The Limits of Self-Knowledge

The injunction to know oneself is ancient and almost universally endorsed. Its difficulty is less frequently acknowledged. We have strong motivations to misrepresent ourselves to ourselves — to protect self-esteem, to maintain consistency, to avoid the costs of confronting qualities we would prefer not to have. These motivations operate largely outside conscious awareness, which is part of what makes them effective.

Psychological research has established that people's accounts of their own mental processes are often inaccurate. When asked why they made a choice, they provide reasons — but controlled experiments suggest that these reasons are frequently post-hoc constructions rather than accurate reports of causal factors. This is not deliberate deception. It is a feature of how self-knowledge is generated: as a narrative rather than an inspection.

The philosopher Eric Schwitzgebel has argued, on the basis of a careful reading of introspective reports and experimental data, that we are very poor judges of our own mental states. We misattribute emotions, confabulate reasons, and systematically overestimate our consistency. This is deeply counterintuitive, because introspection feels authoritative. It feels like direct access.

What is perhaps most interesting about this finding is what it implies for the aspiration to self-improvement. If self-knowledge is unreliable, on what basis do we identify the aspects of ourselves we want to change? And if our accounts of our own behaviour are inaccurate, how do we assess whether we are improving or merely telling ourselves a different story?

These are not reasons for passivity. They are reasons for seeking external feedback — from people who will give honest accounts — and for holding self-assessments loosely, with something like the scepticism one would apply to an unreliable source.