Rebecca Brown and Julian Savulescu
Cross-posted from the Journal of Medical Ethics blog, available here.
There is a rich literature on the philosophy of responsibility: how agents come to be responsible for certain actions or consequences; what conditions excuse people from responsibility; who counts as an ‘apt candidate’ for responsibility; how responsibility links to blameworthiness; what follows from deciding that someone is blameworthy. These questions can be asked of actions relating to health and the diseases people may suffer as a consequence. A familiar debate surrounds the provision of liver transplants (a scarce commodity) to people who suffer liver failure as a result of excessive alcohol consumption. For instance, if they are responsible for suffering liver failure, that could mean they are less deserving of a transplant than someone who suffers liver failure unrelated to alcohol consumption.
These are challenging practical questions, but philosophy – in combination with other disciplines – can help. This involves a combination of intuition-mining through the use of thought experiments, and logical reasoning about justifiable principles, concepts and theories, a process the political philosopher John Rawls called reflective equilibrium. Gradually we can make our judgments about what responsibility is, what conditions must be present, why it matters, and so on, more consistent. One upshot of this is the identification of two conditions thought necessary by many philosophers to judge an agent responsible for some action. These are the control condition and the epistemic condition. The control condition requires that, for an agent to be responsible, she must have been able to control her action (she wasn’t forced or experiencing an epileptic fit or something similar). The epistemic condition requires that the agent could foresee the likely consequences of her action, including their moral significance (she knew that pulling the trigger would send a bullet into the victim’s leg, causing him serious injury; she didn’t believe it was a fake gun or loaded with blanks).
If either of these conditions are not fulfilled it is common to judge people as not responsible. But how does this work in the health context? We can take the same approach – ask whether or not the agent had control over her health-affecting actions, and whether or not she understood the health-affecting consequences of those actions. However, lots of health harms result not from a single, discrete action with clear likely consequences (firing a gun at another person). Instead, they result from the accumulation of numerous actions, each of which makes only a small, probabilistic contribution to eventual health harm. Moreover, one agent’s actions are heavily influenced by the environment they inhabit, including the actions of other agents. For instance, people are more likely to smoke if those around them smoke. The same goes for diet – people will tend to eat similar things to those they live and socialise with.
Theories of responsibility have been developed more or less with discrete, morally significant behaviours, performed by identifiable individuals, in mind. But in the health context the behaviours are repeated, may appear morally benign, and performed by an agent influenced by those around them. If we are to assess responsibility for these health-related behaviours, we need to consider how our theories of responsibility should be adapted to apply properly in these contexts.
In a recent paper in the Journal of Medical Ethics, we argue that two areas need particular attention: how should we assess responsibility over time and across agents. The first must tackle the question of how often the control and epistemic conditions must be fulfilled when considering complex (repeated) behaviours. For example, if we are interested in whether or not a smoker is responsible for developing a smoking-related disease such as heart disease, we might first ask ‘was the smoker responsible for smoking?’ But this seems to require consideration of whether the smoker was responsible on each occasion she smoked a cigarette. That means considering whether the control and epistemic conditions were fulfilled every time she smoked. It is not clear what the threshold should be for considering someone responsible for a smoking habit – whether they must fulfill the conditions of responsibility on every occasion they smoke, or most or only some of those occasions.
The second consideration is to look at who the ‘agent’ is that is responsible for smoking. This might seem obvious – we typically identify agents with the human bodies they inhabit, as bounded by the skin. But there are plausible arguments for extending agency outside the skin – if we make use of various prosthetics or assistive technologies to enhance our physical functioning and cognitive powers, why not include these artefacts within the bounds of the agent? This raises the possibility of whether it is possible to identify the ‘agent’ who is responsible for some action as being spread across multiple bodies. If this is the case, the most likely instances where it would happen seem to be intimate dyads – couples who live together, have special obligations towards one another, have significant influence over one another’s behaviour, and who share important goals. In these cases, someone who seeks to quit smoking might be significantly aided in doing so by a supportive partner who supports them to do so. Similarly, the partner may scupper those efforts by continuing to buy them cigarettes or mocking their attempts to quit. In these cases, we think responsibility might plausibly be ‘dyadic’ – distributed across the dyad, rather than located only with the individual.
The answers to questions about how responsibility should be evaluated over time and across people could have practical implications, in terms of targeting healthcare interventions and distributing resources appropriately. There appears to be political will to ‘hold people responsible’ for their health. Taking account of diachronic responsibility could result in those more frequently fulfilling the conditions of responsibility for their behaviour being more blameworthy than those whose capacity and foresight have historically been more variable. Dyadic responsibility might justify requiring those who contribute significantly to a partner’s health outcomes to share in the responsibility for those outcomes. More work is needed to identify what the implications of responsibility are in the healthcare context – most importantly, whether it is ever legitimate to deny or delay treatment on the basis of responsibility. But if responsibility is to play any role in healthcare, it is vital that it reflects the reality of many health-affecting behaviours. Our current accounts of responsibility are poorly equipped to guide us here. It is essential we adapt them in the light of the reality that humans are creatures whose behaviour changes over time and is influenced by others.
This post relates to Brown, RCH and Savulescu, J (2019) ‘Responsibility in Healthcare Across Time and Agents‘ Journal of Medical Ethics, 10.1136/medethics-2019-105382