Apparently, there was a TV programme in Australia the other day in which a there was a discussion of assisted dying. It got reported in The Guardian, largely on the basis that an 81-year-old audience member kept calling Margaret Somerville “darling” and then got mildly sweary. I’ve only seen those clips from the programme that are linked in the Graun‘s report, so I’m not going to comment on the tone of the debate in particular. Rather, I’m interested in one of the responses to the programme, from Xavier Symons, writing in The Conversation.
Symons takes the opportunity to unpick the idea of a slippery slope argument – in this case, the claim that allowing some forms of assisted dying will commit us to allowing… well, that’s open-ended, but it’s sufficient to say that it’d be terrible. We’d want to avoid terrible things; therefore, the argument goes, we shouldn’t allow any of it. This is well-worn stuff in the seminar room, but it’s a mode of argument that refuses to die. Quite correctly, Symons points out that
there is a need for empirical evidence or sound inferential reasoning to support the claim that event B will necessarily (or probably) follow on from event A. Without this evidence, the argument is invalid. I can’t just claim, for example, that the legalisation of medicinal marijuana leads to the legalisation of ice – I need to show some empirical or logical connection between the two.
So far, so standard. (I’d say “unsound” rather than “invalid”, because the validity of an argument doesn’t depend on its evidence – or, at least, not in the same way; but that’s a small matter.) He then makes another move, which is a bit more interesting:
But (and it’s a big but) there is such a thing as a good and valid slippery slope argument. A good slippery slope argument demonstrates a causal or probable relationship between event A and B, such that event B can legitimately be expected to occur if event A is allowed to occur. […] There are, nevertheless, compelling empirical and logical slippery slope arguments available to defend more modest claims about the “normalisation” of assisted dying.
Is this correct?Well, talking about a “good slippery slope argument” does, I suspect, load the dice somewhat: a good slippery-slope argument may be a good example as measured by the standards of that kind of argument; but if that kind of argument is flawed fundamentally, we don’t move very far. And an appeal to logic doesn’t really get us much further: the logical structure of an argument that allowing assisted dying will have all-things-considered undesirable effects is the same as that of an argument that allowing it will make us helpless to resist the state-imposed murder of undesirables. The difference between those arguments is one of degree, rather than structure.
But I think that Symons is likely to be on firmer ground if we put that kind of consideration to one side, and concentrate on what can legitimately be expected to occur; and to do this, we should look at the empirical side of things. Technically, this means we’re moving from a slippery slope kind of argument to a more straightforward inductive argument. The claim being advanced is that rates of assisted dying have increased steadily in countries where the law has been changed to accommodate it. At first glance, the statistics he provides – dealing with the Netherlands and with Quebec – would seem to support that claim. In terms of brute numbers, the claim looks plausible.
The chart on the right is taken from Symons’ source, and it shows rates of euthanasia in Holland. The number of cases does, indeed, rise year-on-year. However, the chart begins in 2002; this is when the procedure was legalised. And that does make an important difference. It seems perfectly reasonable to expect that reported cases of euthanasia will increase year-on-year in the period immediately after legalisation. It may become more common, and even if actual occurrences don’t become more frequent, reporting will be more common. (In roughly the same way, the number of same-sex marriages in the UK shot up in 2014. That doesn’t tell us much once we remember that the relevant legislation only came fully into force in that year.)
Now, we might expect rates to plateau after a little while; there’s no evidence of that happening yet. But even that shouldn’t necessarily worry us too much. For one thing, demographics may play a part. As the population of a country ages, so the elderly population grows. This gives us a reason to think that the number of people suffering from a chronic illness is also likely to grow, since chronic illnesses like diabetes, COPD, and Parkinson’s and old age are related. That means that there’s likely to be a steadily growing population of people who might decide that assistance in dying is for them; and that growth may be fairly straightforwardly explained. (I don’t have any hard data on this; I’m just tossing it out as a possible explanation. Symons doesn’t offer any hypotheses either, so tu quoque.) Further, as euthanasia gets more established, one might expect discussion about it to become more frank, and for some people to request it who perhaps would not have – or, at least, would not have publicly – before. So when Symons claims that
[t]here is significant evidence from the Benelux countries (Belgium, Netherlands, Luxembourg), as well as the US and Canada, to support this claim. Around 3.7% of all deaths in the Netherlands in 2015 were by virtue of euthanasia or assisted suicide, up from 1.3% when the procedure was legalised in 2002.
And while Dutch legislation changed several times during that period, the steady rate of increase continued even in years when there was no legislative change.
– well, we could still have room to ask further questions.
What seems more important is the claim that
there is significant evidence to suggest that if we do legalise assisted dying in Australia jurisdictions, the practice will be normalised, and we will see a steady but significant increase in deaths by such means.
It’s normalisation that appears to carry the most moral weight in the argument here. That’s fair enough; but then we face a further question about whether that should bother us in the slightest.
Sometimes normalisation of an activity should worry us. If the law is changed to allow for, say, explicit discrimination on grounds of race when it comes to employment, then such discrimination may become normalised. That would be a proper source of moral concern. But what matters is not the process of normalisation, so much as what is being normalised. The root of the worry in Symons’ case is that the discrimination is wrong in its own right – take that away, and normalisation is neither here nor there. We aren’t concerned when laws help normalise things of which we approve.
In other words, even if the normalisation interpretation of the slippery slope argument holds, it doesn’t follow that we have to worry about it: as Symons observes, “some proponents of assisted dying might see ‘normalisation’ as a positive development”. But that just leaves me wondering: if Symons’ reconstruction of the slippery slope is mainly concerned with normalisation, but he’s willing to remain agnostic about whether normalisation is a good or bad thing, then it does seem to mean that things are neither very sloped nor very slippery. Symons does call the initial figures from Quebec “alarming” – so maybe he’s not so agnostic after all – but whether or not we should be alarmed has nothing to do with any putative slope, and he’s not given us any reason here to be alarmed.
One would expect a huge leap in numbers of an activity being reported between the time when it was illegal and the time when it was legal. On its own, that’s not enough.