The Art of Reaching the Conclusion You Started With
Motivated reasoning, its six instruments, and the single question that exposes them all.
There are two ways to hold a belief, and they are not variations on a theme; they are opposites wearing the same coat. The first is to arrive at the belief. One begins in ignorance or in doubt, gathers what can be found, weighs it without knowing in advance which way the scale will tip, and then accepts the verdict the evidence delivers — including, and especially, when that verdict is unwelcome. The second is to begin with the belief already in hand, fixed and cherished, and to spend one’s ingenuity afterward constructing the road that appears to lead there. The paving stones are laid backward from the destination, so that the journey looks, to anyone arriving late, like an honest walk. The first is thought. The second is decoration. And the cruelty of the human mind is that, from the inside, the two feel precisely alike.
This essay is about the second, and about the family of small, well-mannered treasons by which it is carried out. The general name for the disease is motivated reasoning. The six instruments by which it operates have names of their own — confirmation bias, cherry-picking, post hoc rationalisation, begging the question, special pleading, and the Texas sharpshooter fallacy — and each deserves to be examined in turn, because each is a separate trick of the hand, and the conjurer is more easily caught once you know how many hands he has. But they are not six diseases. They are six symptoms of one. Beneath every one of them lies a single inversion, the inversion that defines the whole pathology: the conclusion comes first, and the reasoning is built afterward to defend it.
The disease itself
Motivated reasoning means that a person begins with the conclusion they want, and then selects, interprets, or distorts the available reasons so that the conclusion appears justified. Notice the three verbs, because the disease works through all of them, and to understand it you must refuse to let it hide behind only the gentlest one. It selects: it gathers the agreeable facts and quietly declines the rest. It interprets: it takes ambiguous evidence — and most evidence is ambiguous — and reads it, every time, in the direction that flatters the wanted answer. And when selection and interpretation are not enough, it distorts: it bends the meaning of a finding, softens an inconvenient definition, stretches a category until it covers what it was never meant to cover, or misremembers the strength of what was actually shown. The reasoning is not truth-seeking. It is conclusion-preserving. Its purpose is not to discover what is the case but to protect what has already been decided.
The cleanest way to see the structure is to picture a court. In an honest inquiry, the evidence is heard first and the verdict follows from it; the verdict is the output of the process. In motivated reasoning, the verdict is reached first, in private, and the trial is then staged for the look of the thing — the witnesses selected, the testimony arranged, the inconvenient documents lost on the way to the courtroom — so that the foregone judgement may be pronounced with the dignity of due process. The order is reversed. The conclusion is no longer the destination of the argument; it is the argument’s hidden starting point, and everything that follows is reverse-engineering.
What makes this so difficult to detect — and far more difficult to confess — is that motivated reasoning is not the absence of logic. It is the abuse of it. The reasoner is not being stupid; very often he is being brilliant. The machinery of inference is fully intact, every gear meshing, every step formally respectable. Only the direction is reversed: instead of running forward from evidence to conclusion, the apparatus is set running backward from conclusion to a tailored selection of evidence, and then the tailored selection is presented as though it had come first. This is why intelligence is no protection against the disease, and frequently makes it worse. A clever man is simply a better advocate for whatever he already wished to believe. He builds smoother roads to the same false country. The capacity for reasoning that ought to be his defence against error becomes, in the grip of a motive, the engine of his self-deception, because he can always think of one more reason, one more distinction, one more way of reading the data, and he never runs out of them before he runs out of patience.
There is a tidy diagnostic for the asymmetry at the centre of all this, and it is worth carrying in the pocket. When a claim is welcome — when it supports what we already want to be true — we ask of it only, Can I believe this? And the bar for “can” is generously low: a single supporting study, a plausible mechanism, one sympathetic authority, and the welcome claim is waved through. But when a claim is unwelcome, we ask instead, Must I believe this? And the bar for “must” is set cruelly high: now we demand replication, we interrogate the methodology, we hunt for confounds, we note the funding, we recall that correlation is not causation, we discover a sudden and rigorous scepticism that was nowhere in evidence a moment before. The standard of proof is not fixed. It floats — low for the agreeable, high for the disagreeable — and it floats in whichever direction protects the conclusion. The reasoner experiences this floating standard as ordinary good judgement. That is the whole of the danger: the man defending a position he is determined to keep does not feel that he is lying. He feels that he is being reasonable. The feeling of reasonableness is not evidence of reasoning; it is, far too often, merely the sensation of a motive being satisfied.
The motive itself need not be sinister, and usually is not. It is rarely a love of falsehood. More often it is the wish to think well of oneself, or to remain consistent with what one has already said in public, or to keep faith with one’s tribe, or to avoid the cost and humiliation of having been wrong, or simply to spare oneself the discomfort — the genuine, physical discomfort — of holding two contradictory things in mind at once. That discomfort has a name, cognitive dissonance, and it functions as the engine room of the whole apparatus. When a belief we are attached to collides with a fact we cannot easily dismiss, the collision hurts, and the mind reaches, automatically and beneath the level of awareness, for the cheapest available relief. The honest relief — abandoning the belief — is expensive. The dishonest relief — reinterpreting, discounting, or quietly forgetting the fact — is cheap and immediate. The disease is, in the end, a flight from pain, and it is precisely because the flight is so comfortable that it is so rarely recognised as a flight at all.
With the genus described, we can examine the species — the six instruments, the six well-mannered moves by which the conclusion-first mind does its work.
First instrument: confirmation bias
Confirmation bias is the quietest of the six, and the most pervasive, because it requires no deliberate dishonesty whatsoever. It is simply the mind’s habit of noticing what supports the preferred conclusion and failing to notice what contradicts it. It operates on attention and memory before it ever reaches argument. The supporting fact is registered, dwelt upon, filed where it can be found again. The contrary fact is glanced at and lost, or never quite registered at all — not suppressed, merely unattended, which is more dangerous because it leaves no trace and prompts no guilt. One cannot feel dishonest about evidence one never consciously processed.
The asymmetry shows up at every stage of dealing with information. In search, the believer looks in the places likely to confirm him and not in the places likely to refute him; he reads the authors who agree, follows the accounts that nod along, asks the questions whose answers he can predict. In evaluation, supporting evidence is admitted on sight while contrary evidence is subjected to an interrogation it was never designed to survive. In memory, the confirming instances are recalled vividly and the disconfirming instances fade, so that the remembered record of the world is systematically tilted toward whatever the rememberer wanted. A man who believes a certain misfortune always befalls him will remember every occasion it did and forget every occasion it did not, and his memory will then “prove” the belief that distorted the memory in the first place — a closed loop, sealed against correction, mistaken from within for the plain testimony of experience.
The agreeable fact, in short, is received like an invited guest: shown in, made comfortable, believed. The inconvenient fact is met at the door and shown out again before it can take off its coat. And the host, attending only to the guests who were admitted, sincerely reports that everyone he has spoken to agrees with him. He is not lying. He has simply never let the dissenters into the room. Confirmation bias is what makes motivated reasoning feel effortless, because it does the work of distortion silently, upstream of conscious thought, so that by the time the reasoner sits down to “weigh the evidence,” the weighing has already been rigged by what the scale was ever permitted to hold.
Second instrument: cherry-picking
If confirmation bias is the passive failure to notice contrary evidence, cherry-picking is its active, deliberate cousin — confirmation bias that has learned to use its hands. Here the contrary evidence is noticed; it is simply not presented. The cherry-picker has seen the whole tree and chooses, with full knowledge, to show you only the ripe fruit, leaving the blighted fruit on the branches where you will not look. He quotes the one favourable line and lets the qualifying sentence that immediately followed it die in silence. He cites the single study that came out his way and omits the dozen that did not. He reports the three years in which the figures rose and is quiet about the seven in which they fell. The evidence, in his handling, is not consulted; it is harvested.
The cruelty of cherry-picking is that every individual thing the cherry-picker says may be perfectly true. He need not utter a single false statement. The favourable line really does appear in the source. The supporting study really was conducted and really did find what he says it found. The deception lives not in any one of the claims but in the silence around them — in the suppressed denominator, the unmentioned remainder, the context amputated so cleanly that no scar shows. This is what makes it so effective and so hard to confront: you cannot catch the cherry-picker in a lie, because he has told none. You can only catch him in an omission, and omissions are invisible to anyone who does not already know what was left out. To expose him you must possess the whole tree yourself, and most audiences do not, which is exactly the vulnerability he is exploiting.
Selective quotation is its most genteel form. A sentence is lifted from its surroundings, and the surroundings — which qualified it, limited it, or even reversed it — are left behind. A writer who concluded that a thing was possible but improbable is quoted as having said it was possible; the second half, where the improbability lived, is allowed to vanish. The author’s words are used against the author’s meaning. Cherry-picking is the technique by which a man can build a wholly false picture out of nothing but true pieces, and then, when challenged, point to each piece and ask, with wounded innocence, which of them you dispute. None of them. That was never the point. The lie was in the arrangement, and in everything left out of it.
Third instrument: post hoc rationalisation
Post hoc rationalisation is the alibi. Here the order of events is the whole of the matter, and the matter is this: the position came first, and the reasons came afterward. The position was adopted for some cause that has nothing to do with the reasons later offered — appetite, vanity, fear, loyalty, habit, self-interest, the simple wish to have been right all along — and then, once the position was in place, the mind set about manufacturing a respectable justification to stand in front of the real cause and hide it. The reasons given are not the reasons that operated. They are a cover story, composed after the fact, and the more articulate and elaborate the cover story, the more completely it conceals the unargued impulse that actually did the deciding.
The structure is that of a man who acts on a whim and then constructs, retrospectively, an account in which the whim becomes the conclusion of a careful deliberation that never took place. He bought the thing because he wanted it, and then assembled the case for why it was a prudent investment. He disliked the person on sight, and then discovered the principled grounds for his disapproval. He reached the verdict by feeling, and wrote the judgement by reason, and presented the judgement as though the reason had come first. It is the speech composed after the decision has already been made, written so that the decision might pose as the speech’s conclusion rather than its cause.
What makes post hoc rationalisation so resistant to challenge is that the rationaliser frequently believes his own cover story. The retrospective justification is not always a cynical fabrication offered to others; very often it is a sincere fabrication offered to oneself, and believed. The mind genuinely loses track of the order in which things happened. It remembers the reasons as having preceded the conclusion because that is the flattering order, the order in which a rational creature is supposed to operate, and so the memory rearranges itself to match the self-image. Ask such a man why he holds his view and he will give you the rationalisation in perfect good faith, never suspecting that it was assembled to defend a decision already taken on other grounds entirely. He is not deceiving you about his reasons. He has deceived himself first, which is what makes him so convincing.
Fourth instrument: begging the question
Begging the question — petitio principii, the request for the starting point — is the most genteel of the six offences, and the one most likely to slip past an educated audience, because it carries the full outward form of a valid argument. The trick is to place the conclusion, quietly and in disguise, among the premises, and then to announce, with the air of a man arriving somewhere by hard work, that the premises imply the conclusion. Of course they do. The conclusion was sitting inside them from the beginning. The argument assumes precisely what it claims to prove, and the proof is therefore no proof at all but a closed circle, beautifully drawn, that travels a great distance and returns to exactly where it began.
The circularity is rarely as crude as “It is true because it is true.” That form fools no one. The effective version stretches the loop wide enough that the two ends — the premise and the conclusion — are no longer visibly the same statement, only the same statement wearing different clothes. The premise restates the conclusion in other vocabulary, with a synonym here and a paraphrase there, so that the reader, seeing two different sentences, supposes that one has been used to establish the other, when in truth the second is merely the first translated. A claim is “supported” by a premise that could not be believed by anyone who did not already accept the claim. The whole apparatus depends on the audience not noticing that the foundation and the building are made of the same stone.
For the motivated reasoner, begging the question is the perfect instrument, because it allows him to prove his conclusion using a premise that he was only ever willing to grant because it led to that conclusion. He smuggles the destination into the luggage at the start of the journey, travels the appropriate distance to maintain appearances, and then produces it at the end as though it had been waiting for him there — a discovery, when it was only ever a deposit. It is the formal embodiment of conclusion-first thinking: the conclusion is not derived from the premises; the premises are selected because they already contain the conclusion. The reasoning is valid and entirely empty. It can never tell you anything you did not put in, and it was designed by a mind that had no intention of learning anything it did not already wish to hold.
Fifth instrument: special pleading
Special pleading is the double standard wearing a single, untroubled face. It is the application of one rule to the evidence that threatens the preferred conclusion and a different, gentler rule to the evidence that supports it — and the refusal, when this asymmetry is pointed out, to supply any principled reason for the difference. The rule that convicts one’s opponent is suspended at the precise instant it would convict oneself. A standard of evidence that was treated as decisive a moment ago, when it told against the other side, is discovered to be inapplicable, or premature, or subject to a special exception, the moment it begins to tell against one’s own.
The hallmark of special pleading is the unprincipled exception — the “ah, but this case is different” offered without any account of why this case is different that would not also apply to the cases one wishes to keep. Everyone’s anecdotes are dismissed as anecdotes, except one’s own, which are experience. Everyone’s funding is a disqualifying conflict of interest, except the funding behind the studies one likes, which is support for important work. Every prediction that failed for the other theory is a refutation; every prediction that failed for one’s own is a special circumstance, an anomaly, a case where the theory was “not really meant to apply.” The standard is real and rigorous — right up until the moment it points the wrong way, whereupon a custom-built escape hatch appears in the floor, with no explanation of why the hatch should exist here and nowhere else.
What unmasks special pleading is the demand for consistency: state the rule, and then apply it everywhere it reaches. The honest reasoner who claims an exception must say what makes the case exceptional, in terms general enough that the same exception would be granted to an opponent in the same position. If he cannot — if the only thing that distinguishes the exempted case is that it is his — then the exception is not a principle but a privilege, claimed for the conclusion and denied to its rivals. Special pleading is the move by which the motivated reasoner keeps his cherished belief safe from the very standard he cheerfully wields against everyone else’s, and it is detectable precisely because the privilege, once you look for it, has no justification that survives being stated aloud.
Sixth instrument: the Texas sharpshooter
The Texas sharpshooter is the man who fires his rifle at the side of the barn and then walks up and paints the target around wherever the bullets happened to land, so that every shot, by careful subsequent arrangement, falls in the bullseye. The fallacy named after him is the fallacy of drawing the target after the shot — of finding a pattern in data already collected, and then presenting that pattern as though it had been the prediction the data were gathered to test. The pattern is real; the bullet holes are genuinely there. The fraud is in the order: the hypothesis was fitted to the results, and then dressed up as a hypothesis the results confirmed.
The crucial distinction the sharpshooter erases is the distinction between prediction and accommodation. To predict is to specify, in advance, what you expect to find, and then to look — and a theory that survives this is genuinely strengthened, because it staked itself on something that could have gone wrong and did not. To accommodate is to look first, see what is there, and then construct a hypothesis that fits it perfectly — which is no achievement at all, because a hypothesis built to match the data it was built from is guaranteed to match, and has been tested by nothing. The sharpshooter takes an accommodation and sells it as a prediction. He searches a large field of possibilities until he finds some configuration that looks striking — a run of coincidences, a cluster, a correspondence between two lists — and then he draws his circle tightly around that configuration and declares that the odds of such a thing arising by chance are vanishingly small, conveniently forgetting that he had the whole field to search and stopped only when he found something that looked good.
This is the engine of every spurious correlation paraded as proof, every prophecy that turns out to have been written after the event it foretold, every “pattern” that emerges only because someone went looking through enough noise to find a shape they liked. Give a man a large enough body of data and the freedom to define his target afterward, and he will always find a bullseye, because with the target undrawn, everything is a near miss waiting to be circled. For the motivated reasoner the appeal is obvious: he can take the raw mess of what actually happened, comb through it until he locates the fragment that supports the conclusion he wanted, draw his target snugly around that fragment, and present the result as though the universe had obligingly confirmed a prediction he never actually made. The shot came first. The target was painted to fit it. And the marksmanship is announced as though the order had been the other way around.
One disease, six costumes
Set the six instruments side by side and the family resemblance is unmistakable, because they are not six independent errors that happen to keep company. They are one error in six costumes. Confirmation bias rigs the intake of evidence; cherry-picking rigs its presentation; post hoc rationalisation rigs the story of how the belief was formed; begging the question rigs the structure of the argument; special pleading rigs the standard of judgement; and the Texas sharpshooter rigs the direction of fit between hypothesis and data. Six different stages of inquiry, six different points of attack — and at every one of them, the same hand at work, pulling in the same direction, toward the same predetermined end. They are the symptoms; motivated reasoning is the disease. Treat them as separate fallacies and you will play whack-a-mole forever, refuting one only to watch the same impulse surface as another. Recognise them as one thing, and you can name the thing itself.
In practice they rarely appear alone; a single motivated episode usually deploys several at once, in concert, each covering for the others. Consider the most ordinary case, the one that describes nearly everyone, nearly all the time:
Someone has a position they want to protect, then fits the evidence to get the answer they already wanted.
That single sentence is the whole disease in miniature, and it is, in the first instance, simply motivated reasoning — the conclusion fixed first, the evidence fitted afterward. But watch how the fitting is done, and you will see the instruments arrive. It is expressed, most commonly, through confirmation bias — the man notices and remembers everything that supports his protected position and overlooks everything that threatens it — and through cherry-picking — when he comes to make his case, he presents the supporting fragments and omits the rest. Confirmation bias selects the evidence inside his head; cherry-picking selects what comes out of his mouth. Press him on the gaps and special pleading appears, exempting his evidence from the scrutiny he applies to yours. Ask him how he arrived at the view and post hoc rationalisation supplies a noble origin story for a conclusion adopted on far less noble grounds. Examine his argument closely and begging the question is often holding it up, his premises quietly assuming the very thing in dispute. And if he points to data, the Texas sharpshooter may well have been there before you, circling the convenient fragment and calling it a prediction. One impulse; six hands; a single direction. That is what the disease looks like when it is actually running, and it is running, somewhere, in all of us, more or less constantly.
The clean formulation
If the whole of it must be compressed into a single statement — the kind one can hold in mind as a test, and turn upon one’s own thinking as readily as upon anyone else’s — then this is it:
Motivated reasoning is the practice in which the conclusion is fixed first, and the argument is reverse-engineered to protect it. The evidence is not being used to discover the answer; it is being filtered to manufacture one.
Read that twice, because the two halves are doing different work and both are essential. The first half names the inversion: conclusion first, argument afterward, the natural order of inquiry turned exactly backward. The second half names the function of the evidence, which is the part most people miss, because they imagine the disease consists in ignoring evidence, when in fact it more often consists in using it. The motivated reasoner does not refuse to look at evidence. He is frequently buried in it, fluent in it, able to cite it by the yard. But the evidence is not performing the function evidence is supposed to perform. It is not an input that determines the conclusion; it is a filter, applied after the conclusion is already fixed, whose job is to let the convenient facts through and strain the inconvenient ones out. Evidence used to discover an answer and evidence used to manufacture one look identical from a distance — the same citations, the same studies, the same confident marshalling of fact. The difference is invisible in the output and lives entirely in the order of operations: whether the conclusion was the destination the evidence led to, or the starting point the evidence was recruited to defend.
Why it is invisible from the inside
The final difficulty — and it is the one that makes the whole subject more than an academic catalogue — is that motivated reasoning is almost perfectly invisible to the person committing it, and almost perfectly obvious in everyone else. We are all of us superb detectors of the bias in our opponents and nearly blind to the identical bias in ourselves. When they reach a convenient conclusion, we see the motive behind it at a glance; when we reach a convenient conclusion, it presents itself to us as the plain and obvious truth that any reasonable person would accept. The asymmetry is not hypocrisy, exactly. It is structural. We have access to the feeling of our own reasoning — which feels, from the inside, exactly like honest inquiry, because the machinery really is running and the conclusions really do seem to follow — and we have access only to the results of theirs, which we are free to judge by their suspicious convenience. We grade ourselves on intention and everyone else on outcome, and we are graded by everyone else on outcome in return.
And intelligence, to repeat the most uncomfortable point, is no shield. It is closer to an accelerant. The capacity that ought to defend a mind against error — the ability to reason, to argue, to find considerations and distinctions and counter-examples — is exactly the capacity that, once a motive has set the target, is conscripted to defend the error instead. The more powerful the intellect, the more formidable the advocate it makes, and an advocate is precisely what the motivated reasoner has turned himself into: a man whose reasoning serves a client, and the client is the conclusion he started with. The cleverest people are not the least susceptible to this. They are the most dangerous when susceptible, because they can build a defence so elaborate, so erudite, so genuinely impressive, that neither their audience nor they themselves can find the seam where honest inquiry ended and motivated construction began.
The cure, and the question that administers it
There is no clever technique that defeats motivated reasoning, because cleverness is the disease’s preferred instrument, and any technique built from cleverness can be turned to the disease’s service. The only cure is a disposition, and it is an uncomfortable one: the genuine willingness to lose. A belief one is not willing to surrender to contrary evidence is not, whatever it may feel like, a conclusion. It is a possession — held the way one holds property, defended the way one defends property, and reasoned about the way a man reasons about a thing he has already decided to keep. And a man who cannot be reasoned out of a position was never reasoned into it; he arrived by some other route and merely furnished the place with arguments afterward.
The disposition can be made operational, and reduced, in the end, to a single question — the question that administers the cure, and the one diagnostic that exposes every instrument at once. Put it to any belief you hold, and especially to the ones you hold most dear, the ones that feel least like opinions and most like simple facts about the world:
What would change my mind?
Demand a real answer, and a specific one. Name the evidence — concrete, observable, possible — that would oblige you to give the belief up. If you can name it, and if you would genuinely follow it should it ever arrive, then you are reasoning: your belief is the output of inquiry, exposed to the world, prepared to lose. But if the honest answer is nothing — if no evidence, however clear, however direct, could in principle move you, if every conceivable disconfirmation has already been pre-equipped with an exemption, a reinterpretation, a special circumstance, a reason to look away — then you are not holding a conclusion at all, and no amount of argument you can produce in its favour changes that fact. You are not reasoning. You are protecting. You have a position you want to keep, and you are fitting the evidence to keep it, and all the brilliance you bring to the task is merely the brilliance of a man decorating a destination he never had any intention of leaving.
The evidence is not there to win. It is there to be heard, and to be obeyed when it speaks against you. Anything less — however learned, however fluent, however sincerely felt — is not thought. It is decoration. And the first and hardest discipline of an honest mind is to know, of itself, which of the two it is doing.