Reinterpreting Locke's Latent Fideism — Epoche C2
The chapters at issue, and what they say Locke devotes the closing chapters of Book IV of An Essay Concerning Human Understanding to one question: how strongly may a person believe a proposition that cannot be demonstrated? Chapter xv, "Of Probability", defines the relation; chapter xvi, "Of the Degrees of Assent", supplies the rule for regulating belief by it; chapter xvii distinguishes propositions according to reason, above reason and contrary to reason; chapter xviii, "Of Faith and Reason, and their distinct Provinces", applies the apparatus to revelation; and chapter xix, "Of Enthusiasm", added only in the fourth edition of 1700, rules out belief that offers its own intensity as its credential. The argument of this note concerns the fit between the rule in xvi and the account of faith in xviii — and, specifically, the claim that they do not fit. Before that argument can be made, two things in the earlier version of this note have to be corrected, because the thesis as previously stated rested on both. First, the phrase 'uncertainty of probability' is not Locke's. It is not a term of art in the Essay , and I can find no chapter in which he uses it or anything answering to it as a technical notion. The vocabulary Locke does use in these chapters is "grounds of probability", "degrees of assent", and a scale that runs, in his own listing, from full assurance and confidence down through belief, conjecture and doubt to disbelief. An interpretation cannot be built on a term its subject did not employ, and the version of this thesis that turned on 'uncertainty of probability' as a licence for evidentially ungrounded belief has to be withdrawn. What replaces it is a stronger claim, and one that can be checked against the text. Second, the definition of faith previously attributed to IV.xvi.14 — that faith is a firm assent of the mind which, upon the testimony of the proposer, is undoubtedly assented to as true — is a garbled version of a passage that stands elsewhere. Locke's definition of faith is at IV.xviii.2: faith is assent to a proposition not made out by the deductions of reason, but "upon the credit of the proposer, as coming from God in some extraordinary way of communication". IV.xvi.14 says something different and, for the present purpose, more important: that faith, where it genuinely is faith, is an assurance which "leaves no manner of room for doubt or hesitation" — subject to the proviso that we be sure the thing is a divine revelation and that we understand it rightly. Those two sections, read together, are the whole difficulty. The rule, and where it bites Locke's evidentialist rule is not a slogan but a procedure. The grounds of probability, he says in IV.xv, are two: the conformity of a proposition with our own knowledge, observation and experience; and the testimony of others, vouching their observation and experience. Faced with a probable proposition, the mind that proceeds rationally is to examine all such grounds, weigh what tells for and against, and receive or reject the proposition with a firmness proportioned to the preponderance. IV.xvi.1 then makes those same grounds the measure of the degrees of assent. In modern shorthand — the shorthand is mine, since Locke, writing before any calculus of belief was in general use, gives no numbers — the rule is $$\text{firmness of assent to } p \;\propto\; \Pr(p \mid e),$$ with $e$ the total evidence the believer possesses. Nothing in what follows requires more of the formalism than this, plus two further Lockean commitments stated below. Now apply the rule where Locke himself applies it. He distinguishes original revelation, made to the first recipient, from traditional revelation, which reaches everyone else as a report. For the second, IV.xvi.10 lays down a rule about testimony transmitted through hands: the further a report stands from its first source, the less force it has, and — this is the clause that matters — no probability can rise higher than its first original. So the relevant evidence for a believer today is not the revelation but the transmitted report of one, and the credibility of the report is capped by the credibility of its source and degraded at each remove. Let $R$ be the proposition that God revealed $p$ to the original recipient, and $T_k$ the report as it reaches a believer after $k$ transmissions. Locke asserts only that $\Pr(R \mid T_k)$ is non-increasing in $k$ and bounded above by its value at the source; he proposes no decay law, and it would be a misrepresentation to give him one. But a decay law makes the size of the problem visible. Suppose each link transmits faithfully with probability $r$ and that failures are independent — an assumption of mine, not his, and a generous one, since testimony is not usually independent across a chain. Then fidelity through $k$ links has probability $r^k$. With $r = 0.95$ and five removes, $0.95^5 \approx 0.774$. To keep the transmitted probability above 0.99 across those same five links, each link would have to be faithful with probability about 0.998. Above reason, contrary to reason, and the province of faith That is the state of the evidence. The next question is what Locke thinks assent to a revealed proposition is for , and this is settled by his threefold division at IV.xvii.23. Propositions according to reason are those whose truth we can discover by examining the ideas we have from sensation and reflection; his example is the existence of one God. Propositions above reason are those whose truth or probability we cannot derive from those principles; his example is the resurrection of the dead. Propositions contrary to reason are inconsistent with our clear and distinct ideas. The division does real work. Revelation is superfluous for the first class, since reason reaches those propositions unaided. It is prohibited from the third: IV.xviii.5 holds that nothing contradicting our clear intuitive knowledge can be received as divine revelation, on the gro