Epistemic Innocence¶
A costly cognition can have context-dependent epistemic benefit when no less costly available alternative supplies that benefit.
Core Idea¶
Epistemic innocence is Lisa Bortolotti's proposal that a cognition can be epistemically faulty yet make an otherwise unavailable contribution to an agent's epistemic life. The test has two conditions: a significant epistemic benefit for that person at that time, and no available alternative cognition that would provide that benefit.[^ref-7c734ed6cd26] “Innocent” does not mean true, justified, harmless, or therapeutically desirable.[^ref-2cd9286c75c7]
The counterfactual is not an ideal belief an observer can state; it is a less costly cognition the agent could actually use then. A false belief or memory may preserve attention, information exchange, or some self-knowledge while retaining serious epistemic costs.
Scope of Application¶
The original paper examines elaborated, systematized delusions in schizophrenia and argues for their potential to meet the conditions, with qualifications. The later book considers a wider range of irrational or distorted beliefs. Neither source says every such belief is innocent.[ref-7c734ed6cd26][ref-2cd9286c75c7]
A second original article tests clinical memory distortions: a wrongly dated autobiographical recollection may retain truthful self-related content when a more accurate memory is inaccessible. The authors treat that as context-specific reasoning, not an instruction to leave every inaccuracy unchallenged.[^ref-2984d4e59368]
Clarity¶
The framework forces two questions often collapsed into one: is the belief faulty, and did it enable some epistemic good that the person could not otherwise obtain then? A positive second answer does not erase a positive first answer.
In the delusion argument, relief from anxiety is relevant only if it helps restore attention, inquiry, or learning; the same explanatory belief may later distort new evidence. In the memory argument, knowing the correct date from outside is not necessarily equivalent to the person retrieving a comparably useful accurate recollection.[ref-7c734ed6cd26][ref-2984d4e59368]
Manages Complexity¶
Instead of treating cognitions as simply rational or irrational, the account separates cost, benefit, and alternatives. The last condition matters: if a sounder available explanation could have preserved the same inquiry, the special defense weakens.
The benefits are not assumed to outweigh costs. The assessment isolates what the cognition helps the person do, whether that help is epistemic rather than merely comforting, and whether the same help could be obtained without the error.
Abstract Reasoning¶
Assess a candidate cognition's epistemic shortcomings; identify a concrete benefit rather than a merely comforting feeling; then compare real alternatives available to the agent at the relevant time. The argument is person- and time-indexed. It does not license guessing that a benefit exists in every delusion or that the least costly option was accessible.
Bortolotti's model of a person overwhelmed by highly salient anomalous experiences proposes a possible restoration of attention after an explanatory belief forms, while acknowledging the belief's falsehood and later filtering effects. Bortolotti and Sullivan-Bissett's stylized beach-walk memory is false about when the walk occurred but may preserve a true childhood connection. Both require a separate alternatives test.[ref-7c734ed6cd26][ref-2984d4e59368]
Knowledge Transfer¶
The two-condition evaluation travels from the original delusion analysis to motivated, optimistically biased, or distorted-memory beliefs in Bortolotti's wider project. Transfer is of the evaluative test, not of a clinical verdict about those very different cognitions.[^ref-2cd9286c75c7]
The mechanism does not transfer unchanged: the delusion paper concerns attention and learning under anomalous experience, whereas the memory paper concerns autobiographical retention and communication amid impaired retrieval. “Innocence” here is an epistemic status, not moral innocence or clinical endorsement.
[^ref-7c734ed6cd26]: Lisa Bortolotti, “Epistemic Benefits of Elaborated and Systematized Delusions in Schizophrenia”, British Journal for the Philosophy of Science 67 (2016), publisher PDF in the author's university repository, especially §§6–8. [^ref-2cd9286c75c7]: Lisa Bortolotti, The Epistemic Innocence of Irrational Beliefs, Oxford University Press, publisher description. [^ref-2984d4e59368]: Lisa Bortolotti and Ema Sullivan-Bissett, “The Epistemic Innocence of Clinical Memory Distortions”, Mind & Language 33 (2018), especially §§5.1–5.2.
Neighborhood in Abstraction Space¶
Epistemic Innocence sits in a sparse region of the domain-specific corpus (73rd percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.
Family — Self, Identity & Social Cognition Biases (25 abstractions)
Nearest neighbors
- Information Avoidance — 0.85
- Egocentric Bias — 0.85
- Ostrich Effect — 0.84
- Motivated forgetting — 0.82
- Alternative facts — 0.82
Computed from structural-signature embeddings · 2026-10-08