John Stuart Mill was a passionate believer in public debate, and felt with Spinoza that
the "collision" of ideas sharpens the minds of all parties, yielding suggestions no one would have lit upon in isolation and producing decisions more adequate than any proposal presented at the outset. Public opinion is a progressive force only when it is formed in a free-for-all public debate. Without institutional inducements for public criticism and opposition, in fact, political unanimity is likely to be a sign of irrational conformism.
Thus writes Stephen Holmes in his masterly Passions and Constraint. On the Theory of Liberal Democracy.
Beneath these political claims lies an epistemological principle that later came to be known as fallibilism.
[See also Socrates - Understanding Understanding]
Often there is no simple correspondence between reality and the perceptions of our minds.
So how can we know if our beliefs are true? Mill's trenchant answer was that "the beliefs which we have most warrant for, have no safeguard to rest on, but a standing invitation to the whole world to prove them unfounded." [...]
The Political Representative - Delegate or Trustee?
This fallibilist epistemology inspired many of Mill's political proposals. Consider, for example, his support of a trustee as opposed to a delegate theory of representation. A delegate is a mere agent, sent to parliament to express the opinions of his constituents and subject to immediate recall if he deviates an iota from his mandate. A trustee, by contrast, has ampler room for maneuver. He can vote as he thinks best, using his discretion, disregarding occasionally, if only temporarily, the opinions of his electors.
The delegate model is objectionable, according to Mill, because it implicitly rejects the epistemology of fallibilism. It implies that a representative has nothing important to learn from an uninhibited give-and-take with fellow deputies. But this assumption is unrealistic: "If he devotes himself to his duty," a representative "has greater opportunities of correcting an original false judgement, than fall to the lot of most of his constituents."
The decisive superiority of deputies over citizens lies not in higher intelligence, virtue, or education, therefore but in the unusual nature of the legislative situation itself, a situation which, according to Mill, fosters self-correction. Voters are parochial. They are seldom exposed to the clashing viewpoints of fellow citizens which live in remote parts of the country. No one is ever invited to prove them wrong or rewarded for disclosing their follies. Voters should defer to representatives, therefore, although only in the short-run, not because members of an elected assembly are likely to be especially virtuous, but rather because representatives enjoy the eye-opening benefits of exposure to stinging criticism and relentless debate.
A modern legislative assembly is a machine for public learning because it guarantees that rival political proposals will be "tested by adverse controversy." Deputies are encouraged not only to uncover each other's errors but also to change their own minds whenever they become convinced that they have been laboring under an illusion. If recanting is intelligent, then it can be justified publicly, even to the voters back home, at least eventually. Accountability requires that deputies explain their decisions to their constituents. Because explanations of difficult issues take time, however, a system of immediate recall would make a mockery of government by discussion. Far from being antidemocratic, the trustee theory of representation simply recognizes that public learning, or the collective correction of collective mistakes, can never be instantaneous.
(p. 180 - 182)