In the early 20th century, philosopher Karl Popper came up with a new theory of knowledge: a new epistemology. It emphasized the means of error correction over trying to devise theories that lead to guaranteed (or probable) success. Approaches and methodologies that improve the ability to correct errors should be preferred over those that impede this ability. Popper’s discovery led him to his famous criterion of democracy: that a democracy is not necessarily based on votes or majority rule, but is able to remove bad leaders and policies without bloodshed. This is less of a definition and more of a statement of the character of democracy as a political system; a statement of what problem democracy solves.

Physicist David Deutsch, building on Popper’s work, writes:

The ancient Roman ruler Julius Caesar was stabbed to death […]. The interesting question is […] how it came about that other politicians plotted to remove him violently from office and that they succeeded. A Popperian analysis would focus on the fact that Caesar had taken vigorous steps to ensure that he could not be removed without violence. And then on the fact that his removal did not rectify, but actually entrenched, this progress-suppressing innovation.

The word “entrenched” is key. While some errors are benign in the sense that they do not hinder their own correction, others have the ‘special’ property that they ‘fight’ being corrected. Deutsch gives the example of legalizing theft vs banning debate. Both are errors, but the former does not try to prevent its correction: if you ban theft, you can find out through debate that this was a mistake, and then you can debate ways of correcting the mistake. Banning debate itself, however, is harder to undo since the ban effectively includes a ban on its own fixing: without debate, it’s much more difficult to realize that banning debate was a bad idea. Deutsch describes the mistake of destroying the means of error correction as a “rare and deadly sort of error: it prevents itself from being undone” (ibid., chapter 10). Self-entrenching errors are like barbed hooks.

This insight is not limited to politics. Popper originally emphasized error correction in his theory of science, and it applies to morals, too. Deutsch writes:

Could it be that the moral imperative not to destroy the means of correcting mistakes is the only moral imperative? That all other moral truths follow from it?

In other words, Deutsch’s moral imperative states that one should avoid entrenchment of mistakes. That includes avoiding mistakes which self-entrench.

What’s the connection to honesty? The question of whether or not to lie is a moral question, and some lies require more lies to cover up the original lie. A husband cheating on his wife has to lie about where he was last night, why he smells like someone else’s perfume, and so on. One lie quickly turns into a whole web of lies. And the longer he waits before coming clean, the more he has to fear the consequences of not coming clean sooner. The lie prevents him from being honest in the future. Epistemologically speaking, the original error becomes harder to correct: it is entrenched. I have to imagine Popper knew about this connection between his epistemology and honesty, but I am not aware of him (or anyone else) having made it explicit.

In rare cases that are beyond the scope of this article, lying is permissible, but even ‘little white lies’ can become self-entrenching. It’s hard to predict. Just the other day, I heard a story about a man and woman getting coffee on their first date. The man accidentally put salt, not sugar, in his coffee. When she pointed out the error, he was flustered and replied that the taste of salt in his coffee reminded him of growing up by the seaside; it reminded him of home. This qualifies as a ‘little white lie’ because he had no bad intentions, considered the lie inconsequential, and really just told it out of embarrassment. He thought that was the end of it, but she would later put a pinch of salt in his morning coffee every day. It made her happy to have remembered this detail about him, and he did not want to ruin that for her. And, again, if he had decided to tell her the truth at some point, then the longer he had waited, the worse her feeling of betrayal. So he kept pretending for decades that he enjoyed his salted coffee.

Avoiding ‘little white lies’ requires a certain amount of strength: he should have come clean the first morning she put salt in his coffee, at the latest. Any later, and people’s integrity is liable to buckle under the weight of their own lie. So they keep lying, thereby adding more weight.

There are additional reasons lying is generally bad – for example, there’s the Kantian argument that lies rob other people of their ability to make informed decisions and result in treatment of them not as people but only as means to an end. But one should avoid lying on the entrenchment point alone. One should also avoid creating incentives for others to lie – that’s why people sometimes declare that they will not press charges if a thief returns their stolen property.

In summation, one reason people should generally avoid lying is that lies can quickly become entrenched, preventing their own correction.

Thanks to Amaro Koberle and Kitt Johnson for discussing this idea with me before publication.