The Best Way To Derive An 'Ought' From An 'Is'

One of the long-standing moral problems in philosophy, dealt with most famously by David Hume, is the 'is–ought problem' of how to derive an 'ought' from an 'is'. Hume's 'is–ought problem' is the contention that deriving an 'ought' from an 'is' is not possible on deductive grounds - that is, you cannot 'prove' something normative from something descriptive. Or to put it more simply, in moral or ethical terms you cannot prove what you 'ought' to do from what 'is' in the factual scheme of things. So for example, statements like 'John doesn't pick up his dog's mess when he fouls on the park' and 'America spends millions on space exploration' are both descriptive. Hume is saying that nothing in experience enables us to prove that John ought to pick up his dog's mess or that the money America spends on space exploration is either too much or too little.

However, proofs aren't everything, and Hume was quite willing to concede in his An Enquiry Concerning the Principles of Morals that moral propositions are built on a decked succession of experiences related to preferences, feelings and strategies regarding our well-being. This is because, being a strict empiricist, Hume's primary concern is how these actions play out evidentially, i.e., how empirically they affect us. He didn't insist that morality has to be 100% factual; he merely insisted that our moral thinking is derived from general maxims built on analyses of particulars in everyday experience.

The trouble Hume has is that any connection we make is only one link on the chain - a chain that involves other links of connection, which ultimately lead to a brick wall. On this basis Hume is being ultra sceptical about how easy it is to make moral statements. If you recall, Hume's fork says that the whole human interface with reality can be demarcated into either matters of facts (experience of the physical world) or relations of ideas (logical and definitional connections of those experiences). Using Hume's fork we are unable to prove logical and definitional connections of morality (ought) from experience of the physical world (is), not because we can't make valid empirical statements (such as "If John wants to be knowledgeable he ought to study hard) but because under Humean terms being knowledgeable is only a preference humans have.

Similarly, we could say that it is better for the body to eat healthily, but that then leads to the corollary question, why is better to be healthy? A response might be 'To live longer', but that comes with the further question "Why is it better to live longer?". None of these answers are able to be proved deductively, so every answer always begs the question, and brings us eventually to this epistemological brick wall.

Sometimes, though, where philosophy reaches brick walls, economics can dig holes through those walls, particularly given that economics concerns itself with human behaviour, incentives and actions as well as arithmetic, graphs and bell curves. You may recall that studies from Paul Eckman show that the capacity for emotions like fear, joy, distress, anger, surprise and revulsion is not learned, it is innately part of being human. Cultural nuances dictate how people feel about those emotions, but that just about everyone has them is beyond reasonable doubt. What can be derived from this is that there are things that are objectively better and worse for all humans, and David Hume summarised the body of the book with this statement in good company with Eckman's finding --

"The final sentence, it is probable, which pronounces characters and actions amiable or odious, praise-worthy or blameable... depends on some internal sense or feeling, which nature has made universal in the whole species".

Or to put it in economic terms, humans as a species are very alike, and they do demonstrate numerous ways that their preferences turn into behaviour and actions. When it comes to questions of morality and ethics and right and wrong behaviour, rather than looking for any formal proofs, the question regarding deriving an ‘ought’ from an ‘is’ in economic terms is more about whether it is possible to arrive at a general consensus on morality, and if so, how we reach such a consensus. I’ll add that we can only attempt to do this in the first place because our experience of the world shows us that logical systems work in approximating reality. Consider this statement: “Plants take up carbon dioxide in the atmosphere by the process of photosynthesis”. Within the realms of the body of science we have constructed, this statement is a factual statement, because to contest it would involve departing from the standard rubric of the natural order and proposing an alternative statement that could be subjected to scientific scrutiny. Now consider this statement; ‘John shouldn’t bunk off school in order to steal cigarettes from his local newsagent’.  That is not a factual statement – it is the expression of a desire or feeling which requires much more complex analysis, and which, when subjected to intense scrutiny, would take us further and further into the realms of subjectivity and arguments about consensus. 

But pretty much everyone would naturally agree that it is correct that John shouldn’t behave that way. John shouldn’t bunk off school because evidence shows that poor education inhibits people’s chances of progress and it contributes to their downfall. Similarly, John shouldn’t steal because evidence shows that theft is generally bad for the individual and the society as a whole. John would be advised not to take heroin because evidence shows it will be detrimental to his health. We could come to agree that such statements could reasonably qualify as facts, because we have reasonable grounds on which to state it as fact. That's perhaps the best we can do. But arguably more importantly, the evidence for these propositions is based on how humans behave in order to arrive at goals, the values they have, the empirical evidence for what is good for us physically and psychologically, and what's consistent with human behaviour regarding the purposes we construct.

Given that we have limited knowledge of the empirical world, the only way I can see to derive an 'ought' from an 'is' is imperfectly by using our old friend probability. Once one gets one's head around the notion that almost all knowledge is about probability it becomes clear. To illustrate, there is a well known paradox that can shed light on this situation - the ‘sorites’ paradox. The ‘sorites’ paradox is attributed to Eubulides, who had a handful of beans, and in front of his students placed one bean at a time on the table asking them each time whether that particular bean made it a heap of beans.  They continued to say no, and then when the 15th bean was laid down, they said 'yes', that's a heap'.  The paradox asserts that it’s absurd to just declare that 15 is a heap.  Why is 15 a heap and not 14? Why not 16? What about if Eubulides did the same experiment to another group of students and they thought a heap was 13?  What if one or two in the group thought a heap should be 18?  The take home lesson is that although people adhere to systems they believe are assented to with rigorous reasoning, most often what they are actually dealing with is the arbitrary classifications.

Thought about logically, the question 'When does a heap become a heap?' can only be answered in two ways. We can say a heap of sand must exceed n where n equals a designated number for qualification. Or the other way to solve the heap problem is to say that the probability of calling the pile a heap increases with every grain added. The latter is the correct epistemological route to take because the world is full of many comparable examples, where things of which we think we are certain are really feelings we have based on probability estimates. This is perhaps the best rule of thumb for knowledge and for moral axioms; almost all knowledge is probability based, and everything that constitutes knowledge is arrived at in exactly the same way as the Sorites situation – each increase in evidence or data increases the probability of something constituting knowledge or a moral axiom.