Introduction
The ethical analysis of our moral obligations to animals, as well as practical decisions regarding many issues in veterinary medicine, both presuppose some account of "animal welfare" - that is, of what makes an animal's life go well or poorly. Therefore, it is essential that we think critically about the concept of animal welfare. This paper briefly explores some philosophical issues in the definition and assessment of animal welfare.
Two Different Senses of "Animal Welfare"
The first thing to clarify is the way in which I'm using the term "animal welfare." In both moral philosophy and ordinary language (as concerns humans), the term "welfare" is roughly synonymous with "wellbeing." To say that an individual "has a welfare" is to say that her life can go better or worse from her perspective, and to talk about her welfare is to talk about her interests - what are the things that make her life go better or worse for her? How should we conceptualize her wellbeing in the abstract? This is the sense in which I am using the term "animal welfare" throughout the article. There is a second sense of the term often used in veterinary medicine, where "animal welfare" means a normative ethical position regarding our obligations, often contrasted with "animal rights." I discuss this second sense of animal welfare in its own section below.
"Welfare" is an Evaluative Concept
While it is often claimed in veterinary medicine that animal welfare is a "science-based" concept, this is not strictly speaking true. Rather, "welfare" is an evaluative concept at its core: it refers to what makes for a good life, and terms such as "good" or "bad" are paradigmatic terms of evaluation. Another way of appreciating this is to observe that no amount of scientific data by itself will answer questions about animal welfare. Such data can always be met with the question "Why is that bad for the animal's welfare?" For example, data that supports the conclusion that an animal is in pain do not in themselves "prove" that the animal is experiencing poor welfare unless joined to the evaluative judgment that pain (distress, fear, etc.) is bad. So, too, for judgments that an animal cannot express natural behaviors, that it is not "coping" with its environment, and so on. It is important to recognize the evaluative nature of the concept of animal welfare, since the value judgments at play in the concept may be contested between parties. Furthermore, the kind of scientific information that is relevant when assessing an animal's welfare will depend on how we define welfare.
There are at least a few reasons why welfare is often thought to be "science-based." First, scientific studies are sometimes crucial in assessing an animal's welfare. Second, for much of the 20th century science was gripped by the philosophy of logical positivism, which separated science and ethics. Logical positivism has long since fallen in philosophical circles, but it has left a legacy in veterinary medicine. Third, if welfare is thought to be "science-based," then this bestows expertise upon veterinarians and animal scientists in political disputes about animal welfare - hence there is a political element to how welfare is conceptualized (Rollin 1990; Carbone 2004).
Which Definitions of Welfare Are Plausible?
Welfare has been defined in a number of ways, including "feelings-based" definitions, which focus on an animal's mental states and how it feels; "natural-living" definitions, which focus on an animal's ability to express natural (i.e., "species-typical") behaviors or live in natural conditions; and "function-based" definitions, which focus on such things as normal physical functioning, the absence of disease, survival and reproductive fitness, and adequate coping (adaptation) to various environmental stimuli - often as measured by certain hormonal axes (Fraser et al. 1997). Which of these definitions is most plausible? One important point, often obscured by science-based views of welfare, is that any plausible definition of welfare will likely have to refer to what an animal subjectively experiences. When we talk about human wellbeing, we are talking about what makes our lives go well for us - our quality of life - and it is hard to say that something advances or detracts from our welfare when it never touches what we experience. So, too, for animals.
However, the fulfillment of certain natural behaviors and physiologic measures may be indirectly important in that they relate to what an animal is experiencing. Furthermore, even if we say that a plausible definition of welfare must reference what an animal subjectively experiences, there are more and less plausible ways to develop this. For example, some subjective definitions of welfare focus only on the absence of pain, distress, fear, and other unpleasant feelings. But good welfare is arguably more than the absence of bad experiences; it is the presence of good experiences, too. It follows from this that animals can be harmed if they are deprived of positive experiences that they otherwise would have enjoyed, even if in being deprived they do not experience pain, fear or distress.
The Harm of Death
A related issue to how we define welfare is whether death harms animals. Many persons in the veterinary, agricultural, biomedical and animal welfare science communities, both historically and presently, have implicitly assumed that death does not harm animals. A few philosophers have also argued this, holding that a being must understand the concept of death and have a desire to stay alive in order to be harmed by death, since on their view we can only be harmed by having our desires thwarted. Hence, on this view it is at least conceptually possible that animal agriculture and most biomedical research, which result in animal death, can be conducted in a way that does not harm animals. However, this view is challenged by those who argue that death can still harm animals who do not have a concept of death, since it deprives them of the enjoyments that future life could bring. Indeed, if we must have a self-conscious desire to stay alive in order for death to count as a harm, then infants, the senile and the severely mentally disabled are not harmed by death - a conclusion that most of us find counterintuitive (DeGrazia 2002; Haynes 2011).
Value Judgments in Animal Welfare Science Experiments
The way in which we define "animal welfare" will condition which kinds of data are relevant to assessing an animal's welfare, but we should also be aware that the scientific experiments that furnish these data are themselves value-laden, as are the conclusions that we draw from such experiments. For example, research assessing the welfare impact of sow housing systems will involve methodological value judgments relating to the choice of alternatives that are compared. Similarly, preference-testing experiments also involve a similar judgment about choice options and how these are constructed. Such judgments can affect the conclusions that we draw from the studies. For example, an experiment offering sows 1 kg of straw that they could obtain with an operant response concluded that the sows were not motivated to obtain the straw, since they rarely performed the behavior required to obtain it. However, when another experiment offered 18 kg of straw, the sows were strongly motivated to access it (Fraser 1995). Other experimental variables that may influence results include the type of aversive stimulus, the timing and duration of an aversive stimulus, an animal's psychological state when being subjected to various conditions, the time course of the response being measured, interspecies and individual differences, and others (Mason, Mendl 1993).
Yet another type of methodological value judgment concerns the correlation of measurable attributes, such as physiological variables, with mental states. For example, an animal's cortisol levels may be taken as a measurable proxy for mental states such as anxiety. The basis for the correlation between measure and mental state is always inferential, since we cannot ever directly access an animal's mind to prove that the measure is exactly correlated with the mental state in question. It is important to be aware of such value judgments, since they may be contestable. For example, while elevations in cortisol levels may sometimes be plausibly associated with aversive mental states, such elevations may also be associated with novel and even pleasurable conditions. In addition, cortisol levels may sometimes be normal in conditions that are detrimental to animal welfare as evinced by other criteria (Fraser 1995).
Finally, when evaluating the cumulative welfare impact of many management practices, risks and benefits to the animal must be traded off against each other. These risk-benefit assessments are inherently value-laden. For example, some persons may emphasize the benefits of controlling feed and temperature, preventing attacks from mates, and lowering the risks of infectious disease as speaking in favor of indoor, intensive confinement of farm animals. Other persons may argue that infectious disease risks of outdoor confinement are small, that animal feeding, temperature and physical protection can be adequately accounted for by non-intensive conditions, and that living freely and in accordance with the animal's nature outweighs any putative benefits of intensive confinement (e.g., Fraser 1997; McGlone 2006).
Animal Welfare Versus Animal Rights
A final issue to discuss is the use of "animal welfare" as a normative ethical position, often contrasted with "animal rights." Typically, what seems to be entailed by the position of "animal welfare" (or, sometimes, "animal welfarism") is that while animals have some moral standing and should be treated respectfully or humanely, it is nevertheless acceptable for humans to use animals in various ways that harm them. Animal welfarism so defined tends to be contrasted with "animal rights" because it is thought that an animal rights position is incompatible with such uses of animals.
While it is certainly possible to elaborate a philosophically coherent position that accords some moral standing to animals, but lesser moral standing than humans, my own view is that we should move away from using "animal welfare" in the normative sense elaborated here. First, in practice it tends to conflate our obligations to animals with "welfare" in the sense of what makes for a good animal life, which can make discussions confusing. Second, it is not strictly true that animal rights positions proscribe all human uses of animals, or posit that humans and animals are moral equals. Some do, but some do not - it all depends on how the rights position is specified. Hence it is misleading to oppose animal welfare and animal rights. Third, the position of animal welfarism as it is usually sketched is both vague and under-supported. It ultimately gives us little guidance on how we should treat animals, since terms like "kind," "humane" and "respectful" are very vague. While most developed moral positions start with a set of principles and derive conclusions from them, the welfarist position starts with a set of conclusions - that various kinds of animal use are permissible - and does not provide any supporting argument. Furthermore, it is presumably not true that any and all decisions associated with various kinds of animal use are morally permissible. If we have obligations to protect animals' interests to some degree, then the human use of animals will only be permissible when it proceeds in a way that honors these obligations. But the welfarist position tends not to spell any of this out. What is needed is a moral position that is more concrete regarding our obligations, and also provides reasons to support its assertions. I propose that a framework of "moral considerability" (described elsewhere) is the best candidate for this.
References
1. Carbone L. What Animals Want: Expertise and Advocacy in Laboratory Animal Welfare Policy. New York: Oxford University Press; 2004.
2. DeGrazia D. Animal Rights: A Very Short Introduction. New York: Oxford University Press; 2002.
3. Fraser D. Science, values and animal welfare. Animal Welfare. 1995;4:103–117.
4. Fraser D, Weary DM, Pajor EA, Milligan BN. A scientific conception of animal welfare that reflects ethical concerns. Animal Welfare. 1997;6:187–205.
5. Haynes RP. Competing conceptions of animal welfare and their ethical implications for the treatment of non-human animals. Acta Biotheoretica. 2011;59:105–120.
6. Mason G, Mendl M. Why is there no simple way of measuring animal welfare? Animal Welfare. 1993;2:301–319.
7. McGlone J. Comparison of sow welfare in the Swedish deep-bedded system and the US crated-sow system. J Am Vet Med Assoc. 2006;229(9):1377–1380.
8. Rollin B. The Unheeded Cry: Animal Consciousness, Animal Pain, and Science. New York: Oxford University Press; 1990.