Nirvana fallacy

“A lock won’t stop a determined thief”

“There’s no point buying a good bike lock. Anyone with an angle grinder can cut through it in a minute.”

True, but the alternative to an imperfect lock isn’t a perfect one; it’s no lock, or a flimsy one. A lock that stops opportunists and slows down the determined is far better than either. This is the clear-cut case: the option is dismissed for failing a standard that nothing on offer meets.

“It won’t reach everyone”

The town library proposes letting people renew books online. At the meeting, a board member objects: “Plenty of our older patrons don’t use the internet, so this won’t help everyone. I don’t think we should do it.”

The objection sounds principled, and it names a real limitation. But the current system, renewing in person or by phone, doesn’t go away; online renewal adds an option for the people who can use it and takes nothing from those who can’t. Measured against the realistic alternative (today’s system), the proposal helps some people and hurts nobody. Measured against an ideal that reaches everyone, it fails, and so does every other option, including the one the board member is defending.

The alternative imagined at its best

Volunteers run a snow-shoveling rota for older residents on their street, and this winter three driveways were missed after one storm. A neighbor writes to the residents’ group: “The rota clearly isn’t working. The town should take this over.”

This looks like the reasonable kind of rejection, one option in favor of another. But the neighbor compares the rota’s actual record with the town’s service as it would ideally run, not with how the town’s service actually performs: how long it takes to come, which streets it covers, what it costs. The economist Harold Demsetz identified this as the core of the error: an existing, imperfect arrangement is judged against an ideal instead of against the real alternatives. The town’s service may well be better, but the three missed driveways don’t show it.

Brooke Noel Moore and Richard Parker’s textbook classifies it as a kind of false dilemma, one that ignores the options between “perfection” and “nothing”:

• Either this option solves the problem perfectly, or it isn’t worth doing.
• It doesn’t solve the problem perfectly.
• Therefore it isn’t worth doing.

As with other false dilemmas, the logic is valid and the first premise is false: there’s a wide range of options between perfect and worthless.

Economists, following Demsetz, frame the same mistake less as a false choice than as a wrong standard of comparison: the ideal is used as the benchmark, when the benchmark should be the next-best real option.

• Perfect or nothing: an option is rejected because it leaves some of the problem unsolved, as in the bike lock example. This is Moore and Parker’s perfectionist fallacy. Bradley Dowden’s Internet Encyclopedia of Philosophy entry defines it as rejecting a proposal “solely because it doesn’t solve the problem perfectly, in cases where perfection isn’t really required.”
• The idealized alternative: the option is compared with an alternative imagined at its best, as in the snow-shoveling example. This was Demsetz’s target (see below).
• Rejection by objections: listing objections to a plan and concluding it should be rejected, without weighing the objections to the alternatives. Richard Whately’s Elements of Logic called this the “fallacy of objections”, committed when someone shows “that there are objections against some plan, theory, or system, and thence infer[s] that it should be rejected; when that which ought to have been proved is, that there are more, or stronger objections, against the receiving than the rejecting of it.” Whately added that “there never was, or will be, any plan executed or proposed, against which strong and even unanswerable objections may not be urged”, and that the same move is used both against reforms and by reformers against what already exists. (Scott Aikin and John Casey read the same passage as a variant of what they call bothsiderism, the family that includes the middle ground fallacy, because Whately also describes the tactic of urging an audience to suspend judgment while there are reasons on both sides.)
• Demanding perfect evidence: refusing any conclusion until a study with no possible weaknesses exists. This version appears in debates about causation (see Correlation vs. causation) and shades into Moving the goalposts when the required standard keeps rising.

A related wrong comparison points the other way: Relative privation dismisses a problem because worse problems exist, measuring against the worst case instead of the best.

Where the name comes from. In a 1969 article, the economist Harold Demsetz criticized what he called the “nirvana approach” in public policy economics, which he described as presenting “the relevant choice as between an ideal norm and an existing ‘imperfect’ institutional arrangement.” He contrasted it with a “comparative institution approach” in which the relevant choice is “between alternative real institutional arrangements.” Demsetz argued that the nirvana approach is especially prone to three logical fallacies, which he named the “grass is always greener” fallacy, the fallacy of the free lunch, and the “people could be different” fallacy. Later writers turned his phrase into “the nirvana fallacy”. Critical thinking textbooks use other names for the same pattern: Moore and Parker call it the perfectionist fallacy, and M. Neil Browne and Stuart Keeley the searching for perfect solution fallacy.

Pointing out that an option is imperfect is a fair contribution to any decision. It becomes an error only when imperfection alone is treated as a reason to reject.

• A better alternative is actually available. Rejecting an option because a real, attainable option does better is exactly the comparison the fallacy skips. It helps to name the alternative.
• The flaw defeats the option’s purpose. A lock that any key opens, or a backup system that loses the files it backs up, fails at the one thing it’s for. That isn’t a demand for perfection.
• The situation really does call for near-perfection. Dowden’s definition limits the fallacy to cases “where perfection isn’t really required.” Where a single failure is severe (food safety, structural loads, medication doses), holding out for a highly reliable method is reasonable, especially when one exists.
• Waiting is a real option with known costs. If a clearly better option is coming soon and delay costs little, choosing to wait is a comparison, not a rejection of everything imperfect.

The test: compared with what? Does the objection show the option is worse than the realistic alternatives, or only that it isn’t ideal?

A better tool already on hand

A volunteer proposes a shared spreadsheet for booking the community center’s rooms. The coordinator says: “Spreadsheets let two people book the same slot at the same time. The building’s booking system, which we already have access to, blocks double bookings. Let’s use that.”

The coordinator rejects the spreadsheet for a flaw, which sounds like the nirvana fallacy. But the comparison isn’t with an ideal; it’s with a specific system that is already available and doesn’t have the flaw. This is the better alternative is actually available condition, and it’s the kind of comparison Demsetz argued for.

Home canning

A friend shares a method for canning green beans in a water bath and says it “usually works fine.” Pat decides not to use it: “Low-acid vegetables need a pressure canner. ‘Usually’ isn’t good enough when the failure is botulism.”

Pat is rejecting a method for being less than completely reliable, which on its face is a demand for perfection. But the cost of a single failure is severe, and a tested, more reliable method is available. This is the near-perfection is required condition: the standard fits the stakes, and it’s being applied to a real alternative, not an imagined one.

An ideal is easy to picture, and its flaws never show, because it doesn’t exist. The flaws of a real option are concrete and visible, while the flaws of the alternative (often the status quo) aren’t under discussion at all. Judging only what’s in front of you fits what Daniel Kahneman calls “what you see is all there is”.

Whately saw the same asymmetry in debate: objections to a proposed plan are vivid, and unless “the opposite objections be set in the balance on the other side, we can never advance a step.” Since the existing arrangement is what remains when every imperfect proposal is rejected, the fallacy tends to protect whatever is already in place (see Status quo bias). It can also run the other way, as when an existing arrangement is condemned for its flaws and replaced by something whose flaws haven’t been examined, which is the version Demsetz described.

• Ask “compared with what?” Name the realistic alternatives, including doing nothing, and put their flaws next to the ones being raised.
• Separate “not perfect” from “not worth it.” Ask how much of the problem the option solves, and at what cost.
• Check that the alternative is real. If someone proposes replacing an imperfect option, ask how the replacement actually performs, not how it would perform at its best.
• Don’t swing to the other extreme. Rejecting an option can still be right; the imperfection just isn’t enough to show it.