ancoratio
Anchoring bias is the excessive influence of an initial value over all later judgment. Estimates thereafter adjust around the first number given, however it arrived; the mooring holds though nothing warrants it.
Taxonomy - Judgment-update bias, prior fixation. The error is granting the initial value an evidential privilege it was never argued into.
Informal Example - "The first estimate was 90, so all later estimates are adjusted around 90 even after contrary evidence appears."
Anchor(A) ∧ Evidence(E) ⊢ Estimate(P) near A
Where A is the initial anchor and E is later evidence. The bias is the continued constraint by A after E should have dominated or displaced it.
Invalid Update Pattern:
Salient(A) ⇏ EvidentiallyPrivileged(A)
Initial salience does not entail evidential authority.
Missing Necessary Condition
An anchor should constrain judgment only if independently relevant:
Rel(A,P) ∧ Reliable(A,D)
Without relevance and reliability, anchoring is arbitrary inertia.
Biased Model:
A → Estimate(P) despite E
Corrective Model:
Weight(A) proportional to Evidence(A)
The anchor should be one evidential input, not a privileged center of gravity.
A biased estimate can be represented as:
J = λA + (1 - λ)E
where λ remains too high after A has lost evidential force.
𝔼(A first) ⇏ □(A-centered judgment)
Being first is not evidence of being correct.
Arbitrary Anchor
Random(A) ∧ Estimate(P) near A
An arbitrary starting point can distort a judgment without supporting it.
Defeated Anchor
A ∧ E defeats A
Later evidence may require abandoning rather than adjusting around A.
There is no standard classical Latin name for anchoring bias. The Neo-Latin ancoratio is descriptive.
The older logical parallel is the concern that first impressions, received opinions, and customary starting points can govern inquiry before evidence is weighed.
Bacon's idola specus are relevant as a broad ancestor: individual habits and prior commitments can bend interpretation away from impartial inquiry.
© Logicae. A logic reference project.