Inefficacy, Inaction, and Individual Responsibility: An Analysis of Industrial Animal Agriculture in Consequentialism

Context:

I wrote this as one of my Bachelor's thesis in my philosophy undergrad. At the time of writing this paper, I just started to get pretty deep into EA and I think it shows in the themes that I talk about/the sources I use vs. the lack of correct/common terminology and nuance in some/these areas. The paper also went under some review for the Flourish Journal, later on. 

Ever since I wrote it, I have noticed how many ideas (or the framing of some ideas) have influenced various ways in which I think about moral responsibility and effective altruism, but in re-reading it, I always found too many things I don't strongly endorse anymore, find inaccurate, could have used better examples or made a stronger point. At this point I have also thought about this so much that I don't have a sense of whether any of what I wrote about is "new" enough to add value to discourse. It's been probably close to a year that I wanted to give this a revision with my current thoughts, added points, corrected details, etc. and at this point, I am afraid, I have to admit that I just won't get to it. I thus want to take the Draft Amnesty Week as an occasion to post it without overthinking it. I gave it a quick glance to correct some typos and formatting errors but intentionally didn't correct any of the flaws.

1. Introduction

Suppose that the practices of factory farming are morally wrong. Does it follow that the individual consumer is responsible for the harm caused by the industry? It certainly cannot always be shown that they are causally responsible for animal suffering. It often appears that a single person’s choice to abstain from animal products does not have any real-world consequences on the trillion-dollar industry that is industrial animal agriculture. A consumer then can claim that whilst they fully consider non-human animals to be moral patients, their purchase of animal products does not cause any harm and they are therefore not morally responsible for the treatment of farmed animals. This is called the “inefficacy argument” and can be raised in any problem resulting from collective action in which a single person’s actions are neither sufficient nor necessary for the harm caused. A notable advocate of this argument is Walter Sinnot-Armstrong, who used it in the context of individual responsibility and climate change.

This paper intends to explore the inefficacy problem of factory farming and animal product consumption as a way of exploring the position required to deny individual responsibility for factory farming. “Factory farming” will be used in this paper to describe any farm that involves the intense breeding, raising and confinement of non-human animals.^[1] Given the prevalence of actual factory farming within animal agriculture, the term “animal products” will be used to encompass any products coming from said conditions.

In chapter, “2. Problem Profile and Premises”, the issue of factory farming and the inefficacy problem arising from it is briefly summarized to lay out the implicit premises of the paper. In chapter “3. Sinnott-Armstrong Deconstructed”, Sinnot-Armstrong's argument is formalized and discussed with various attempts to adjust the argument in an effort to find a framework for moral responsibility that justifies individual consumption of animal products. A strictly consequentialist position is assumed to be necessary, but it brings up an issue of uncertainty and practical guidance. To address these issues, chapter “4. Probabilities and Expected Harm – Putting Theory to Practice” uses the framework of moral responsibility discussed in chapter 3 and integrates it with the expected value theory as well as a novel, ad hoc developed alternative called “Plausible Probability Theory” to analyze the implications for the consumption of animal products. Lastly, chapter “5. Indirect Harm, Collective Inaction and Academics” examines the role of doing and allowing harm, which is a related aspect that arises from the strict consequentialist position necessary to deny individual responsibility for factory farming – as shown in chapter three. It is argued that highly influential people such as academic philosophers have an even larger responsibility based on the harm they could prevent. 

2. Problem Profile and Premises

The discussion in this paper centers around the inefficacy problem in the context of factory farming. The implicit assumptions therefore are that (1) factory farming is problematic, (2) that the consumption of animal products on a collective level can be linked to factory farming and (3) that the consumption of animal products on an individual level cannot be as easily linked to factory farming. In this chapter, I will briefly summarize and argue for these three premises.

More than 70 billion land animals and considerably more aquatic animals are slaughtered per year. Most of these animals are held in conditions that are commonly referred to as “factory farming”.^[2] A 2021 survey by the Sentience Institute suggests that around 75% of respondents reported discomfort with the animal industry.^[3] While moral considerations of non-humans date back to Greek writer Plutarch, it was only in the 1970s that animal ethics gained traction within moral philosophy.^[4] Approaches defending non-human animals come from a variety of ethical theories. Whilst there have also been attempts to justify animal product consumption in certain circumstances, there is no notable attempt to justify the conditions of factory farming. After analyzing the most prominent positions within animal ethics, John Nolt concludes that “[...] there are many points on which most animal ethicists agree: [...] Factory farms cause unjustifiable levels of suffering [...]; we should withdraw support from them by excluding their products from our diets.”^[5]It should be noted that many farms that are not classically referred to as “factory farming” use practices that could still be considered immoral or at least morally questionable such as the separation from the mother cow and her calf in the dairy industry or the slaughter of male chicks in the egg industry. It should further be noted that almost every farm uses practices that would be considered immoral if one would take a non-speciesist position since the sheer objectification and premature slaughter of an animal would be widely condemned if the animal in question was human. The link between mass consumption and factory farming appears to be clear. Most of the commonly considered morally reprehensible practices within factory farming were developed in response to the growing demand for animal products by our society in large. However, this does not necessarily imply that individual consumption of animal products can be linked to factory farming. Considering the size of animal agriculture and the complexity of the production chain, the link between the purchase of a singular person rarely seems to make a relevant difference for any particular animal. Since the purchase of an animal product is typically neither sufficient nor necessary for the actual suffering of any given animal, one could argue that while factory farming is wrong and we have a collective responsibility not to fund it, there might be no individual responsibility for any one person to boycott the industry. This is analogous to the critique of mere participation, according to which (some) consequentialist frameworks cannot explain why it would be immoral to be part of a collective that does harm, as long as one is not causally responsible for the harm done. As these positions are sometimes criticized as excessively permissive, this paper assumes a consequentialist framework.^[6] 

3. Sinnott-Armstrong Deconstructed

The argument against individual responsibility has sparked a lot of discussion within the debate around climate change. Walter Sinnot-Armstrong made the case against such responsibility in his 2005 article “It’s Not My Fault” by pointing out how different principles for assigning moral responsibility either prove that a wasteful sports-car driver is not morally responsible for climate change or run risk of proving too much and are therefore implausible to begin with. In this chapter, Sinnot-Armstrong’s argument will be analyzed to identify a minimal position to deny individual moral responsibility for factory farming.

3.1 The Harm Principle

One of the least controversial principles suggested in Sinnot-Armstrong’s paper “It’s Not My Fault” is the “harm principle”, defined as: “We have a moral obligation not to perform an act that causes harm to others.”^[7]  His paper has since been discussed frequently and there have been several attempts to argue that wasteful driving is in fact morally wrong. Attempts to reconcile this issue are usually dependent on additional premises such as a virtue ethical position or concepts from moral mathematics and expected harm.^[8] The inefficacy argument in climate change can be analogously applied to factory farming.^[9] Rather than working with additional premises, this chapter intends to identify the minimal position one needs to hold to deny individual responsibility and what it would entail.

When we question whether we cause harm, usually we depend on concepts of sufficiency and necessity. An individual can be said to cause x, if and only if the individual’s actions were sufficient or necessary to the outcome x. In discussing his harm principle, Sinnot-Armstrong notes a situation in which someone is neither necessary nor sufficient for the harm caused, but is still considered morally responsible:

Imagine that it takes three people to push a car off a cliff with a passenger inside, and five people are already pushing. If I join and help them push, then my act of pushing is neither necessary nor sufficient to make the car go off the cliff.^[10]

He goes on to argue that what makes him morally responsible in this situation is the intention of doing harm and the fact that their act was unusual. The argument about the usualness of the action is more of a pragmatic than a moral argument. Sinnot-Armstrong suggests that judging “usual” actors (e.g. individuals causing an average amount of emissions) would result in more extreme actors (e.g. individuals causing an excessive amount of emissions) feeling less motivation to change. The factor of intentionality, however, is certainly subject to discussion. In response to this argument, one could either (1) consider intentionality as a relevant factor for moral responsibility, (2) reject it and attempt to identify alternative factors that make the car pushing morally wrong or one could (3) reject it and deny moral responsibility for the sixth car-pusher. I will discuss these three options in the following subchapters.

3.2 Intentionality and Foresight

In his paper “Climate Change, Individual Emissions, and Foreseeing Harm” (2017) Chad Vance argues for an extended concept of moral responsibility. Building on Sinnot-Armstrong, he argues:

An action is prima facie morally wrong if:

(i) it is a cause (or part of a cause) of a harm, i.e., if it is either a necessary or a sufficient condition (or both) for that harm or

(ii) the action is neither necessary nor sufficient for any harm, but is an action that makes one a member of a group whose actions collectively cause harm, and the agent either (a) acts with the intent to produce harm (as a means or end), or (b) does not intend harm, but strongly foresees that harm will result from the group’s action.^[11]

Vance removes the requirement of unusualness and includes the factor of strong foresight in the definition. He argues that insofar as intending harm is taken as a relevant factor for moral responsibility, strongly foreseeing harm would have to be a relevant factor as well. This equivalence of strongly foreseeing harm and intending harm appears to be uncontroversial in situations where the benefit of an action is minor or non-existent in comparison to the wrongness of the foreseen harm. Imagine the sixth car-pusher simply intended to get some light exercise in but was fully aware that the car would be pushed off the cliff. He would then, presumably, be just as responsible as someone who intended to push it off.^[12] Vance then adds more stipulations to avoid proving too much with his theory (i.e., covering too much ground by additionally assigning moral responsibility to actions we don’t typically consider problematic). He argues that (1) if the group of people benefiting from an action are the same as the ones who are harmed, (2) if the individuals are aware of the risks involved and consent to remain in the situation, and/or (3) if there are regulatory systems established to avoid harms as far as possible established, the prima facie wrongness of the action can be overridden.^[13] Since the benefit derived from consuming animal products is minimal for many people in the WEIRD countries (often mainly providing gustatory pleasure), and the three additional factors outlined by Vance do not apply either, anyone who does not depend on animal products for non-trivial purposes can be held morally responsible for the harm caused by factory farming.^[14][15] 

3.3 Rejecting Intention and Contributory Effects

Vance proves that if one considers harmful intention as a significant factor in assigning moral responsibility, by means of strongly foreseeing harm, the same moral responsibility can be assigned to the consumption of animal products, insofar as the consumer is aware of the harm caused by factory farming. Following his argument, to avoid this conclusion, one would have to discard intentionality as a factor and attempt to find a different relevant factor that distinguished the sixth car-pusher from the person consuming animal products.

One such distinction might be the total number of people involved. Intuitively, if we imagine a person contributing to a three-person job where five people are acting, it is very easy to imagine a possible world in which the person pushing the car does in fact make a relevant difference, since the surplus of people is low. Assume a different situation, in which it would take three people to push the car and there are already one thousand people pushing when our subject joins in. Contrast that with a situation in which there are only three people pushing the car and our subject is the fourth person involved. One might intuitively feel that in the thousand people scenario, the additional car-pusher is less morally responsible for the damage done than in the three people scenario. This can either be explained by the surplus or by a contributory effect explanation. According to the contributory effect explanation, the responsibility of the people involved is shared and therefore diminishes with every person pushing the car. The same applies in the surplus explanation, but the counting of individuals would start with the last person necessary (and thereby causally responsible). To identify which one appears more plausible, assume a different situation. Imagine it would take a hundred people to push the car and the subject is the 101st person to join. Now, in contrast, imagine it would take three people to push the car and the subject is, still, the 101st person to push it. The surplus interpretation would assign more responsibility to the 101st person in the first scenario than in the second, while the contributory effect explanation would consider these situations equal.

In the surplus approach the calculation is the same as the contributive effect calculation, but it tends to assign more responsibility to the people involved since the counting of people starts at a later stage. If the threshold of people needed is precisely met, everyone would bear full responsibility for the harm. If exactly 100 car-pushers push a car that takes 100 people to push, everyone is responsible for precisely one car going off the cliff. If, however, a 101st person joins in, the responsibility would diminish, and everyone would be responsible for e.g., half the damage.

Both theories have difficulties. Consider this argument against the contributory effect explanation, inspired by Derek Parfit, who adapted it from Tom Regan^[16]:

Assume that you and three other people are forced to push one of two buttons. If all four of you press Button A, 15 people will be tortured. You can, however, press button B and torture 4 people instead. Everyone before you has pressed button A – presumably hoping that if you press B, then only four people will be tortured in total, which is the least amount of people tortured possible in this situation. You calculate that if you join the group, you will be responsible for the suffering of 15 people divided by 4, which equals less than 4 people, so you decide to push button A as the fourth person.

Situations like this highlight how the focus of our intuition tends to be on marginal effects – meaning that we would in fact consider the actual difference that a person makes by either being sufficient or necessary to the harm caused.^[17] One could, analogous to Parfit, revise this argument by claiming that you would be additionally responsible for enabling the harm caused by the others. This, however, appears to either strip their responsibility away (making you responsible for all 15 people being tortured and them for none) or assume non-shareable responsibility, making each one of you responsible for the suffering of 15 people. In this case, a surplus approach would make all of you responsible for 15 people as well, but only for 7.5 if another person joins in, assigning you more responsibility than the contributive approach in any case. Either way, this concept of non-shareable responsibility would typically increase the harm caused by an individual as part of a collective and is therefore disadvantageous for a claim that attempts to deny individual responsibility.

Furthermore, sharing responsibility in general has been criticized by Michael Zimmerman in his 1985 paper “Sharing Responsibility”:

This [adding of 5 people more than necessary to cause harm] surely would not have diminished responsibility of any of the original ten [necessary to cause said harm]. (What an easy way ‘out’ that would be! Just invite a few more friends to participate.) And there seems to be no reason to think that there is necessarily anything which distinguishes, in a morally relevant way, any one of the fifteen from any other.^[18]

Whether one accepts the contributory effect interpretation or the surplus interpretation, both interpretations encounter further difficulties when taken into the context of factory farming. Dividing the total harm caused by factory farming by the number of people who consume animal products would still result in each person being morally responsible for the suffering and death of approximately 13 to 28 vertebrates per year.^[19] With the surplus interpretation, that result would be the best-case scenario (i.e., being one person behind the threshold), meaning that the responsibility ascribed by that interpretation tends to be larger than in the classical contributory effect explanation.

3.4 Rejecting Intention and Rejecting Responsibility

If one rejects both suggestions and is unable to find a relevant difference between the sixth car-pusher and the person who buys an animal product, rejecting the moral relevance of the intentionality whilst not holding the person who buys an animal product responsible, implies that the sixth car-pusher is also not morally responsible for the damage done. Thus, when Sinnot-Armstrong suggest that “[t]here are, admittedly, special circumstances in which an act causes harm without being either necessary or sufficient for that harm”^[20], one could simply reject that claim. The argument might go as follows:

P1: A sixth person pushing the car is the same as a thousandth person pushing the car. (This follows from denying the relevance of the number of people involved)

P2: A thousandth person pushing the car is prima facie not responsible for the damage done to the car. (This follows from prima facie rejecting the moral responsibility of a person purchasing animal products)

C: The sixth person pushing the car is prima facie not responsible for the damage done.

This line of argumentation is not necessarily implausible. If one takes a strictly consequential position, the morality of an action does only depend on its consequences. If it takes three people to push a car down a cliff, the third person then is completely and exclusively responsible for the damage done to the car, since they are the only person whose action has actual consequences.^[21] This position would view the different perspectives on shared responsibilities as mere heuristics. It should be noted however that this interpretation of consequentialism does not only reject these heuristics, but it also questions the criteria of sufficiency and necessity. The third person pushing the car is technically neither sufficient (since the car would have not been pushed off the cliff by this person alone) nor are they necessary, if more people join to push the car (assuming that the others would have joined even without the third person being present). The third person is only responsible because there is a type of causal link that can be traced to them. They did push the car and the car did fall merely as a result of their actions. This criterion can be called the mere causation criterion. One could of course reject this criterion and remain with only the sufficiency and the necessity criteria, which would imply different consequences. If for example a group would intend to push a hundred cars down the cliff, without mere causation, the only way in which anyone can be held responsible is if there are exactly three hundred people involved, because only then it could be claimed that someone was necessary for the harm caused (assuming it takes three people per car). Furthermore, said identifiable harm however would only be the last car since no one person is necessary (or sufficient) for any of the other 99 cars being pushed off the cliff. This would imply that there is a one in three chance that every single person is responsible for the damage of exactly one car being pushed off the cliff. The outcomes are similar to those of the contributive effect calculation, with the difference that the harm that someone is responsible for is multiplied by the number of people involved per unit of damage (in this case one unit of damage is one car and the number of people involved is three) and is only considered if the exact threshold needed for the damage is hit. 

Whether one accepts mere causation or not has different implications when it comes to individual consumer choices and factory farming. It is estimated that the largest producers in the poultry industry can respond to a set of approximately 900 chickens at a time.^[22], ^[23] Rejecting the mere causation criterion, would imply that if the total number of chickens being bought is a multiple of 900 (e.g. 1800 or 2700), every person who contributes to these purchases is responsible for exactly 900 chickens being raised and slaughtered. If the total number of chickens being bought is not a multiple of 900 however, no one would be held responsible.^[24] If on the other hand one accepts the mere causation criterion, regardless of the total number of consumers, every person purchasing the equivalent of every 900th chicken is responsible for the harm done to 900 chickens, but no one else is.

3.5 The Direct Harm Principle Summarized

Sinnot-Armstrong proposes three criteria in his direct harm principle: sufficient/necessary causation, harmful intentions, and unusual behavior. The latter two were proposed to justify assigning moral responsibility to a person who joins a five-person group in pushing a car off a cliff, under the assumption that it would only take three people to successfully push the car. The chapter discards the usualness of the action as a factor, leaving harmful intention to be analyzed in three possibilities: (1) accepting the criterion, which lead to also assigning moral responsibility in the case of factory farming; (2) rejecting it and proposing a different criterion to distinguish the sixth car-pusher the person consuming animal products, which has not yet been found; and (3) rejecting it and denying responsibility for the sixth car-pusher.

Furthermore, it has been shown what denying responsibility of the sixth car-pusher might entail, depending on whether one would accept the mere causation criterion. Either way, in this framework people purchasing animal products take enormous risks in their actions, since they might be responsible for the suffering and death of 900 birds. It is unclear however, what these risks imply for our moral judgment. The following chapter will discuss the role of probabilities and expected harm in practice.

4. Probabilities and Expected Harms – Putting Theory to Practice

In the previous chapter it has been shown that a minimal position on the harm principle in theory leads to a risk heavy position in practice. Whether one rejects mere causation or not, when purchasing chicken, one runs a 1 to 900 risk to be morally responsible for the suffering and death of 900 chickens. Theoretically, the strict consequentialist might argue that this has no significance, since probabilities are not a matter of actual consequences, but one of information. One could remain with the theoretical implications laid out in the previous chapter and conclude that, at least in retrospect, the purchase of a chicken is usually unproblematic but potentially disastrous in other, albeit rare, situations. This, however, is only relevant in retrospect. Insofar as an ethical position should give normative practical guidance, strictly consequentialist theories must work with probabilities to identify the best way to act. Uncertainty is an important consideration in practical ethics and there are different decision theories to address them. This chapter will discuss two ways of doing so and their implications.

4.1 Expected Value Theory

When discussing decision theories, two relevant factors seem intuitively uncontroversial. Assuming equal (potential) harm, a decision theory should account for a preference of lower risks over higher risks. Similarly, assuming equal risk, a decision theory should be able to account for a preference of lower potential harm over higher potential harm. One intuitively promising way of accounting for these is Expected Value Theory (EV). According to EV, the relevant factor for practical decision making is the harm potentially done multiplied by the probability of it occurring.^[25] Thus, the 1 to 900 probability of causing harm to 900 chickens is equivalent to a 100% probability of causing harm to one chicken. Usually, EV would include the possible benefit and its probabilities as well as all other possible outcomes and then put these factors together to decide. Since there are only two relevant possible outcomes to a chicken purchase (i.e. the purchase causing a new set of chickens being processed or not) and the benefits are minimal in a western society, the focus of this paper will be on the harm caused.

A similar issue has been raised in response to Sinnot-Armstrong. He claims that the risk principle, defined as “We have a moral obligation not to increase the risk of harm to other people”, is too restrictive when taken literally and too vague when distinguishing significant from insignificant risks.^[26] Naturally, EV can counter the vagueness argument by providing a structural approach and calculative approach. Here, the expected harm, the suffering and death of one chicken, is weighed against the benefit of eating it (e.g., gustatory pleasure). On the other hand, boiling water in a kettle appears to have such a low probability of meeting the threshold for a climate catastrophe that the expected harm of doing so is minor. Avram Hiller raises a similar point. He argues that “it is prima facie wrong to perform an act which has an expected amount of harm greater than another easily available alternative”.^[27] He refers to a calculation done by John Nolt according to which driving a car for a twenty-five-mile distance amounts to causing “¼ of a day’s worth of serious harm” to one or two people.^[28]

This decision theory is intuitively plausible and commonly used in daily practice but seems to run into issues when it comes to tiny probabilities and enormous values. For example, consider this issue presented by Nick Beckstead and Teruji Thomas:

On your deathbed, God brings good news. Although, as you already knew, there’s no afterlife in store, he’ll give you a ticket that can be handed to the reaper, good for an additional year of happy life on Earth. As you celebrate, the devil appears and asks, ‘Won’t you accept a small risk to get something vastly better? Trade the ticket for this one: it’s good for 10 years of happy life, with probability of 0.999.’ You accept, and the devil hands you a new ticket. But then the devil asks again, ‘Won’t you accept a small risk to get something vastly better? Trade that ticket for this one: it is good for 100 years of happy life – 10 times as long – with probability 0.9992 – just 0.1% lower.’ An hour later, you’ve made 50,000 trades. (The devil is a fast talker.) You find yourself with a ticket for 1050,000 years of happy life that only works with a probability of 0.99950,000, less than one chance in 1021.^[29]

In their paper, Beckstead and Thomas highlight a problem common in many types of value theory that works with uncertainty. According to EV, every single one of your trades was better than the previous one, but intuitively it seems that 1050 000 years of happiness with a 1 to 1021 chance is considerably worse than e.g. 100 years of happiness with a 99% chance.^[30] 

Applying EV would conclude that the individual responsibility for the purchase of an animal product is in exact correspondence with the product in question (e.g. the purchase of one chicken caused the harm of one chicken). If one were to reject EV, one could either ignore decision theory altogether or attempt to develop/find an alternative approach in which the consumption of animal products remains unproblematic.

4.2 Plausible Probabilities

An approach to justify animal product consumption would have to justify disregarding the probabilities with which one would cause harm. It might be claimed that, at least in the case of the chicken, a 1 to 900 chance is so small that it is not relevant in day-to-day life, regardless of the harm it causes. This approach can be called the Plausible Probability approach. Assume that there is a threshold probability p for which it is plausible enough for someone to be expected to know that a given outcome would occur, thereby making the acting subject blameworthy. Under this approach, any action with a probability below p is not to be blamed. In the first few trades with the devil, you risked your life – assumed to be something of great value – but since the probability of you dying was only 0.1%, it would not seem plausible for you to be blamed for taking the risk (especially considering the benefit). However, at some point along the trades you would have taken a risk of dying with which it seems reckless to do so (regardless of the potential benefits).

The probability of someone causing the suffering and slaughter of 900 chickens might be 1 to 900 – assuming they buy a whole chicken – but if one were to do so 624 times, the probability of this threshold being met at least once would be more than 50%.^[31] In the case of 900 chickens, this plausible probability might be considerably lower than 50%. If p were, for example, 0.02, i.e., that the plausible probability for someone to be considered blameworthy is 2% or higher, it would take approximately 18 chickens purchased – the average US-Citizen consumes around 20-25 chickens per year - to be considered blameworthy in this framework.^[32], ^[33] Similar to the problem of collective and individual responsibility, whilst the collective purchases over a given timeframe might be considered blameworthy, assessing individual purchases might prove difficult.^[34] In contrast to the collective action problem however, the assessment that a person does something morally wrong in their lifetime/in a given time frame rather than in a singular act is not as comforting as the assessment coming from collective responsibility problems (i.e. being part of a group that collectively does something wrong rather than doing something wrong as an individual). Overall, this framework may seem intuitively plausible in moral reasoning, but it fails to justify animal product consumption over an extended period/beyond a certain amount. Furthermore, the identification of p proves to be difficult, especially in cases in which the harm caused is so large that it may be challenging to accurately comprehend its extent.

4.3 Probabilities Summarized

It has been demonstrated that a classical decision theory such as EV would suggest that individuals have a moral responsibility to abstain from consuming animal products - assuming one considers the harm done to non-human animals in factory farms as morally reprehensible. Furthermore, rejecting EV itself might be plausible, but an alternative approach such as one of plausible probabilities, which attempts to ignore negligible risks in decision making, is not guaranteed to justify animal product consumption either, since the risk of meeting a given threshold and being responsible for e.g., the suffering and death of 900 chickens is not negligible in the long run. To argue that the consumption of animal products is morally unproblematic, one could simply reject decision theories. However, simply rejecting decision theories is not a viable solution as it would leave individuals without systematic practical guidance on how to make moral decisions. This lack of guidance, however, is not the only issue that arises in remaining purely consequentialist. By disregarding intentions, consequentialists will also have to deal with the implications from inaction, which is particularly problematic for individuals with high influence, such as academics like Walter Sinnot-Armstrong.

5. Indirect Harm, Collective Inaction and Academics

As shown in the previous chapters, one could deny individual responsibility by maintaining a strictly consequential framework, but one would still run a rather large risk in consuming animal products for an extended period of time/in larger amounts. However, there is a different principle which the inefficacy argument has to face: in remaining strictly consequentialist and ignoring intentionality, there seems to be no justification left to distinguish between doing and allowing harm. This would imply that not only one’s actions, but also one’s inactions are subject to moral evaluation. Collective inaction and shared responsibility has been discussed by Larry May (1990), he writes:

For inaction to be collective there must be some sense in which the group failed to act. There must be practical plausibility to the counterfactual claim that the group could have done otherwise. If this condition is met, and if it is determined that the group should have acted, then the group is collectively responsible for harms that result from its inaction.^[35]

May goes on to argue that the responsibility each member of the group has is dependent on their potential influence on other members, meaning that people with high influence (e.g. people with good social and leadership skills) bear more responsibility for collective inaction than people with low potential influence.^[36] In the context of factory farming and due to globalization, people in high influence positions, such as writers, academics, politicians or celebrities, would bear more responsibility for not taking action to address the issue. Not considering shared responsibilities, as could be implied by a strictly consequentialist position, the argument might go as follows:

P1: x is morally responsible for a harmful outcome y if x’s actions are sufficient or necessary for y (or if y is a result of x’s actions by mere causation).

P2: Since only the consequences are relevant to the morality of an action, every inaction is also an action (e.g., one could actively fail to prevent harm from happening by not doing anything to prevent it).

C: x is responsible for inactions that are sufficient or necessary for y.

Insofar as it is possible that x could have acted in a way to prevent y, x’s inactions were sufficient to cause y.^[37] In terms of factory farming, this means that if it is possible for someone to influence enough people to not buy the equivalent of 900 chickens (i.e. making 40 people vegan for a year) and they fail to do so, that person is responsible for the suffering and death of 900 chickens. This, of course, creates enormous pressure on anyone who can influence other people to do as much as they can, since the argument demands action to the fullest extent possible. If someone were able to prevent the purchase of 1800 chickens but only does so for 1700, they remain responsible for the second set of 900 chickens.

When addressing the indirect harm principle (“The indirect harm principle: We have a moral obligation not to perform an act that causes harm to others indirectly by causing someone to carry out acts that cause harm to others.”)^[38] Walter Sinnot-Armstrong claims that most people are not as influential as they think; that wasteful driving is not habit-forming; that wasteful driving won’t undermine his other efforts in environmentalism; and that the scale of climate change is too big for him to cause it, even “with a little help from [his] friends”.^[39] Evidence suggests that these claims are not as apparent as they seem, especially when it comes to meat consumption and when the harm to prevent isn’t “the scale of climate change” (i.e., all of factory farming), but the smallest possible unit of harm preventable (i.e., 900 chickens). In contrast, research repeatedly indicates that a driving factor of meat consumption is its normalization and that social influence on consumer choice and behavior is widely neglected and underestimated.^[40], ^[41] Factors that would fall under the category of normalization appear as common issues for former vegans.^[42] One way of modeling the effect of social influence is Granovetter’s Threshold Model of Collective Behavior (1978) according to which, individual behavior is dependent on the strength of one’s intention and the number of “supporters”, i.e., people who portray the intended behavior, that one is aware of (with varying degrees of influence depending on the relationship between the individuals). Since every person has different thresholds, it appears to be plausible that a single individual openly communicating behavior change starts something like a snow-ball effect.^[43], ^[44] The impact of this effect is inversely correlated with the amount of people already on board with the spreading idea, which indicates high potential impact for early stage advocacy. Therefore, individual consumption of animal products has relevant effects beyond the direct causal chain of consequences within the market. Furthermore, individual consumption doesn’t only influence others, but also shows significant effects on the consumer. Studies show that participants who consume meat products as snacks during the experiment ascribe less sentience to non-human animals than those who had nuts.^[45] A 2019 report by Trent Grassian suggests that people who adhere to stricter dietary goals - e.g., going vegan - are more likely to succeed in their diet than people who set more flexible goals - e.g., reducing meat consumption.^[46]

In summary, even with a strictly consequentialist position, one would remain responsible for all harm done to non-human animals that could have been prevented - the scale of which is especially high for people with a lot of influence. Furthermore, due to the importance of social influence in consumer choices and behavior, it is plausible that most people are able to prevent some harm done to animals and bear some degree of responsibility as well.

6. Conclusion

In an attempt to defend the consumption of animal products while acknowledging the harm caused by factory farming, it appeared that following the harm principle offers a promising approach. Since that principle didn’t seem sufficient to account for other cases in which one would commonly assign responsibility (like the sixth car-pusher), it had to be modified, but every modification seemed to also assign responsibility to the consumption of animal products. Alternatively, one could deny responsibility in said other cases. This, however, would only work in retrospective, since every normatively guiding view would need an appropriate decision theory to act under uncertainty. It has been argued that a classical decision theory, such as EV, would condemn the consumption of animal products. Furthermore, in an attempt to resolve issues around tiny probabilities, an alternative decision theory has been introduced, which would still assign moral responsibility to the harm done to non-human animals but would only do so after a threshold of animal products consumed. It has furthermore been shown that in denying any modification to the harm principle, one remains responsible for any amount of harm caused by factory farming that could have been prevented. This especially applies to academics, writers, and other highly influential people, but is also a factor to consider for people without extraordinary social influence and status, since social influence and normalization are significant factors for the consumption of animal products.

Bibliography

Ali Ladak and Jacy Reese Anthis, “Animals, Food, and Technology (AFT) Survey: 2021 Update,” PsyArXiv, 2022, https://doi.org/10.31234/osf.io/3v6ku. 

Avram Hiller, “Climate Change and Individual Responsibility,” The Monist 94, no. 3 (2011), 349-368.

Chad Vance, “Climate Change, Individual Emissions, and Foreseeing Harm,” Journal of Moral Philosophy 14, no. 5 (2017), 562 – 584.

Cheryl Abbate, “Virtues and Animals: A Minimally Decent Ethic for Practical Living in a Non-ideal World,” Journal of Agricultural and Environmental Ethics 27, no. 6 (2014), 909 – 929.

Derek Parfit, Reasons and Persons, reprinted, corrected edition. (Oxford University Press, 1987).

Elliot Aronson, Timothy Wilson, and Samuel Somers, Social Psychology, 10th ed. (New York, NY: Pearson, 2019).

“Farmed Animal Fundamentals," Faunalytics, accessed January 27, 2023, https://faunalytics.org/fundamentals-farmed-animals/.

“Global Animal Slaughter Statistics & Charts: 2022 Update,” Faunalytics, accessed May 22, 2023, https://faunalytics.org/global-animal-slaughter-statistics-charts-2022-update.

“Global Farmed and & Factory Farmed Animals Estimates”, Sentience Institute, accessed May 22, 2023, https://www.sentienceinstitute.org/global-animal-farming-estimates.  

Jared Piazza, Matthew Ruby, Steve Loughnan, Mischel Loung, Juliana Kulik, Hanna Watkins, and Mirra Seigerman. “Rationalizing meat consumption. The 4Ns,” Appetite 91 (2015), 114-128. https://doi.org/10.1016/j.appet.2015.04.011.

Jessica Nolan, Wesley Schultz, Robert Cialdini, Noah Goldstein, and Vladas Griskevicius, “Normative Social Influence is Underdetected,” Personality and Social Psychology Bulletin 34, no. 7 (2008), 913-923.

Joakim Sandberg, “My Emissions Make No Difference,” Environmental Ethics 33, no. 3 (2011), 229–248.

John Nolt, Environmental Ethics for the Long Term: An Introduction (London/New York: Routledge, 2015).

Kathryn Asher et al., “Study of current and former vegetarians and vegans: Secondary findings,” 2016, 13, https://faunalytics.org/wp-content/uploads/2016/02/Faunalytics-Study-of-Current-and-Former-Vegetarians-and-Vegans-–-Secondary-Findings-.pdf.

Larry May, “Collective Inaction and Shared Responsibility,” Noûs 24, no. 2 (1990), 269-277.

Mark Granovetter, “Threshold Models of Collective Behavior,” American Journal of Sociology 83, no. 6 (1978): 1420-1443.

Matthew Halteman and Steven McMullen, “Against Inefficacy Objections: The Real Economic Impact of Individual Consumer Choices on Animal Agriculture,” Food Ethics (2018), Pre-Publication Draft, https://www.semanticscholar.org/paper/Against-Inefficacy-Objections%3A-the-Real-Economic-of-McMullen-Halteman/8c995c4301d9565d44e3127d8393854b72728033.

Michael Zimmerman, “Sharing Responsibility,” American Philosophical Quarterly 22, no. 2 (1985), 115 – 122.

Nick Beckstead and Teruji Thomas, “A paradox for tiny probabilities and enormous values,” Global Priorities Institute, GPI Working Paper No. 7-2021, July 2021, https://globalprioritiesinstitute.org/nick-beckstead-and-teruji-thomas-a-paradox-for-tiny-probabilities-and-enormous-values/.

“Numbers of farmed fish slaughtered each year,” fishcount.org.uk, accessed May 22, 2023, http://fishcount.org.uk/fish-count-estimates-2/numbers-of-farmed-fish-slaughtered-each-year.

Peter Singer, Animal Liberation, Translated by Claudia Schorcht. 2nd ed. (Germany: Harald Fischer Verlag, 2016).

Steve Loughnan, Nick Haslam, and Brock Bastian, “The role of meat consumption in the denial of moral status and mind to meat animals,” Appetite 55, no. 1 (2010), 156-159, https://doi.org/10.1016/j.appet.2010.05.043.

Trent Grassian, “Meat Reduction & Vegan Promotion. Summary report: Motivators, barriers, and dietary changes from the largest study of participants in meat reduction & vegan campaigns”, 9, https://forusallsite.files.wordpress.com/2019/03/grassian-reductionandveganpromotion-v2.pdf.

USDA National Agricultural Statistics Service. Poultry Slaughter. 2020 Summary. February 2021, https://downloads.usda.library.cornell.edu/usda-esmis/files/pg15bd88s/f1882d39g/j6731z19s/pslaan21.pdf.

Walter Sinnott-Armstrong, “It’s Not My Fault,” Perspectives on Climate Change: Science, Economics, Politics, Ethics, edited by Walter Sinnott-Armstrong and Richard Howarth, (Amsterdam/San Diego: Elsevier, 2005), 285 – 307.

1. ^{^}

    It is difficult to formally define factory farms, since criteria like “intense confinement” can be considered on a sliding scale. However, the purpose of this paper is to explore an argument that presupposes someone considering that a given farm is immoral by their own standards. Therefore, it might be neither necessary nor particularly meaningful to provide a formal definition.

2. ^{^}

    "Farmed Animal Fundamentals," Faunalytics, accessed January 27, 2023, https://faunalytics.org/fundamentals-farmed-animals/.

3. ^{^}

    Ali Ladak and Jacy Reese Anthis, “Animals, Food, and Technology (AFT) Survey: 2021 Update,” PsyArXiv, 2022, https://doi.org/10.31234/osf.io/3v6ku.

4. ^{^}

    Peter Singer, Animal Liberation, Translated by Claudia Schorcht. 2nd ed. (Germany: Harald Fischer Verlag, 2016), 227.

5. ^{^}

    John Nolt, Environmental Ethics for the Long Term: An Introduction (London/New York: Routledge, 2015), 157.

6. ^{^}

    Cheryl Abbate, “Virtues and Animals: A Minimally Decent Ethic for Practical Living in a Non-ideal World,” Journal of Agricultural and Environmental Ethics 27, no. 6 (2014), 914.

7. ^{^}

    Walter Sinnott-Armstrong, “It’s Not My Fault,” Perspectives on Climate Change: Science, Economics, Politics, Ethics, edited by Walter Sinnott-Armstrong and Richard Howarth, (Amsterdam/San Diego: Elsevier, 2005), 289.

8. ^{^}

    Joakim Sandberg, “My Emissions Make No Difference,” Environmental Ethics 33, no. 3 (2011), 229–248; Avram Hiller, “Climate Change and Individual Responsibility,” The Monist 94, no. 3 (2011), 349-368.

9. ^{^}

    One difference between these two might be the argument from imperceptible harm, since there is no such assumption as purchasing an animal product causing a tiny amount of harm to anyone.

10. ^{^}

     Sinnot-Armstrong, “It’s not my Fault,” 289.

11. ^{^}

     Chad Vance, “Climate Change, Individual Emissions, and Foreseeing Harm,” Journal of Moral Philosophy 14, no. 5 (2017), 572.

12. ^{^}

     Ibid., 570.

13. ^{^}

     Ibid., 576-577.

14. ^{^}

     Ibid., 582.

15. ^{^}

     (1) Humans benefit from animal product consumption, but not-human animals are harmed, (2) non-human animals didn’t consent to being farmed, (3) in industrial agriculture, methods are optimized for profit and efficiency rather than animal welfare.

16. ^{^}

     Derek Parfit, Reasons and Persons, reprinted, corrected edition. (Oxford University Press, 1987), 67-69.

17. ^{^}

     Sandberg, “My Emissions Make No Difference,” 235.

18. ^{^}

     Michael Zimmerman, ”Sharing Responsibility,” American Philosophical Quarterly 22, no. 2 (1985), 119.

19. ^{^}

     The number of farmed land animals slaughtered per year were taken from “Global Animal Slaughter Statistics & Charts: 2022 Update,” Faunalytics, accessed May 22, 2023, https://faunalytics.org/global-animal-slaughter-statistics-charts-2022-update/. The number for farmed fish slaughtered were taken from “Numbers of farmed fish slaughtered each year,” fishcount.org.uk, accessed May 22, 2023, http://fishcount.org.uk/fish-count-estimates-2/numbers-of-farmed-fish-slaughtered-each-year. The estimations for the prevalence of factory farming were taken from “Global Farmed and & Factory Farmed Animals Estimates”, Sentience Institute, accessed May 22, 2023, https://www.sentienceinstitute.org/global-animal-farming-estimates.  

20. ^{^}

     Sinnot-Armstrong, “It’s Not My Fault”, 289.

21. ^{^}

     Assuming at least one more person joins. Otherwise, everyone involved would be necessary for the harm caused and therefore morally responsible. 

22. ^{^}

     Matthew Halteman and Steven McMullen, “Against Inefficacy Objections: The Real Economic Impact of Individual Consumer Choices on Animal Agriculture,” Food Ethics (2018), 26, https://www.semanticscholar.org/paper/Against-Inefficacy-Objections%3A-the-Real-Economic-of-McMullen-Halteman/8c995c4301d9565d44e3127d8393854b72728033.

23. ^{^}

     Other industries might be considerably smaller but also respond slower to changes in consumer behavior. It therefore seems unlikely that the chance of causing suffering is even lower than 1 to 900. Since attempts to reconcile with this risk (cf. Chapter 4.2) benefit from higher risks and largely ignore the harm difference in these large scales (i.e. if it were a 1 to 300 risk for 300 chickens, EV – as discussed in Chapter 4.1. – wouldn’t change, but the approach from plausible probabilities – as discussed in 4.2 - would conclude a worse outcome), the 1 to 900 probability is assumed to be the best-case scenario for the person defending the consumption of animal products.

24. ^{^}

     It could be argued that the farmers’ and slaughterhouse workers’ actions are sufficient for the harm they cause, since if they were to do the same action (raising/slaughtering the bird) without anyone else existing, the harm still occurs. This is debatable however, since one could take the position that they would in fact not raise/slaughter the animals in question if no one would buy any poultry. It can also be argued that the core decision makers are responsible, since their call is necessary for the additional chickens being raised and slaughtered. This too is debatable since they would likely just be replaced.

25. ^{^}

     Nolt, Environmental Ethics for the Long Term, 44.

26. ^{^}

     Sinnot-Armstrong, “It’s Not My Fault,” 293-294.

27. ^{^}

     Hiller, “Climate Change and Individual Responsibility,” 353.

28. ^{^}

     Ibid., 357.

29. ^{^}

     Nick Beckstead and Teruji Thomas, “A paradox for tiny probabilities and enormous values,” Global Priorities Institute, GPI Working Paper No. 7-2021, July 2021, 1-2, https://globalprioritiesinstitute.org/nick-beckstead-and-teruji-thomas-a-paradox-for-tiny-probabilities-and-enormous-values/.

30. ^{^}

     The authors go on to argue that both a timid approach (i.e., avoiding certain risks if the harm is too large) and a reckless approach (i.e., ignoring intuition and following expected values unconditionally) have difficulties. It would be beyond the scope of this paper to discuss them in detail, this example simply intends to showcase that one could somewhat reasonably reject EV.

31. ^{^}

     The formula to calculate the required amount of actions x with a probability of 1/n  for any desired plausible probability p would be: x=ln (1-p)/ ln⁡(1-1/n).

32. ^{^}

     USDA National Agricultural Statistics Service. Poultry Slaughter. 2020 Summary. February 2021, 5, https://downloads.usda.library.cornell.edu/usda-esmis/files/pg15bd88s/f1882d39g/j6731z19s/pslaan21.pdf.

33. ^{^}

     It should be noted that the numbers taken here do include turkeys.

34. ^{^}

     One could argue that - analogous to the purely consequentialist criteria in the previous chapter and assuming 2% as plausible probability - a person would be free to buy 17 chickens but would do something morally disastrous in buying their 18th.

35. ^{^}

     Larry May, “Collective Inaction and Shared Responsibility,” Noûs 24, no. 2 (1990), 273.

36. ^{^}

     Ibid., 275.

37. ^{^}

     Depending on how x’s (in)actions are framed, they could also be considered necessary. When the (in)action z is framed as “fails to prevent harm”, it follows that in every possible world in which not z (“doesn’t fail to prevent harm”), y (preventable harm) doesn’t occur. 

38. ^{^}

     Sinnot-Armstrong, “It’s Not My Fault,” 291.

39. ^{^}

     Ibid., 292.

40. ^{^}

     Jared Piazza, Matthew Ruby, Steve Loughnan, Mischel Loung, Juliana Kulik, Hanna Watkins, and Mirra Seigerman. “Rationalizing meat consumption. The 4Ns,” Appetite 91 (2015), 114-128. https://doi.org/10.1016/j.appet.2015.04.011; Jessica Nolan, Wesley Schultz, Robert Cialdini, Noah Goldstein, and Vladas Griskevicius, “Normative Social Influence is Underdetected,” Personality and Social Psychology Bulletin 34, no. 7 (2008), 913-923; Elliot Aronson, Timothy Wilson, and Samuel Somers, Social Psychology, 10th ed. (New York, NY: Pearson, 2019), 10.

41. ^{^}

     Especially factors following from normalization such as the perceived necessity and naturalness of meat consumption are common rationalizations.

42. ^{^}

     Kathryn Asher et al., “Study of current and former vegetarians and vegans: Secondary findings,” 2016, 13, https://faunalytics.org/wp-content/uploads/2016/02/Faunalytics-Study-of-Current-and-Former-Vegetarians-and-Vegans-–-Secondary-Findings-.pdf.

43. ^{^}

     Mark Granovetter, “Threshold Models of Collective Behavior,” American Journal of Sociology 83, no. 6 (1978): 1420-1443.

44. ^{^}

     Assume that my behavior change (and advocacy) impacts five people to change their diet. It follows not only that every person who is influenced by the people whose behavior I changed is indirectly part of my influence, it also follows directly that every person whose threshold for behavior change was ‘the number of people before my influence plus five or lower’ are influenced by my advocacy as well. 

45. ^{^}

     Steve Loughnan, Nick Haslam, and Brock Bastian, “The role of meat consumption in the denial of moral status and mind to meat animals,” Appetite 55, no. 1 (2010), 156-159, https://doi.org/10.1016/j.appet.2010.05.043.

46. ^{^}

     Trent Grassian, “Meat Reduction & Vegan Promotion. Summary report: Motivators, barriers, and dietary changes from the largest study of participants in meat reduction & vegan campaigns”, 9, https://forusallsite.files.wordpress.com/2019/03/grassian-reductionandveganpromotion-v2.pdf.