Key Assumptions and Scope

Several assumptions and scope limitations guide our discussion. First, we assume that all sentient beings have moral status. We understand this to mean that the interests of all sentient beings matter morally, and that they matter for the sake of the beings themselves. We understand sentience as the ability to have positive (pleasant, attractive) and/or negative (unpleasant, aversive) experiences, either physical or emotional, in response to one’s environment. Henceforth, all references to animals used in research should be understood to refer to sentient animals only. Due to this assumption, our arguments do not address those who deny that animals have moral status, or who believe that our obligations to animals are merely indirect—that is, that we have obligations regarding animals only to the extent that our treatment of animals bears on human interests.^{Footnote 11,Footnote 12} Although we assume that all sentient beings have moral status, we do not make any assumptions regarding the relative moral status of human beings and nonhuman animals. Some argue that although all sentient beings have moral status, human beings have higher moral status than nonhumans.^{Footnote 13} We intend our analysis to be neutral between this view and the view that all sentient beings have equal moral status.

Second, we limit our discussion to research with animals that is designed to benefit human beings and does not offer any prospect of therapeutic benefit to the animals themselves. For example, we do not consider veterinary research that is designed to assess a new treatment that might benefit the participants. The reason for this focus on “nonbeneficial” research is that it is unclear, even with human subjects, whether investigators should respect the preferences of individuals who are incapable of informed consent and do not recognize that their participation in a study offers them the prospect for important personal benefit that is not otherwise available.

Third, we bracket the issue—which a comprehensive investigation of animal research ethics would address—of what level of risk and harm is acceptable in the context of nonbeneficial research with animals. This question is especially pressing given that animals are sometimes exposed to research that poses extremely high risks of morbidity and mortality. Rather than addressing this important issue, we focus on the status of assent and dissent with respect to nonbeneficial research with animals that satisfies the relevant limits on risks, whatever they turn out to be.

Fourth, the extent to which animals are able and willing to express their preferences with respect to participation in a given study depends, at least in some cases, on how they are treated prior to the study. Animals who are mistreated or severely confined may become “broken” or depressed, losing the ability to respond and express their preferences. To avoid the need to address this complication, we assume that, apart from research interactions, the animals in question are treated well and housed in ethologically appropriate ways.

Clarifying Assent and Dissent

We understand assent as a prospective indication by an individual of her willingness to participate in research. To solicit a human subject’s initial assent, the investigator explains the study to the extent that the individual can understand it and then asks whether the individual is willing to enroll. To solicit ongoing assent, the investigator explains the upcoming aspects of the study and asks whether the individual is willing to continue to participate. Assent can play a role in animal research only to the extent that investigators are able to communicate with the animal in question, the animal is able to attain a sufficient level of understanding, and the animal is able to communicate her prospective willingness (or lack thereof) to participate to the investigator. Of course, in many and perhaps most instances, these conditions cannot be met.

We understand dissent as an expressed objection, either verbal or behavioral, to what an individual is experiencing as a result of a research procedure. (We recognize that in this context we may employ “objection” and the verb “object” somewhat more broadly than they are usually used.) Objections might take the form of refusing to stay still, crying out, or otherwise expressing discomfort or discontent. Importantly, and in contrast with assent, research subjects can dissent without understanding the aspect of the research to which they are objecting, or without even understanding that they are involved in research.^{Footnote 14} Indeed, an individual might dissent to what she is experiencing precisely because she does not understand what is happening. Thus, among animals (and young human children, for that matter), dissent—or its absence, acquiescence—is more likely to be ascertainable than assent, which requires some understanding.

Investigators should not assume that any objection on the part of a research subject necessarily constitutes dissent to participation. Whether an objection qualifies as dissent depends on why the subject is objecting. For example, a subject might object because she is hungry and not because she is bothered by any feature of the research. Thus it is important for investigators to attempt to identify the source of a research subject’s objection. If the subject’s objection is not directed to the research, it does not constitute genuine dissent and may not provide a reason to halt or modify the procedure.

The following example helps to clarify the difference between assent and dissent. Imagine that an investigator is conducting a study that requires subjects to undergo an electrocardiogram (EKG). Soliciting assent involves either explaining or, if possible, nonverbally conveying to potential subjects that their hearts will be monitored for the purpose of research, as well as what they will experience during the procedure. For some subjects, it might be helpful to model the process of placing the EKG leads on a manikin. A subject provides assent if she demonstrates that she has adequate understanding of the relevant procedure and then agrees to participate.

If the subject is unable to provide assent, as is likely with most animals, the researchers might proceed with the study activity, taking care to be sensitive to the subject’s reaction. The investigators might place the subject on a table, encourage her to lie down, and begin applying the electrodes. The subject need not understand what is going on in order to express dissent to what she is experiencing as a result of the research. The subject dissents to the research if she refuses to stay still, starts pulling the electrodes off, cries out, or otherwise expresses that she is unwilling to undergo the procedure. If the subject cooperates or acquiesces, she shows a lack of dissent.

Two Reasons for Respecting Subjects’ Preferences Regarding Research Participation

In order to determine whether, and to what extent, investigators have moral reason to solicit animals’ assent or respect their dissent, it is helpful to begin by reflecting on the motivation for requiring that we solicit and respect the preferences of human subjects (even those incapable of providing informed consent) regarding their participation in nonbeneficial research. There are at least two reasons that underlie this motivation. First, individuals’ preferences often track their welfare interests. For example, an individual might decline to participate in a nonbeneficial study that involves multiple blood draws because she has a phobia of needles and would suffer as a result of the repeated blood draws. Similarly, a subject might withdraw from a study because it is taking too long and she is suffering from homesickness. The uncooperative behavior of a young child who is expressing dissent might indicate that she is uncomfortable or panicked because she is being forced to do something she does not want to do. We call reasons for respecting preferences that trace to the study’s impact on the subject’s quality of life welfare-based reasons.

A second sort of reason for respecting an individual’s preferences regarding participation in research is one that pertains to the individual’s agency. We understand agency as the capacity for self-directed, intentional action. Call reasons of this sort agency-based reasons. Agency-based reasons for respecting preferences arise when (and only when) there is value in an agent’s being free to do what she prefers and that value arises independently of whether her preferences track her welfare. For example, regulations for human subjects research mandate that investigators respect a competent adult’s decision not to participate in research even when it is in her interest to do so.^{Footnote 15,Footnote 16} Importantly, we are not assuming that merely being an agent is sufficient for it to be the case that an individual’s preferences merit agency-based respect. Rather, we understand agency-based reasons in such a way that it is an open question whether agency is sufficient, although it is obviously necessary. On the view we defend, it is not, in fact, sufficient.

It is important to clarify our reason for focusing on agency rather than on autonomy. When considering what sorts of agents are such that their preferences merit agency-based respect, the key question is as follows: Exactly why do the preferences of some agents merit agency-based respect? Given that many assume that agency is a necessary condition of autonomy and that the preferences of autonomous individuals merit agency-based respect, one might think that the answer is, These agents are autonomous. However, this does not necessarily address the question; even if being autonomous is, in fact, sufficient for an individual’s preferences to merit agency-based respect, this does not explain exactly what it is about being autonomous that makes an autonomous agent’s preferences meriting of agency-based respect. Given the considerable disagreement regarding what account of autonomy is correct, proponents of competing accounts might not even agree on exactly what it is. For these reasons, it is possible to pursue the key question without engaging with influential discussions on the nature of autonomy and freedom of action.^{Footnote 17,Footnote 18,Footnote 19}

Welfare-Based Reasons for Soliciting Assent or Respecting Dissent in Animal Research

All sentient animals have an experiential welfare and therefore welfare-based interests. Given our assumption that sentient animals have moral status, welfare-based reasons apply to animal research. The question, then, is whether, and to what extent, soliciting animals’ assent and respecting their dissent provide a useful means of protecting their welfare.

In contrast with human subjects research, there are currently no stated limits on the amount of risk and harm that animals can be subjected to in nonbeneficial research. The Public Health Service Policy requires that investigators make an effort to minimize animals’ suffering and distress (relative to research goals) but does not set upper limits on risk or harm.^{Footnote 20} As noted in the second section, we do not make any recommendations as to what the proper limits on risks and harms should be in animal research. Rather, we focus on the status of assent and dissent with respect to nonbeneficial animal research on the assumption that it satisfies those limits, whatever they may be.

If we assume that there are upper limits to the amount of risk and harm animals can be subjected to during research, these limits will play the primary role in protecting research animals’ welfare. However, it seems that welfare-based reasons for assessing animals’ dissent still have a role to play. First, forcing an animal to participate in a study against her will has the potential to cause severe anxiety, stress, and possibly lasting trauma. Even if the risk level of a study falls within the allowable range, the act of forcing an unwilling animal to participate in that study might cause the animal serious additional harm. Investigators should pay careful attention to an animal’s dissent and make an assessment of whether the harm caused by forcing her against her will exceeds the allowable limit. One consequence of implementing a requirement that investigators make such an assessment would be that, for any study in which the allowable limits on risk and harm are very low, it might turn out that investigators will be required to respect the dissent of any animal for whom being forced to participate is causing non-negligible harm.

Another case in which there are welfare-based reasons for requiring investigators to respect an animal’s dissent is when the dissent provides an indication that the study activities themselves are exceeding the relevant limits on risk and/or harm. For example, it might be the case that a study testing a new drug on mice is approved on the basis that injecting the drug will not cause more risk or harm to the mice than is allowable. However, this might be false; perhaps the data are incorrect or incomplete, or those who designed the study made an error when estimating the drug’s effect on mice. In such a case, the mice’s dissenting behavior might provide evidence that the relevant limits on risk and harm are being exceeded. One might suggest that this is a case of not so much respecting an animal’s dissent as just discovering, on the basis of the animal’s behavior, that the risk level of the study is higher than allowable. Even if this is true, in this sort of case the relevant behavior will constitute dissent to the study. Accordingly, investigators have an obligation to be sensitive and responsive to animals’ dissent and to take care to assess its source. The study should be discontinued when the nature of the animal’s dissent provides good evidence that the relevant limits on harm or risk have been exceeded.

Finally, there is a reason for investigators to take animals’ dissent into account even when it does not provide a reason for them to respect it (where “respecting it” requires removing the animal from research). In accordance with the Public Health Service Policy, investigators are required to minimize research animals’ suffering and distress.^{Footnote 21} Because an animal’s dissent often indicates that the animal is suffering, attention to dissent provides a means of promoting the goal of minimizing suffering. When an animal objects, the investigator needs to first determine the nature and source of the objection. First, is the animal dissenting to the research, or is her objection to something unrelated? Second, is the objection an indication of significant discomfort or pain, or is it indicative of relatively mild discomfort? In cases in which there is reason to think the animal’s discomfort is significant and related to the research, investigators have an obligation to do whatever is possible, consistent with the conduct of the study, to minimize harm. This might involve altering the study procedures or giving the animal an anesthetic. Although this does not constitute respecting the animal’s dissent, it highlights a welfare-based reason for investigators to be required to take animals’ dissent seriously.

Having found various reasons to attend carefully to research animals’ dissent, we must acknowledge challenges to implementing a successful practice of assessing and respecting dissent. First, as previously noted, it will not always be clear whether an animal’s objection constitutes dissent to the research per se. Second, even when it is clear that an animal is dissenting to the research, it might be difficult to discern exactly what about the research is causing the problem as well as what level of distress is being expressed.

These factors bear on how the investigators should respond when an animal objects during a research procedure. Whether the investigators are required to stop the research depends on why the animal is dissenting as well as the extent of the animal’s discomfort. Due to our limited means of communicating with animals, these factors can be difficult to assess. If the animal’s objection does not constitute dissent to an aspect of the research procedure, it might not provide a reason to remove the animal from the research. For example, if the animal is always hungry and her upset reflects only that she is currently hungry, then there is no special reason to stop the study. Even when the animal’s objection clearly constitutes dissent to the research, it might be in virtue of only one aspect of the procedure and not to participation more generally. For example, perhaps the light is too bright, or the animal dislikes the particular researcher who is holding her. If investigators are able to figure out what is bothering the animal and can modify the procedures to address her objection, then they should be required to do so (and need not discontinue the research). However, it will often be difficult for investigators to accurately assess the nature of the animal’s dissent.

A final difficulty is that there are studies for which it is not possible to respect or usefully assess dissent because the animal will not have the opportunity to object until it is too late to stop the procedure. For example, to assess the toxicity of experimental drugs prior to human testing, investigators might give animal subjects a single oral dose of the drug, with the goal of evaluating whether the drug makes the animals ill. In these studies, the research intervention—provision of the drug—occurs before any possible negative symptoms. By the time the animals have the negative experience that might evoke dissent, it is too late to stop the intervention.

Although these problems are significant, we do not think they undermine the appropriateness of requiring investigators to assess animals’ dissent. To the contrary, we think such assessments are imperative in many research contexts. In parallel contexts of human subjects research, investigators face similar challenges in evaluating the dissent of one-year-olds and adults with severe cognitive impairments. Are the cries of a one-year-old an indication of some passing unpleasantness or intense suffering? When an adult with severe Alzheimer’s states that she wants to leave, is she objecting to participating in the study or is she uncomfortable and distressed no matter where she is? Despite these challenges, investigators nonetheless do their best to assess the dissent of human subjects. Experience can help to determine which inputs cause distress and to what extent they can be ameliorated. We recommend that there be a requirement that investigators working with animals be educated on the specific sorts of animals they work with so as to facilitate assessing their behavior during research studies. Whenever it is unclear if an animal’s distress is severe enough to require halting the study, investigators should err on the side of caution and discontinue the research for the animal’s protection.

In contrast with dissent, we do not think that soliciting assent will provide an effective means of protecting the welfare of animal research subjects. Recall that, to provide assent, a research subject must have some understanding of the research as well as what her participation in it will involve. Most of the animals used in research are not capable of this level of understanding. Further, the lack of a common language makes it very difficult to communicate the relevant information to the animals. Even when investigators are able to “explain” the basics of the study to an animal, it is unlikely that the animal will be able to predict what the experience of participating in the research will be like for her. And, even if some animals are able to do this in some cases, the investigator will not be in a good position to know if the animal’s subsequent behavior is indicative of genuine assent. Finally, it is unclear that obtaining an animal’s prospective assent would be especially useful as a means of protecting the animal’s welfare. These considerations suggest that welfare-based reasons do not provide a reason to attempt to obtain animals’ assent.

The Nature of Agency-Based Reasons

As discussed previously, considerations pertaining to agency give rise to reasons for respecting an individual’s preferences where those reasons arise independently of how the individual’s preferences impact her welfare. In this section, we consider why some individuals’ preferences merit agency-based respect. We begin by considering why there are agency-based reasons for respecting the choices of competent adult human beings.

Consider the case of Mary, a healthy, 25-year-old single mother who suffers from neither physical nor mental disabilities. Mary intends to pursue a master’s degree and is deciding between getting the degree in elementary education or in engineering. It is generally believed that Mary has the right to make this choice—the choice of whether to go back to school at all, as well as what to pursue when she does. Other things being equal, it would be morally wrong for someone to take control of her life and make this decision for her.

Because Mary’s preferences will likely track her welfare interests, protection of Mary’s welfare provides some reason to allow her to make her own choices. However, this does not seem to be the only reason why it would be wrong to interfere with Mary’s choice. Even if Mary is in fact a poor judge of what will make her happy, she has an interest in making these sorts of decisions for herself.^{Footnote 22} Why is this?

We suggest that at least one answer is as follows: For an individual with sufficient decisionmaking capacity, there is value in her ability to play an active role in shaping her life. This ability is valuable because it enables such an individual to make her life her own. An individual might value the lives of others in virtue of what those people have achieved or the virtues they possess. But she cannot feel pride or regret for those lives in the way that she can with respect to her own. Consider the difference between a nice piece of art that someone else created and a piece of art that you created yourself. The latter has the potential to express your beliefs and emotions. It embodies a number of historical facts about how you came to have those beliefs, desires, and emotions and also reflects your idiosyncrasies and flaws. Even if the piece you make is, in certain ways, not as good as the artwork made by someone else, the fact that it is your own is of significant value.

When an individual is able to shape her life through her choices, this not only has consequences for how her life goes but also bears on the development of the individual herself. When we make choices that incorporate our beliefs, desires, and values, and live through as well as learn from the consequences of those choices, we develop as individuals.^{Footnote 23,Footnote 24} As such, the freedom to make choices is crucial to enabling us to shape who we are as individuals.

With these ideas in mind, we suggest that the following condition is sufficient for there being an agency-based reason to respect an individual’s preferences:

• The agency-valuing condition: The individual is an agent who values her ability to play an active role in shaping her life and herself.

We understand an agent as an individual who is capable of acting intentionally and making decisions in pursuit of satisfying her preferences. In order for an individual to satisfy the agency-valuing condition, it must matter to her that she is able to make free choices, and this must matter to her for the sake of her having the ability to contribute to shaping her life and herself.^{Footnote 25}

Two clarifications regarding the agency-valuing condition are in order. First, we are not claiming that satisfaction of the condition requires that the individual explicitly believe that she values her ability to play an active role in shaping her life and herself; rather, it just must be the case that the individual does value this ability. Second, satisfaction of the condition does not require that the individual have well-formed ideas regarding herself or how she wants her life to go. Although competent adults often do have such ideas and often are explicitly aware of these ideas and take time to hone them, this is not necessary in order to satisfy the agency-valuing condition.

For present purposes, the important question is whether any nonhuman animals satisfy the agency-valuing condition. A comprehensive answer to this question would require providing an account of exactly what cognitive capacities are necessary and sufficient for satisfying the condition, and then considering whether any nonhuman animals possess those capacities. Providing such an account would take us beyond the scope of our aims in this article. Rather, we suggest that the rich literature on the behavior and cognitive capacities of chimpanzees provides very good reason to think that normal adult members of this species satisfy the condition. If this is correct, then it is likely that at least those animals whose cognitive abilities are comparable to chimpanzees—in particular, other great apes, dolphins, and possibly elephants—do as well.

Chimpanzees as Deserving of Agency-Based Respect

We begin by enumerating a number of relevant behavioral and cognitive traits of chimpanzees and then explain why possession of these traits provides substantial reason to think that chimpanzees satisfy the agency-valuing condition.^{Footnote 26}

It is well established that chimpanzees have a relatively high degree of agency. They are capable of acting intentionally and do so in accordance with relatively complex plans.^{Footnote 27} One example of this is that chimpanzees plan their activities and travel routes on the basis of where they know food sources are.^{Footnote 28} They show foresight and the ability for long-term planning by fashioning tools that are appropriate for their present and future aims, as well as by selecting the appropriate tools to take with them on long journeys.^{Footnote 29,Footnote 30,Footnote 31} These behaviors also show that chimpanzees deliberate when making choices—they often choose in a way that is controlled and responsive to the specifics of their situation.^{Footnote 32}

In addition to being capable of intentional action, chimpanzees have demonstrated the ability to distinguish between the intentional and unintentional actions of others.^{Footnote 33} For example, chimpanzees became frustrated when a human researcher teased them by offering food and then pulling it away but did not get frustrated, and instead waited patiently, when the human accidentally dropped the food before the chimpanzee could take it.^{Footnote 34} Chimpanzees’ ability to distinguish intentional versus unintentional action suggests that, in addition to being relatively sophisticated agents themselves, they have at least a partial understanding of what is involved in being an agent or actor.

Chimpanzees are self-aware with respect to their existence as persisting agents as well as with respect to their identity within the social structure of their community. The former is demonstrated by the fact that they pass the “mirror test”—they are able to recognize themselves when reflected in a mirror, rather than mistaking the reflection for the presence of another animal.^{Footnote 35,Footnote 36,Footnote 37} Their social behavior demonstrates that they have a keen awareness of where they stand in the social hierarchy and what ways of behaving are appropriate given their standing. Importantly, there is also evidence that they care where they stand in this hierarchy and make an effort to move up the social ladder and improve their standing with those who are dominant.^{Footnote 38,Footnote 39}

Taken together, this evidence provides good reason to think that chimpanzees meet the agency-valuing condition. Chimpanzees’ relatively sophisticated level of self-awareness, in conjunction with the fact that they have some understanding of intentional action and agency, suggests some awareness of themselves as agents. They make deliberate choices aimed at fulfilling long-term goals, and their complex social interactions indicate that they have social ambitions and personal values. All of this suggests that chimpanzees not only are agents but also value their agency in virtue of the fact that it enables them to shape their lives and selves through their choices.

As noted previously, satisfaction of the agency-valuing condition does not require that an individual be capable of thinking or caring about herself and her life with the degree of self-reflective awareness that competent adult humans possess. The fact that chimpanzees do not have well-worked-out life plans and likely do not reflect on the sorts of individuals they want to be does not entail that they fail to satisfy the agency-valuing condition.

If chimpanzees satisfy the agency-valuing condition, then at least several other animals—namely, those with comparable cognitive abilities—probably do as well. However, we are not claiming that animals with significantly lesser cognitive abilities do not satisfy it, nor have we claimed that satisfaction of the agency-valuing condition is necessary, as well as sufficient, in order for there to be agency-based reasons for respecting an individual’s preferences and choices. Rather, our aim is to argue that there are agency-based reasons for respecting the preferences of at least some nonhuman animals.

Agency-Based Reasons for Soliciting Animals’ Assent and/or Respecting Their Dissent

Given the assumption that there are agency-based reasons for respecting the preferences of at least some nonhuman animals, including chimpanzees, how does this bear on our obligations to solicit assent and/or respect the dissent of these animals in research?

The answer to this question depends, at least in part, on whether agency-based reasons come in varying strengths. If they do not and, rather, all agency-based reasons are of equal strength, it follows that agency-based reasons for respecting these animals’ preferences regarding whether they participate in research are as strong as agency-based reasons for respecting competent adults’ preferences of the same sort. Because the latter reasons are very strong, it follows that researchers should be required to solicit the assent of these animals whenever possible and to respect their dissent whenever soliciting assent is not possible.

The other possibility is that the strength of agency-based reasons varies as a function of the strength of the relevant cognitive capacities of the individual. For example, competent adult humans are more self-aware than chimpanzees. Competent adults are also capable of explicit thought and reflection regarding themselves and their lives and have explicit ideas about who they are, who they want to be, and how they want their lives to go. One might think that these features suggest that competent adults merit stronger agency-based reasons than chimpanzees and, perhaps, stronger reasons than some children and cognitively disabled adults. If there are other nonhuman animals to whom agency-based reasons apply and they are weaker than chimpanzees with respect to the relevant cognitive capacities, then one might think that these animals’ preferences merit weaker agency-based reasons for respect than do the preferences of chimpanzees.

The most promising defense of the idea that agency-based reasons come in varying strengths appeals to the idea that the strength of an agency-based reason for respecting a being’s preferences depends on the strength of the being’s interest in shaping herself and her life. For example, one might argue that those who are more reflective and self-aware, or those who have richer conceptions of what they want their lives to look like, have a comparatively high stake in being free to make choices that allow them to shape their lives and themselves. This would allow that there are stronger agency-based reasons for respecting the preferences of, for example, a competent adult human than those of a chimpanzee. If it turns out that there are agency-based reasons for respecting the preferences of certain breeds of dogs, then there might be stronger agency-based reasons for respecting the preferences of chimpanzees than the preferences of those dogs.

If agency-based reasons do come in varying strengths, then it is possible that investigators do not always have an obligation to solicit the assent/respect the dissent of an animal to whom agency-based reasons apply. When, and under what conditions, investigators would be obligated to respect such an animal’s preferences would depend on the strength of the relevant reason. However, at a minimum, it seems plausible that investigators have an obligation to make sure that no animal to whom agency-based reasons apply has her dissent disrespected on a consistent basis. The idea is that if an animal is able to choose freely in most contexts in her life—and in the ones that are most significant—then disrespecting her preferences in a limited number of cases might not compromise her ability to shape herself and her life. For example, if an animal is occasionally forced against her will to participate in research but, for the most part, is left to live her life in the animal sanctuary, the harm might be negligible.

When it comes to animals who are sophisticated enough that agency-based reasons for respect apply to them, obtaining assent, and not merely respecting dissent, might be possible in some cases. In cases in which it is possible, doing so is preferable to merely respecting dissent because it shows respect for the animal if we provide her with a prospective choice, rather than proceeding with the study and waiting for her to object. If there is a reasonable possibility of obtaining meaningful assent, or at least of providing the animal with a general idea of what the research procedure involves, investigators should attempt to do so. Such attempts can be valuable even when they fail; they are valuable because it is more respectful to the animal if she is given some idea of what will happen in advance of its happening. For example, even if an investigator cannot provide an adequate explanation of the study procedures to an animal, she can make gestures indicating that she would like the animal to enter the lab rather than stay in the sanctuary. This provides the animal with a choice about whether to see what is happening in the lab, even if she cannot actually provide meaningful prospective assent to the study procedure.

Although there is value in obtaining the assent of some animals to whom agency-based reasons for respect apply, the difficulties involving understanding and communication that were outlined earlier are still present. The lack of a common language makes it difficult to obtain sufficient evidence that the animal understands, and it will be difficult to know whether the animal’s reaction constitutes genuine assent. We encourage investigators to attempt to identify ways to better communicate their aims to animals who satisfy the agency-valuing condition.

Summary and Conclusion

We have considered the extent to which there is moral reason for investigators to solicit the assent and/or respect the dissent of nonhuman animals who are used in research intended to benefit human beings. In doing this, we have provided an explicit ethical justification for the conclusions that were reached in the 2011 IOM report on the scientific necessity of using chimpanzees in biomedical and behavioral research.^{Footnote 40,Footnote 41} We have also discussed some of the practical challenges to obtaining animals’ assent and assessing their dissent. In spite of these challenges, we have argued that there are welfare-based reasons to require that investigators respect animals’ dissent in certain circumstances as well as consider the dissent of all animals. (As stated earlier, considering the dissent of an animal involves assessing the nature of the dissent in order to determine whether the research should be discontinued or whether anything can be done to make the experience better for the animal. We claim that all dissent should be considered, even if it is not ultimately respected.) Given the difficulties of communicating with animals and the relative inability of animals to understand what participating in a given study will be like for them, we claim that an assent requirement does not provide an effective means to minimize the harms caused to research animals.

We also have argued that there are agency-based reasons to respect the preferences of at least some animals regarding whether they participate in research. In particular, there are agency-based reasons to respect the preferences of at least those animals who satisfy the agency-valuing condition. We have argued that there is good reason to think that this includes chimpanzees, and any other animals with comparable cognitive capacities. We have left it open whether any other animals qualify. When agency-based reasons apply to an animal, they provide a pro tanto reason to respect the animal’s dissent, regardless of whether the dissent tracks the animal’s welfare. The question regarding what requirements should be implemented for investigators might depend on whether agency-based reasons come in varying strengths. If they do vary in strength, then whether the animal’s preferences must be respected might depend on the strength of the relevant reason. If agency-based reasons do not come in strengths, then investigators should be required to solicit the assent (when possible) and respect the dissent of all animals to whom agency-based reasons apply.