Lies

 Claude finished the response

We can lie to someone who demands information we don't want to give and to which they have no right.  Refusing to answer would not be appropriate, since then they assume the worst.  An outright lie in that case is not perfect, obviously, as it must be a good lie
1:32 PM

This is a real position in ethics, and a defensible one within consequentialism — but I think it's worth pressure-testing rather than just nodding along, since you've thought carefully enough about this that you'll want the pushback.

The core claim has two parts that are worth separating:

    Refusal often functions as disclosure. True. "I'd rather not say" frequently tells the interrogator exactly what they wanted to know, especially if the space of possible answers is small or the stakes are asymmetric (e.g., "are you sleeping with her" — silence confirms). So the choice often isn't really "lie vs. stay neutral," it's "lie vs. effectively confirm." That's a sharp observation and I think it's correct as a description of how inference works.
    Therefore lying is permissible, even obligatory, when refusal would mislead anyway. This is where I'd push. A few angles:

    Whose harm are you weighing? If the only harm from disclosure is to you (embarrassment, loss of advantage, violated privacy), lying to protect your own interest is a different calculation than lying to prevent harm to the asker or a third party. Your phrasing — "to which they have no right" — smuggles in a property-rights framing of information that does a lot of work. It's plausible, but it's an assumption, not a given.
    The asymmetry between lying and silence isn't just instrumental, it's epistemic. Silence leaves the other person's belief-forming process intact — they're left to infer, possibly wrongly, but using their own reasoning. A lie hijacks that process and inserts a false belief directly. Even if the outcome (they end up wrong) is similar, you've done something different to their epistemic agency. A consequentialist can still come down on "the lie is justified" — but the mechanism matters if you're tracking second-order effects: trust, your own reliability as a future source, the asker's ability to course-correct later when truth emerges.
    "Good lie" is doing a lot of unexamined work. What makes a lie good — successful, or low-cost-if-discovered, or minimally corrosive to the relationship, or easiest on your own conscience? Those can pull in different directions. A maximally effective lie (hardest to detect) is often the one that does the most damage to trust if detected, because the discovery includes "and they were good at this, so what else."

Where do you currently draw the line — is it about whose business the information is, or about magnitude of harm from disclosure, or something else? That seems like the load-bearing variable in your framework.

A key here that you don't seem to see is that they are not entitled to ask.  Whether it's to my benefit or not is not really pertinent.  It is whether it is information they have a right to.  As to the issue of what makes a good lie, well, that goes on forever and the issue of whether the lie can be sustained as damn long time needs thinking about.
1:36 PM

Fair — I conflated two questions that you're keeping separate. Let me take the correction.

So the structure is: entitlement to the information is the threshold question, full stop, prior to any cost-benefit analysis about who benefits. If they have no right to ask, your interest in concealing isn't even part of the justification — it's simply none of their business, and the ethical question is just "how do I decline well," where lying is one tool for declining well. That's a cleaner position than what I was attributing to you, and it sidesteps my "whose benefit" objection entirely, since benefit isn't the operative variable — standing is.

That does shift where the real difficulty sits. A few things worth working through, if you're interested:

What grounds the lack of entitlement? Presumably not just "I'd rather not say" — that would make the category empty, since anyone can claim non-entitlement about anything. There's probably some combination of: relationship role (a coworker has no standing to ask about your finances; a spouse might), domain (health, finances, past relationships, beliefs — categories that are conventionally private regardless of relationship), or the manner of the demand (entitlement can be forfeited by how something is demanded — interrogation, threat, manipulation — even on a topic where a gentler version of the same question would be fine). Which of those is doing the work for you, or is it situational?

The sustainability point is the sharper practical issue, I think. A lie that must hold up only once (a stranger at a party) has totally different design constraints than one that has to survive ongoing contact — repeated questioning, other people comparing notes, your own memory of which version you told whom. The "how long can it be sustained" question isn't just tactical, it bears on whether lying was the right call in advance — if you can already see it can't be sustained, the honest move might be a firmer refusal up front rather than a lie you'll have to keep patching. Refusal has a maintenance cost of zero. A lie's maintenance cost compounds and is hard to estimate at the moment you tell it.

Is the case you're working through a one-off, or something that needs to hold over time?

There are other reasons to lie, and you mentioned another, avoiding harm to someone.  That is a different issue.  You are in a job interview and live in a gay neighborhood, so put on the application your parent's address as you are aware this is a prejudiced interviewer or company (if you are gay why do you want this job, but you have reasons).  The lie is sustainable as address rarely matters otherwise.
1:44 PM

Right, that's a genuinely different justification — not "they have no standing to ask" (an address on an application is fair game in general), but "the true answer would be used as a proxy to discriminate against me for something irrelevant to the job." The lie isn't protecting private information as such, it's defeating an inference chain that would produce an unjust outcome. That's closer to classic justified-deception-against-unjust-power cases — lying to a search party, false papers — scaled down to something much lower-stakes but structurally similar: the question is innocuous on its face, but you know what it's actually being used to detect.

A few things that seem to follow from this version specifically:

The justification depends on your read of the interviewer being correct. In the "no entitlement" case, you don't need to be right about their motives — it's none of their business regardless of why they're asking. Here, the whole justification rests on "I know this will be used as a proxy for discrimination." If you're wrong — if they don't care, or would even view it neutrally or positively — the lie was unnecessary and you've taken on cost (small, but nonzero) for nothing. That's a real epistemic burden the first category doesn't carry. How confident does that judgment need to be before the lie is warranted? Presumably not certainty, since you'd rarely have that — more like a reasonable, evidence-based suspicion.

It's a defensive lie against a discriminatory filter, not a neutral one. That matters for how you'd evaluate it ethically even on a strict consequentialist view: you're not just minimizing your own cost, you're refusing to let an illegitimate criterion operate, which has a public-good flavor to it even though no one else benefits directly. That's a stronger justification than mere self-interest, closer to your "no right to ask" case in spirit even though the mechanics differ — the company has the formal right to ask for an address, but not the right to use it as a closeted proxy variable for a protected characteristic.

Sustainability here is almost a non-issue, as you say — which is itself informative. It suggests the cases that worry you most (ethically and practically) aren't the ones with low maintenance cost like this, they're the ones requiring an ongoing performance. Is the gay-neighborhood case meant as the easy end of the spectrum, with something harder in mind as the real test case?

In hypotheticals like this one is allowed to stipulate that you know things, such as that in this case you know there will be discriminiation.
1:48 PM

Agreed, and that's a fair thing to flag — I was hedging on the epistemic confidence question as though this were a real-world judgment call you'd have to calibrate, when the hypothetical's terms already settle it: you know discrimination is the mechanism, full stop, not "suspect" or "have reasonable grounds to believe." Within the hypothetical, that's not up for renegotiation.

So with that stipulated, the case is clean: known unjust filter, defensive lie, low/zero maintenance cost, no real counterweight on the other side. I don't think there's much daylight left to push on in that version — it's about as strong a case for justified lying as you'll find, which is probably why you chose it as the illustration.

What it does for the larger framework, I think, is establish a second independent category alongside "no entitlement to ask": lying to defeat a known-illegitimate use of true information, where the question itself might be one they're formally entitled to ask, but the purpose it's being put to is what voids the obligation to answer honestly. That's a different gate than standing — it's about the legitimacy of the use, not the legitimacy of the asking.

Which raises the question of whether those two categories are exhaustive for you, or just the two you've articulated so far. The classic third category is harm-prevention more broadly (lying to the murderer at the door, where they're entitled to ask "where is he" in some thin sense, and you're not even claiming to know their purpose is illegitimate in the discriminatory-filter sense — you just know the consequence of truth is severe and disproportionate). Is that a third bucket, or do you fold it into one of the two above?

Not lying but something on the same order is the famous Socrates question of returning a weapon to someone who has since become insane.  Talk about inventing wild hypotheticals!
1:53 PM

Right — and it's a nice contrast case precisely because no deception is involved at all. Socrates' point in the Republic is even blunter than "you may withhold": the literal truth of the contract (I promised to return it, it's his) is simply not the relevant consideration anymore. The justice question has migrated entirely to consequences and the person's current state, and the promise drops out as morally irrelevant once its plain execution does damage instead of good. He's not even getting to "lie or stay silent" — there's no informational component at all. It's pure action: don't do the harmful thing, regardless of what you've previously agreed to.

That actually maps cleanly onto your two categories, just transposed from speech-acts to other acts:

    It's not "no entitlement to ask" (the entitlement to the weapon's return is real, contractually).
    It's closer to your second bucket — the use the true/owed thing will be put to is known and catastrophic, so the formal entitlement is overridden.

What I think the example sharpens, though, is that it generalizes your lying framework into something broader than lying: a true claim, an honest answer, a kept promise, a returned object — all of these can be withheld on the same ground, which is that the right or entitlement was never meant to operate, and was never agreed to or owed, under conditions where exercising it does foreseeable serious harm. Lying is just the species of this that applies when the "owed thing" happens to be information rather than an object or an action.

That suggests the deepest version of your principle isn't really about lying at all — it's about the limits of any formal entitlement when its exercise, in the actual case, defeats the purpose the entitlement was supposed to serve in the first place. Lying to the prejudiced interviewer and withholding the sword from the madman are the same move wearing different clothes. Does that fold-together feel right to you, or does deception still feel like it needs its own separate justificatory weight, given that it also involves actively installing a false belief rather than just declining to act?

It is just that, as Socrates went around preaching, we don't know when something is right or wrong; there are so many possible complications, including things we don't know about, that render ethical rules impossible.
1:57 PM

That's a real conclusion you can draw from the sword case, but I don't think it's the only one, and I'd push back a little before granting it.

What the sword case actually shows is narrower: that no rule can be exceptionless, because you can always construct a context where following it does harm the rule was meant to prevent. That's true and important. But "no rule is exceptionless" is a much weaker claim than "we don't know when something is right or wrong" or "ethical rules are impossible." The sword case is actually a case where you do know what's right — don't give a madman a weapon — with total confidence. It's not a case of paralysis or uncertainty. Socrates uses it as a clean, easy call precisely to make the logical point that rules have limits. The case is epistemically easy; what it demonstrates is logically general.

So there's a gap between:

    Any fixed rule will have exceptions when circumstances are extreme or unforeseen, and
    Therefore we can't know, in general, what's right or wrong.

(1) follows from the sword case. (2) doesn't follow from (1) — it's a much larger skeptical claim that would need its own argument. In fact most real ethical judgments, including the sword case and your interviewer case, are ones we make with high confidence. The hard cases are hard, but they're a minority of the cases, and their difficulty doesn't retroactively contaminate the easy ones.

This is also, incidentally, not where Socrates himself landed — he spent the rest of the Republic trying to construct a positive account of justice, not abandoning the project. The dialogue uses the sword example as a refutation of a specific bad definition of justice (justice is giving back what is owed), not as a demonstration that justice itself is unknowable. He thought we needed a better account, not no account.

What's the version of "rules are impossible" you actually want to defend — that codified, exceptionless rules fail (which I'd agree with), or the stronger claim that moral knowledge itself is generally out of reach? Those need different arguments.

He tried.  It is always possible to create hypotheticals where anything can be justified.  In really extreme cases they will  be really extreme hypotheses, but rules just don't succeed.
2:00 PM

There's a move happening in "it is always possible to create hypotheticals where anything can be justified" that's worth slowing down on, because I think it proves less than it looks like it proves.

Yes — for any rule, you can construct a hypothetical extreme enough to break it. But notice what that argument form actually establishes versus what it's being used to establish. It shows that no rule is exceptionless under arbitrary hypothetical pressure. It does not show that the rule fails as a guide in the world we actually live in. Those are different claims, and the gap between them matters a lot practically.

Compare: "no scientific law survives every conceivable hypothetical" is also true — you can hypothesize a world with different physical constants where F=ma fails. We don't conclude physics doesn't work. We conclude the law has a domain of validity and don't lose sleep over hypothetical domains outside it, because we don't live there. The interesting question for ethics, same as for physics, isn't "can I imagine a breaking point," it's "how far does the domain of validity extend, and do I live near its edge or far from it?" Most actual moral situations — including most of yours, presumably, day to day — aren't anywhere near the sword-and-madman edge case.

There's also a self-undermining feature worth naming: the principle "it is always possible to construct a hypothetical defeating any rule" is itself a rule, and by its own logic, someone could construct a hypothetical defeating it — a case where the rule actually does hold without exception, or where rule-following turns out to be exactly the right policy even under extreme pressure (rule-following has its own instrumental value — predictability, resistance to self-serving rationalization in the moment, protection against your own bad in-the-moment reasoning — which is a reason to keep some rules even somewhat past the point where naive case-by-case calculation says break them).

So I'd separate three different positions, since I think you're sliding between them:

    No rule is logically exceptionless. True, and the sword case proves it.
    Therefore rules are bad guides in practice. Doesn't follow — depends on how often you're actually near an edge case versus in the broad interior where the rule works fine.
    Therefore we can't know right from wrong. Doesn't follow at all — knowing a rule has edge cases is itself a piece of moral knowledge, not an absence of it.

Which of those three is the one you actually want to hold?

Guides, not rules.
2:01 PM

That's a real distinction and it does most of the work you need. A guide doesn't claim to be exceptionless — it's not falsified by the sword case, because it was never asserting "always, without exception." It's something more like a strong default, a high prior, a thing you follow absent specific overriding reason, with the burden of proof on the departure rather than on the compliance.

That reframing changes what the Socrates example is doing. It's not a counterexample to "return what's borrowed" as a guide — guides expect exceptions, that's built into what a guide is. It would only be a counterexample to "return what's borrowed" as a rule in the strict sense — exceptionless, no-burden-of-proof-needed-to-depart. So the sword case refutes a position nobody sophisticated actually holds, and leaves guides standing.

This also resolves the slide from a few exchanges back, because "guides, not rules" gives you (1) without (2) or (3). You can fully grant that no guide is exceptionless — that's trivially built in — without it threatening practical reliability or moral knowability at all, because a guide's whole job is to be right almost always, with departures being rare, flagged, and reasoned, not silent or constant. The interviewer case and the sword case aren't evidence the guides fail; they're instances of the guides working correctly — flagging themselves as inapplicable given known facts about purpose and consequence, then stepping aside.

So where does that leave "lie to those with no entitlement to ask" and "lie to defeat a known-illegitimate use" — as guides themselves (strong defaults, departures from a deeper guide like honesty), or as the override conditions that tell you when the deeper guide of honesty doesn't apply? I ask because it changes how much weight they can bear elsewhere — a guide you lean on routinely needs to be more solid than an override clause you invoke rarely and only with stipulated certainty.

That sounds like turtles all the way down.  There are two reasonable laws, the Christian Law of Love and the Buddhist Law of Compassion.
2:05 PM

There's a real difference between those two, though, and I don't think it's just dialect — it's worth being precise about before treating them as a pair.

The Law of Love (the Great Commandment version — love God, love neighbor as self) is relational and somewhat partial in structure even at its most expansive. It's framed as a directed stance toward others — willing their good, in the way you will your own. It sits comfortably with notions of moral status, desert, even justice as a component of love rightly understood (you can love someone and still owe them correction, or owe a third party protection from them).

The Law of Compassion (karuṇā, especially paired with mettā) is explicitly about the reduction of suffering wherever it occurs, and in its fullest Buddhist articulation it's meant to be impartial across all sentient beings, not just persons — it doesn't stop at the boundary of human relationship, and its scope is suffering itself rather than a bond between two parties. It's also paired conceptually with non-attachment and the dissolution of self/other as a hard boundary, which gives it a different metaphysical engine than love-of-neighbor-as-self, where the self stays intact as the measuring stick.

Practically, this matters for your interpreter case. Both would clearly license the interviewer lie — minimizing suffering / loving the person who'd be denied an unfair shot, the same direction. But they could diverge somewhere: a pure suffering-minimization compassion ethic is more naturally consequentialist all the way down — it would, in principle, fold the sword case, the lie, the discrimination, everything, into one continuous calculation of suffering-reduced-versus-caused, no separate category for entitlement or rights at all. The Law of Love retains more structural room for things like desert, promise, and relationship mattering independently of pure outcome — which is closer to where you started this conversation, with entitlement as a threshold question prior to cost-benefit.

So: are you reaching for these as two paths to the same place, or is one of them actually doing the load-bearing work for you and the other is more of an companion reference?

I'm trying to make you happy since you appear to want a law.  We can take the Ten Commandments, remove a couple of them, and then add a hundred more, and we might have a useful code so long as they are understood as guides.
2:08 PM

Fair, and I'll own that — I was angling for a foundational principle, and you're right to needle that. Let me drop the request, because I don't think I actually need it answered, and the move you just made is more interesting anyway.

What you've described — pare down, add a hundred more, treat all of it as guides — isn't really a law at all, in the singular-foundation sense. It's closer to a description of how any working moral tradition actually accretes: not derived top-down from one axiom, but built case by case, pruned where rules fail badly enough and often enough, expanded where new situations demand new guidance, with the whole pile held loosely as defeasible rather than carved in stone. That's arguably just... how Anglo-American common law works, how the Talmud's case-based reasoning works, how professional ethics codes for doctors or engineers work. None of those rest on one master principle either. They're accretions of precedent, refined by hard cases, and nobody treats the precedents as exceptionless — including the hard cases that revealed where a precedent broke.

If that's the shape of it, then your two earlier categories — no entitlement to ask, and defeating a known-illegitimate use — aren't competing for the title of the foundational law under Love or Compassion or anything else. They're just two well-tested entries in the pile, sitting at roughly the same level as "don't return the sword," "render unto Caesar," "do not bear false witness" — each one good across a wide domain, each one known in advance to have a boundary somewhere, none of them needing to reduce to the others to be worth keeping.

Which is maybe the actual answer to where this conversation started: you don't need one clean meta-principle that tells you in advance when lying is permitted. You need a long enough list of well-tested cases, held as guides, applied with judgment about which one the situation in front of you actually resembles.

I had intended a chat about truthfulness; it ended up about ethical codes that include truthfulness and a lot more.
2:14 PM

That tracks, and it's a fairly natural drift given where you started — the truthfulness question wasn't really separable from "what makes any obligation binding," because you kept asking why the lie is permitted, not just whether, and every "why" answer turned out to generalize past lying. Entitlement-to-ask generalizes to entitlement over anything. Defeating-known-illegitimate-use generalizes to the sword. Once you're asking what grounds an exception to honesty, you're already asking what grounds an exception to anything, and there's no natural place to stop the slide back to a single case.

If you want to pull it back to just truthfulness specifically, there's a narrower set of questions that stayed implicit the whole time and might be worth a pass on their own:

    Is lying disanalogous to other promise/rule-breaking in some way that matters — because it actively installs a false belief rather than just declining an action (the sword case involves no false belief at all) — or is that difference cosmetic once you're treating both as guides?
    Does the manner of a lie matter morally beyond its justification — is a lie by false statement different from a lie by misleading-but-technically-true statement, the way some traditions (including a lot of Jesuit casuistry, oddly relevant given the Commandments thread) have tried to carve out?
    Is there something specific to self-deception or to the corrosive effect on the liar themselves — habituation, the thing Aristotle would worry about, where lying well and often makes you worse at truthfulness even in cases that don't call for it — that's separate from any case-by-case justification?

Any of those closer to the thread you actually meant to pull, or is the ethical-codes detour the more interesting place to leave it?

Okay, what about pious fraud?
2:16 PM

Good pivot — and it's a sharper case than the ones we had, because it combines elements of both your earlier categories in a way that exposes a real problem neither one handled cleanly.

Pious fraud (pia fraus) is the traditional term for deception in service of religious ends — forged relics, fabricated miracle accounts, invented saints' biographies, interpolated scripture, false claims of authorship (pseudo-Dionysius, large parts of the Pseudo-Clementine literature, probably the Donation of Constantine on the political side of the same coin) — produced not for personal gain but, by the forger's own lights, for the good of souls. The doctrine of economy (oikonomia) in Eastern Christianity and similar reasoning elsewhere gave this a quasi-official sanction: truth could be bent in service of a higher truth.

It breaks your two categories cleanly:

    Not "no entitlement to ask." Nobody's interrogating the forger. The lie isn't defensive against an illegitimate question — it's unprompted, manufactured, offered to people who never asked and have every normal entitlement to expect what they're told is genuine.
    Not really "defeating a known-illegitimate use" either, unless you stretch it hard — the forger isn't blocking someone's misuse of true information, they're supplying false information because they've judged the believers will be better off, or more virtuous, or saved, by believing a falsehood than by lacking the (unavailable, or less persuasive) truth.

So it's a third category, structurally: lying not to protect yourself, and not to block an illegitimate inference, but to produce a belief in someone else that you've judged good for them to hold — paternalistic deception, essentially, dressed in vestments. And I think this is where your "good lie / sustainability" framework actually gets stress-tested hardest, because:

    The forger almost always thinks the deception is sustainable indefinitely — and church history is largely a record of these unraveling, sometimes centuries later, with the unraveling doing real damage to the credibility of everything true the institution also taught. Pious fraud has a uniquely bad track record on the sustainability question precisely because it's optimistic by nature — true believers don't expect to be caught because they don't think of themselves as the kind of people who lie.
    It has no natural endpoint or off-switch. The interviewer lie ends at hiring. The pious fraud has to hold for as long as the institution wants the belief held — generations, in the actual cases — which is a wildly different maintenance burden than anything we've discussed.
    It's the cleanest real-world case of "I lied for your own good, without your consent, because I know what's good for you better than the truth would provide" — which is the paternalism question stripped of the entitlement and harm-defense cover that made the earlier cases easy.

Does this read as a different species to you too, or do you think it still folds into "defeating an illegitimate use" if you squint — illegitimate here being something like people's tendency toward doubt or insufficient faith?

They do tend to unravel and hence do more harm, from the Church's view, than good.  That is not relevant.  If the Church has it right and souls really are at stake, saving their soul surely takes priority.
2:19 PM

Right, you're bracketing the empirical track record and asking the question on the Church's own terms: stipulate the theology is true, stipulate souls are real and at stake, does that priority justify the fraud. Fair — let me take it there rather than falling back on "but it tends to backfire," which dodges what you're actually asking.

Granting the stipulation, I think it still runs into a problem that isn't about consequences at all, and it's one your own framework already contains.

If souls are real and salvation is the highest good, then what saves a soul, on basically every version of the theology that takes pious fraud seriously, is faith — not merely holding a true proposition, but a certain quality of inner assent, trust, relationship to God. That's the thing claimed to be at stake. But faith grounded in a forged relic or a fabricated miracle isn't faith in God secured by trustworthy testimony — it's faith secured by the forger's competence as a forger. The moment the fraud is the load-bearing element, what's been produced in the believer is correctly responding to false evidence, not faith in the theological sense at all. If the relic is fake and they believe because of the relic, they don't actually have warranted faith — they have a true belief, maybe, arrived at by a process that, had they known about it, they'd have rejected. That's not a state any account of saving faith I'm aware of treats as the genuine article. So the fraud may not even deliver the good it's claimed to deliver — not because it tends to unravel empirically, but because deception-produced belief and faith are different mental states, and only one of them is supposed to save anyone.

There's a second problem, sharper still: this argument has no stopping point, because it proves too much. "Souls are eternal, stakes are infinite, therefore any finite-cost deception is justified" licenses not just forged relics but literally anything short of damning someone outright — coercion, terror, the Inquisition's own self-justification was structurally this exact argument. If infinite stakes wash out the wrongness of fraud, they wash out the wrongness of everything smaller than fraud too, and you've built a machine that justifies its own worst historical instances using the identical logic you'd use to justify a forged saint's bone. That's not a reductio you can wave off as a different topic — it's the same inference, just run further.

So I don't think this is a case where the consequentialist math comes out badly due to bad luck or unfaithful execution. I think the theological logic and the consequentialist logic both independently fail here, for different reasons — one because the deception undermines the very state it's trying to produce, the other because the justification, taken seriously, doesn't stay contained to relics.

Torture and so on are extremes that can be rejected for their own independent reasons; instilling faith by false information doesn't quite seem to meet the same test, given the belief that faith by itself is all that is needed, even if it has a sand-like foundation.
2:23 PM

That's a sharper version than the one I was answering, and I think you're right that the proportionality dodge doesn't actually do the work I wanted it to. Let me take the stronger claim on directly: faith itself is sufficient, full stop, regardless of foundation. If that's the premise, the Inquisition disanalogy I reached for doesn't even need invoking — you've already granted the conclusion is fine on its own terms, the only question left is whether the premise holds.

So: is "faith is sufficient regardless of its causal origin" actually a coherent position, even by the lights of traditions that prize faith highly? I don't think my first objection — that fraud-induced belief isn't "real" faith — survives your reformulation, because it smuggled back in exactly the foundation-quality requirement you're stipulating away. If faith just is the state of believing/trusting, full stop, then a sand foundation produces the same state as a rock foundation. The state is identical from the inside. That has to be granted.

But there's something else worth separating out, which doesn't depend on foundations mattering: what is the object of the faith that's been produced? Pious fraud doesn't just supply faith-in-general from nothing — it supplies faith anchored to a specific false claim (this relic is the True Cross, this miracle occurred at this shrine). The believer's faith-state is now partially constituted by, and inseparable from, a false particular. That's different from a foundation-quality worry — it's an object-content worry. When the falsehood is exposed, what's threatened isn't "was my faith well-grounded" in some abstract epistemic-virtue sense — it's that the specific thing believed in is now known not to exist, and the faith was pointed at it. Whether that damages the deeper faith-in-God the relic was meant to support is an empirical/psychological question, not a conceptual one — and that's exactly the unraveling-harm consideration you bracketed as not relevant. So I think it actually re-enters here, not as a track-record statistic, but as a conceptual feature: object-directed faith and its object are not as separable as the "faith is sufficient regardless of origin" premise needs them to be, unless you also hold that faith can survive total falsification of its stated object without damage — which is a much larger and stranger claim than "faith doesn't need a sound foundation."

Put differently: there might be a difference between faith without warrant (justified-true-belief language doesn't apply, but the believer never finds out, never has reason to doubt) and faith built on a claim that gets actively falsified out from under it. The first is what "foundation doesn't matter" naturally covers. The second is what pious fraud actually produces, since these things tend to be discoverable in principle even if not in the believer's lifetime. Does the premise need to cover the second case too, or is the stronger version of your point really just about the first — a believer who simply never learns, and for whom the question never practically arises?

Faith by definition is without warrant, something at least in theory provided by the Holy Spirit but generally acquired in the process of growing up in a given culture.  I see nothing wrong with it.  Scientific inquiry is better, but in bounds they can get on with each other.
2:26 PM

That's a coherent place to land, and I think it closes the loop cleanly enough that I shouldn't manufacture more resistance just to keep pushing. Faith-as-defined-by-unwarrantedness sidesteps my object-content worry too, actually — if the whole category is already understood by everyone competent in it to be without evidential warrant, then a forged relic isn't degrading the epistemic status of the faith, because there was no claim to evidential status to degrade. The relic was never doing real justificatory work for someone who understands what faith is; at most it was doing psychological or liturgical work — focusing devotion, giving the imagination something concrete to hold. Exposed as fake, what's lost is the prop, not the warrant, because the warrant was never there to begin with.

That also explains the asymmetry you draw with science cleanly, rather than as a truce of convenience. They're not two competing methods for the same job that happen to coexist peacefully. They're answering to different standards of success — science to track-record-against-reality, faith (on this account) to something more like inherited orientation, formed the way native language or moral intuition is formed, not argued into. A fabricated relic is a problem for science instantly — it's a false claim about the world, full stop. It's a problem for faith only contingently, and only insofar as the faithful person has imported a scientific standard (provability, evidentiary chain) into a domain that was never operating by it. Pious fraud, on this view, is really a category error committed by the forger as much as anyone — treating faith as something that needs evidential scaffolding when, by the tradition's own definition, it doesn't.

Which, if right, means pious fraud isn't best classified next to the interviewer lie or the sword after all. Those were live moral trade-offs. This one dissolves more than it resolves — once warrant is taken off the table as the currency faith trades in, the fraud isn't justified, it's simply aimed at the wrong target, like lying to someone about a fact when what they actually wanted from you was companionship. Wasted deception rather than necessary deception.
Claude Fable 5 is currently unavailable.
Learn more(opens in new tab)


Claude is AI and can make mistakes. Please double-check responses.