriculture is embedded in social conditions makes it difficult to see past the welfarist paradigm, and this partly explains why the full range of counterfactual possibilities is not considered in welfarist evaluations. This means that the wrongness of the treatment of nonhuman animals in human care is not necessarily a deep-seated moral failing on the part of individuals and so may be partly excusable (c.f. Hadley, 2009). Nevertheless, if moral progress is to be made, humans will need to develop new sensitivities towards nonhuman animals and that requires more carefully attending to the moral features of our practices. In particular, there needs to be a realisation that we have hitherto undervalued both care and the importance of affective relationships of love and friendship between domesticated nonhuman animals.

Conclusion
To conclude then. Within contemporary farming systems, the importance to nonhuman animals of their relationships has wrongly been accorded little weight. At the same time, and with some degree of irony, researchers have been forced to acknowledge that relational harms justify reparative action. Techniques such as gentle touching have been developed for this purpose. Unfortunately, the primary intention manifest in their use has been to secure instrumental benefits for humans. In the process of ameliorating and mitigating relational harms, tactile expressions of intimacy and care have been used to secure and betray trust. In this way, these techniques demonstrate a failure to care for vulnerable others in the right way. For touch to be used in praiseworthy ways requires an appreciation for the relational needs and respect for the associative rights of nonhuman animals, an appreciation that is sadly often lacking.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.