Richard Smith: Stop jumping from “is” to “ought”

Richard SmithLast week for the first time I examined a PhD, and one of my co-examiners, a moral philosopher, told us of “Hume’s guillotine” and taught us a lesson that all doctors should know.

The defence of the PhD was in Copenhagen and in public, as is the custom in Denmark. Around 80 people were there, and it was an excellent occasion. The PhD, Conflicts of Interest in Biomedical Publishing, was by Andreas Lundh and very strong. It comprised seven papers, six of them published (one each in the BMJ, Lancet, and PLOS Medicine) and one accepted for publication on the morning of the defence.

The studies showed the existence of many kinds of conflict of interest in biomedical publishing, and Lundh ended his clear presentations with a series of recommendations, all including the word “should.” Many of them were about increasing the transparency of biomedical publishing with recommendations like “medical journals should publish their financial statements” and “more drug trials should be conducted with public money.”

All of the recommendations seemed sensible to me, but Peter Sandøe, professor of bioethics at the University of Copenhagen, told us about Hume’s guillotine, which can be crudely summarised as “never derive a ‘should’ from an ‘is’.” It is, of course, the creation of David Hume, the Scottish enlightenment philosopher. Lundh had given us lots of excellent observations on how things are, but then he jumped, as is very much the custom in medical articles, to the recommendations.

As Sandøe pointed out, there was nothing in the middle—no spelling out of assumptions, no discussion of the arguments for and against the recommendations, and no recognition than the recommendations might do more harm than good.  His view was not that Lundh should abstain from making “should” claims, but that he should spell out the underlying moral principles and assumptions.  It was Lundh jumping straight from “is” to “should” that he objected to. Sandøe also said that it was ironic that the origins of recommendations of increased transparency should be so lacking in transparency.

Lundh and the rest of us recognised that this was a good point, and when I teach on writing scientific papers I tell people to stay away from making recommendations on which they have no evidence. Even with a clinical trial that shows treatment A to be better than treatment B in condition X it doesn’t follow that everybody with X should be treated with A: it might, for example, be very expensive, unobtainable, or toxic. Nevertheless, it is normal, sometimes even a requirement, to say in medical papers what you think should follow from your recommendations.

When I got home I looked up Hume’s guillotine and discovered that he took an extreme position, arguing that you should never move from an “is” to an “ought” or “should. Hume published his thinking in 1739 in his A Treatise on Human Nature. He wrote:

“In every system of morality, which I have hitherto met with, I have always remarked, that the author proceeds for some time in the ordinary ways of reasoning, and establishes the being of a God, or makes observations concerning human affairs; when all of a sudden I am surprised to find, that instead of the usual copulations of propositions, is, and is not, I meet with no proposition that is not connected with an ought, or an ought not. This change is imperceptible; but is however, of the last consequence. For as this ought, or ought not, expresses some new relation or affirmation, ’tis necessary that it should be observed and explained; and at the same time that a reason should be given; for what seems altogether inconceivable, how this new relation can be a deduction from others, which are entirely different from it. But as authors do not commonly use this precaution, I shall presume to recommend it to the readers; and am persuaded, that this small attention would subvert all the vulgar systems of morality, and let us see, that the distinction of vice and virtue is not founded merely on the relations of objects, nor is perceived by reason.”

I found this clear example of the problem in the Philosophy Index.

“One may consider the following moral argument as an example of an is-ought problem:

1.    Sam is stealing money from work.
2.    Losing money by theft causes harm to Sam’s employers.
3.    (One ought to not cause harm to his employers.)
4.    Therefore, Sam ought to stop stealing money from work.

Premises 1 and 2 are “is” statements, describing facts of what is happening. Premise 3 and conclusion 4 are “ought” statements that describe how things should be happening. But what is the source of this knowledge? This argument appears to be valid if the premises are true, but unless we can logically support premise 3, it is not sound. What can possibly give us rational knowledge that things ought to be different than the way things are?

Hume argues that we cannot, and that ought statements, and other supposed moral knowledge, are not rational.”

Breaking the law myself, I recommend you all to apply Hume’s guillotine with all the articles you read in medical journals and to use it to subvert the “vulgar systems of morality” that fill our newspapers.

Competing interest: RS was paid a fee for his work and had his expenses paid to travel to Copenhagen by the University of Copenhagen. While there he had some of the best food he’s ever eaten, some of it paid for by Andreas Lundh (after the defence) and some by Peter Gøtzsche, Lundh’s superviser. He’s very grateful to both and recommends Copenhagen to everybody as a funky city.

Richard Smith was the editor of the BMJ until 2004 and is director of the United Health Group’s chronic disease initiative.