Recently, Bonnie Steinbock has argued that there is still not a convincing case to support the legalisation of doctor-assisted suicide.¹ The argument is framed in consequentialist terms: rather than contend that there is something intrinsically wrong with mercy killing itself, caution is recommended because of the risk that a system may be open to sufficient abuse to warrant its non-implementation. A welcome criticism is made of partisanship that obstructs useful progress in the debate, which she suggests should be based on objective, empirical evidence on both the need for, and the risks associated with, legalised doctor-assisted suicide.¹ The main presumption underlying Professor Steinbock’s argument is that a proportionately large number of people would become vulnerable to great and unacceptable harm under a regime that was not strictly regulated. She asserts certain values, which, on a consequentialist calculation, support her conclusion that currently the case for legalisation is not made.

Here I would like to question how useful objective, empirical data can be. Although data that show some assertions to be objectively right or wrong may be obtained and are doubtless valuable, I will show that the debate on assisted suicide is unavoidably partisan and that the protagonists will therefore receive data differently, necessarily because of their moral and philosophical convictions. Steinbock would like the question of legalisation to “turn on objective, empirical evidence about both the need for PAS [physician-assisted suicide] and the risk of mistake and abuse.”¹ But to be understood, the concepts of “need” and “abuse” require pre-existing value judgements. Furthermore, even if the values of these concepts are quantifiable, they are probably parts of distinct metrics, so offsetting one against the other will become a matter of qualitative assessment. This does not entirely disvalue the use of objective data, but it means that the …