S UBJECTIVE E XPERIENCE INTO W ELFARE 7.1 Introduction – the problem
7.4. A pragmatic proposal
I have flagged throughout this chapter that we might remain unconvinced about the metaphysical claim, regarding the existence of a common currency by which we can combine or weight different affects. However, even in this case, we would still want to make management decisions regarding welfare and resource allocation. Here, as mentioned, we might think instead of welfare as being a multi-component construct more like health. Here the different components may individually exist, but do not form an integrated state. Instead, they are brought together because of our categorisations, and interests. In this case, it would not be surprising that there is no metaphysical entity forming the common currency, or a single correct answer about the interactions between affects, and their weightings. Instead, we may have multiple accounts, each of which lays out a different way of combining affects into a single score, based on different background assumptions, normative commitments, and individual preferences. In this case, the problem of determining weightings is no longer an empirical one – there is no single privileged set of weightings that we are trying to discover. Instead, we are looking for a solution which best fits our purposes.
When considering why we want a method for integrating welfare components, we have the two reasons discussed in Section 7.1 – to make assessments of overall animal welfare, and to make considerations of trade-offs between different components of animal welfare. Most often, what we want is a way of determining how we should use our resources to maximise increases in welfare. To this end, we may consider using something like robustness reasoning (see Chapter Five). In these cases, we could compare the decisions recommended by multiple
different aggregation procedures and prefer those that are recommended by many or most different processes. Alternately, if simply considering whether a single proposed intervention is an appropriate use of allocated resources, we could examine whether it results in an increase in welfare across most or all of our different aggregation procedures. In cases where it does, we would proceed, and where it does not, we may instead look for an alternative. Something like the SOWEL model previously discussed could be extremely useful in these applications, as it allows for variation of input weightings, and could be used to test results across a range of acceptable parameters. In this way, even without the strong metaphysical commitment to a common currency and an integrated state of welfare as a real entity, we would still be able to make relevant decisions regarding animal husbandry and management.
This of course still leaves us with the issue of deciding on which set of aggregation procedures we consider reasonable. There are an infinite number of ways of weighting the different components of welfare and their contribution to the overall score, and so robustness reasoning of the type described above will not be of much use if we were to take all of these under consideration. Instead, we need to decide on a reasonable subset of those procedures. I will not here advocate for any definitive way of deciding on what would count as reasonable for these purposes, but I will raise a few relevant considerations.
Firstly, we would want to discount any procedure that places too much, or too little, weight on each component. If we are considering any particular condition or affect as a component of welfare, it is because we think it makes a significant contribution to welfare, and so we don’t want excessively low weightings. Similarly, it does not seem right that any single affect will be primarily determinate for welfare, and so aggregation procedures with excessively high weightings should also be discounted.
This ties in to the second consideration, which is that our weighting procedures should be intuitively plausible. We have some sense from our own experience of which types of affects have greater and lesser impacts on our own feelings of wellbeing. Similarly, from our observations of other people and animals we have an idea of what influences them the most. An aggregation procedure that weighted transient boredom much higher than chronic pain, for example, would not be convincing. Intuitive plausibility must obviously be taken carefully. As discussed in Section 7.3.3, weightings set by expert opinion run the risk of being overly anthropomorphised, and missing those things which matter to the animals under consideration. However, when supplemented with appropriate knowledge about the animals, this can help us narrow down the set of appropriate procedures.
The primary drawback of treating welfare as a construct and using these procedures is that, compared to the case in which we believe we are measuring welfare using a real common currency, we would have weakened confidence in the applicability of our results. Our aggregation procedures would be based on human intuitions about plausibility, and the process of arriving at these intuitions is opaque (think of the Welfare QualityÒ framework). Where we are basing our sets of reasonable aggregation procedures on these considerations, we might fail to capture those things that matter more to the animals.
The comparison here is again with health. Health is a construct of this type, and we are frequently able to make sense of trade-offs between different components – say, taking a medication to improve symptoms of a disease, that at the same time will impact kidney function. However, there are two reasons why this case is not exactly analogous to animal welfare. The first is that we are making decisions for humans, and so our intuitions about how to weight acceptable trade-offs are much more likely to hit the mark than for animal species, particularly those quite distantly related and dissimilar to us. The second is that even in the human health cases, we might think that we are actually often appealing to some other common currency; weighting components of health in relation to our preferences for different types of disease or incapacity, or relating to our overall lifespan or quality of life. If this is true, then we still need common-currency thinking. This is not to say that the results of the procedures described above will not get us some way to achieving the ends we desire in welfare measurement and decision-making, but that they might be much more limited than we would ideally prefer.
7.5. Conclusion
Scientific measurement of subjective animal welfare is fraught with a number of problems, one of which is how we can make sense of integrating or comparing different positive and negative mental states. Despite the heterogeneity of these different affects, I have argued that we have good reason to think that these can be integrated into an overall welfare experience. Further, we can measure this overall experience using whole-animal measures, and use changes in these measures - along with preference testing - to see how animals make their own trade- offs to determine relative weightings of different experiences. This will thus allow us to make effective management decisions. Even if we remain unconvinced about the presence of a real common currency, we can still use robustness reasoning on different aggregation procedures to identify the best decisions under these circumstances.