Are Voters Rational? (2)

Following on from part one (here), this section explores possible meanings of rationality:

II. Understandings of Rationality

Judgements of rationality will depend largely on what is meant by ‘rationality’, and it can have many different meanings. For some, rationality is merely a matter of internal consistency, requiring for example that one doesn’t believe both Q and not-Q. Sen, however, points out consistency alone doesn’t seem to be enough – if one consistently acts so as to frustrate one’s own desires, then one hardly appears rational.

A radically different account is offered by Plato, who believed the world was governed by a natural principle of order, the Form of the Good. He supposed that the Good is pursued by reason, and that it was in everyone’s interests to live in accordance with what was objectively Good – even if they had to be controlled by more rational Guardians in order to do so. For Plato, there are things that are just rational or irrational, regardless of one’s desires – for example, he thought acting justly was rational, while (to use a modern example) the non-instrumental desire for a saucer of mud may be irrational.

In between mere consistency, and living according to some objective heavenly plan, accounts of rationality employed in the social sciences often suppose agents act in order to maximise some quantity, e.g. firms act to maximise their profits, and individuals may seek to maximise their well-being or preference-satisfaction (I avoid terming either of these ‘utility’ as it is ambiguous between either of these, and other, meanings).

If we require that one act in order to realise one’s desires, accepting Hume’s dictum that “Reason is, and ought only to be the slave of the passions”, then we avoid Sen’s worry about self-frustration. This, however, seems to licence almost anything – after all, Hume insisted “’Tis not contrary to reason to prefer the destruction of the whole world to the scratching of my finger”. If what one desires is beyond rational criticism, then it seems RCT will be unfalsifiable – we need simply point out that those who vote desire to vote, and act so as to maximise the votes they cast. Moreover, one’s desires could be based on false beliefs. As Green and Shapiro observe “The constraints of logic in no way prevent one from imputing to voters the belief that their failure to cast ballots will dislodge the earth from its orbit”. Assuming one doesn’t want this, it will be rational to go to great lengths to avoid it – yet on this means-ends approach, it would be entirely consistent with someone’s rationality for them to want this.

Capturing what is meant by rationality seems difficult. We clearly mean something more than – or perhaps other than – consistency (it seems rationality must allow us to change our mind/preferences to reflect learning or circumstances). It is easy, and tempting, to think of it as simply a means-ends issue, but there are some ends we intuitively regard as irrational. To simply insist people maximise some unknown and unobservable value is not a useful explanatory or predictive theory, and true only by tautology (if we define maximising ‘utility’ to be simply doing what you prefer). We want rationality to be able to evaluate – and criticise – ends, yet without going down’s Plato’s route of supposing there is one uniquely best rational life plan for everyone.

Swift suggests numerous ways in which we might avoid totalitarian perfectionism – for instance, we may deny that there is one best life plan for everyone, or even any individual, or that anyone else can better know it, or be justified in imposing it. It seems we should recognise a plurality of legitimate goals, and that rationality will ordinarily allow us considerable discretion for many different actions – it is merely a minimal threshold that rules out certain self-conflicting or -destructive actions. I will return to these themes in the fourth section, but first I turn to the standard components of RCT.