Believing you understand how complex things work far more deeply than you do — a gap only exposed when asked to explain in detail.
Imagine you see a magic trick every day and you think, 'Oh yeah, I totally know how that works.' But then someone says, 'Okay, explain the trick step by step.' Suddenly, you realize you have no idea how the magician actually does it — you just thought you knew because you'd seen it so many times.
The Illusion of Explanatory Depth describes the systematic tendency to overestimate the precision, coherence, and depth of one's own understanding of how things work — from everyday devices like zippers and toilets to complex policies like healthcare systems. Unlike general overconfidence, this illusion is specific to explanatory or causal knowledge, where multilayered mechanisms create many hidden gaps that people fail to notice until they attempt to articulate them. When asked to generate a step-by-step mechanistic explanation, people consistently revise their self-assessed understanding sharply downward, revealing that what felt like thorough knowledge was in fact a shallow, skeletal representation. The effect is strongest for objects and systems with visible parts, where familiarity with the surface appearance is easily confused with understanding of the underlying causal architecture.
The same glitch looks different depending on the terrain. Finance, medicine, a relationship, a team — same mechanism, different costume.
Investors frequently overestimate their understanding of complex financial instruments like derivatives, CDOs, or cryptocurrency protocols. They feel knowledgeable because they can name the products and describe their purpose at a high level, but lack the mechanistic understanding of pricing, risk exposure, or settlement processes needed for sound decision-making.
Patients often believe they understand their medical conditions or treatments deeply because they can describe symptoms and name medications, but they lack causal understanding of how the drug interacts with their physiology. Clinicians themselves may overestimate their grasp of the mechanisms underlying diseases outside their specialty, relying on surface-level textbook summaries.
Students routinely rate their understanding of material as high after reading a textbook or watching a lecture, confusing recognition of concepts with genuine explanatory knowledge. Teachers may overestimate how well students understand material because the students can correctly repeat terminology without being able to explain the underlying processes.
People feel they deeply understand their partner's emotional motivations or family dynamics, but when pressed to explain specifically why their partner reacts a certain way in conflict situations, they find their 'understanding' was a vague narrative rather than a genuine causal model.
Designers and product managers frequently claim to understand how a system's back-end architecture works because they interact with its front-end daily. This false confidence leads to unrealistic feature requests, underestimated complexity in sprint planning, and resistance to engineering pushback on technical feasibility.
Managers overestimate their understanding of operational workflows they oversee, assuming that familiarity with outcomes equates to knowledge of the step-by-step processes. This leads to poorly scoped projects, underestimated timelines, and misinformed restructuring decisions.
Voters and commentators hold extreme positions on complex policies — tax reform, immigration systems, trade agreements — while lacking mechanistic understanding of how those policies actually produce outcomes. Media coverage reinforces this by emphasizing positions and values rather than causal explanations, allowing the illusion to persist unchallenged.
Leonid Rozenblit and Frank Keil, 2002, Yale University. Published in Cognitive Science as 'The misunderstood limits of folk science: an illusion of explanatory depth.'
In ancestral environments, most practical tasks required only shallow functional knowledge — knowing that a sharp rock cuts, or that fire cooks food — rather than deep mechanistic understanding of underlying physics. Maintaining a confident sense of understanding promoted decisive action and resource use without the paralyzing cost of admitting ignorance. Additionally, our ancestors often operated within communities where knowledge was distributed across individuals, making it adaptive to feel one had access to expertise even when the detailed knowledge resided with others.
LLMs can amplify the IOED by producing fluent, confident-sounding explanations that give users an even stronger false sense that they (or the AI) deeply understand a topic. Users may read an AI-generated summary and mistake its coherent narrative for genuine depth, reducing their motivation to seek true mechanistic understanding. Additionally, AI systems trained on surface-level internet text may themselves encode shallow causal models, generating plausible but mechanistically wrong explanations that are difficult for non-experts to detect.
50 cards are free to preview. Buyers unlock the rest of the deck plus the interactive training — Spot-the-Bias Quiz unlimited, Swipe Deck with spaced repetition, My Blindspots, Decision Pre-Flight, the Printable Deck + Cheat Sheets, and the Field Guide e-book. $29.50$59.
