Belief bias - logicae.org

About

Belief bias is the judging of an argument by the believability of its conclusion rather than by whether the conclusion follows. The verdict on the inference is read off the destination; the route is never inspected.

Taxonomy - Validity-assessment bias, conclusion-driven evaluation. It is the confusion of plausibility with inferential form; truth and validity, made interchangeable.

Informal Example - "The conclusion sounds true, so the argument must be valid."

Minimal Symbolic Form

Canonical Form:

Plausible(C) ⊢ Valid(Premises ∴ C)

Where:

C: the conclusion
Plausible(C): C seems believable
Valid(Premises ∴ C): C follows from the premises

Whether C is likely and whether C follows are different questions; the bias answers the first and files it under the second.

Logical Analysis

Invalidity Statement:

Truthlike(C) ⇏ Valid(A)

The plausibility of a conclusion does not entail the validity of an argument.

Missing Necessary Condition

Validity requires preservation of truth from premises to conclusion:

¬∃M (Premises true in M ∧ C false in M)

Believability is not a substitute for model-theoretic consequence.

Argument-Evaluation Model

Biased Model:

Plausible(C) → Accept(Valid(A))

Corrective Model:

Form(A) ∧ Consequence(Premises,C) → Accept(Valid(A))

The conclusion's believability should be screened off when evaluating validity.

Semantic Formalization

Argument validity is semantic:

Premises ⊨ C

Belief bias substitutes:

Believable(C) ⊨ C

This changes the question from consequence to prior plausibility.

Epistemic Representation

𝔼(Believable(C)) ⇏ □(Valid(A))

A believable conclusion is not evidence that the inference is valid.

Countermodels

True Conclusion, Invalid Argument

C ∧ ¬(Premises ⊨ C)

A conclusion can be true while the argument for it is invalid.

Unbelievable Conclusion, Valid Argument

Premises ⊨ C ∧ Implausible(C)

An argument can be valid even when the conclusion is surprising.

Historical Background

The distinction between truth and validity is central to ancient and medieval logic, even though "belief bias" is a modern label.

Aristotelian syllogistic and later scholastic logic repeatedly separate consequence from the material truth or apparent plausibility of the terms.

The Neo-Latin phrase praeiudicium credulitatis is only descriptive. The older logical ancestor is the formal distinction between valid consequence and persuasive conclusion.