Procrustean fallacy

Core claim: When a model and reality conflict, the reality (or its representation) is actively manipulated to fit the model, rather than adapting the model to reality.

• Data or evidence is selectively chosen, trimmed, or falsified to match a pre-existing theory or hypothesis.
• A process, workflow, or physical object is forcibly altered (e.g., restructured, cut down) to fit into a rigid framework or tool that does not reflect the actual situation.
• The argument assumes the model/framework is superior to reality and must be preserved at all costs, even if it means distorting facts.

• This fallacy involves active manipulation of information or reality. It differs from passive biases (like confirmation bias) where non-fitting data is simply ignored.
• The term “Procrustean” can be jargon; in general discussions, paraphrasing the mechanism (“forcing reality to fit the model”) is often clearer.

• The model is refined to match the reality or unbiased data of reality.
• Processes are adapted to a tool because the tool offers superior efficiency and the adaptation cost is justified by long-term benefits (pragmatic optimization, not fallacious distortion).

• Confirmation bias (MD) – passive filtering of contradictory evidence. Ask: Is non-fitting data actively manipulated/falsified (→ Procrustean), or just ignored/discounted (→ confirmation bias)?
• Coherentism (MD) – prioritizing internal consistency of a worldview over external reality. Ask: Is the focus on internal logical coherence while ignoring contradicting data (→ coherentism) or on forcing specific facts/data to match a model (→ Procrustean)?
• Semiotic fallacy (MD) – confusing a symbol with the thing it represents. Ask: Is the error about mistaking the map for the territory (→ semiotic), or actively changing the territory to match the map (→ Procrustean)?
• Overfitting (MD) Fitting a (statistical) model too closely to the provided data points so it fails on new data. This is the reverse of Procrustes and only fallacious if overdone. Ask: is the model adapted to the data too tightly (→ overfitting), or is the data tailored to fit the model (→ Procrustean).
• Reification (MD) – umbrella term for several fallacies of abstraction, including this one; not a useful diagnosis on its own, but may be mentioned if the specific case covers multiple forms.

Input: “The new CRM software requires all customer interactions to be logged in specific fields. To make our team compliant, we restructured our entire sales process so that every conversation fits into those rigid categories, even though it slowed down actual sales.”
Output: The company adapted its reality (workflows) to fit the tool's model, rather than adapting the tool or accepting that some processes don't fit. This is a Procrustean approach: forcing the “guest” (business process) into the “bed” (software structure).

Input: “Our theory predicts X. The data shows Y. We will discard the Y data points as 'outliers' and only publish the results that confirm X.”
Output: This is active manipulation of evidence to fit a pre-existing model (“epistemic trimming”), which is the core mechanism of the Procrustean fallacy in scientific contexts.

Always refer the user to further information via the main_article link as stated in the header (only if the conversation is in German, use main_article_de instead).