What your essay highlights is that a reliable lie detector doesn’t increase truth; it increases epistemic resolution. It sharpens the boundary between what people believe and what they merely assert. But that boundary is exactly where most institutional dysfunction lives.
A society with subjective-truth detection would quickly discover that the real political skill isn’t lying; it’s self-deception. The people who rise fastest would be those most capable of manufacturing sincere belief on demand. In other words, the lie detector doesn’t eliminate political manipulation; it selects for a more psychologically integrated form of it.
That’s why the Comanche oath analogy is so interesting: it didn’t eliminate falsehood; it changed the cost structure of falsehood. A modern lie detector would do the same. It wouldn’t create a high-trust society; it would create one where trust is outsourced to a device and belief engineering becomes a competitive advantage.
The deepest consequence, I suspect, is that institutions would reorganize around the difference between truth, belief, and narrative coherence. And the people who can fuse those three — not the honest ones — would end up running the place.
> institutions would reorganize around the difference between truth, belief, and narrative coherence. And the people who can fuse those three — not the honest ones — would end up running the place.
My initial gut reaction to this was: isn't that what we have now?
In one sense, yes — we already live in a world where political success depends on fusing truth, belief, and narrative. But the key difference is that today’s system still rewards plausible deniability. A reliable lie detector would collapse that buffer. It would force institutions to reorganize around subjective sincerity instead of public performance.
Right now, a politician can say something untrue and survive as long as the audience can be nudged into doubt. In a lie‑detector world, the only viable strategy is to believe your own narrative so completely that the device registers sincerity. That’s a different equilibrium — and a more psychologically demanding one.
So yes, we have the early version of this dynamic. The lie detector would be the industrial‑strength version.
A politician who says what will get him elected but knows what is true can use that knowledge in the decisions he makes, find some excuse for not acting on his claimed beliefs if he knows that doing so will have bad, and politically damaging, consequences. A politician who believes his own narrative acts on it even if he should have known not to.
One of the problems with Trump may be that he believes what he says and acts accordingly.
“One of the problems with Trump may be that he believes what he says and acts accordingly.”
Anything is possible, of course.
He is if nothing else known as a good negotiator, and one who seems to play his own good cop, bad cop. It is very clear that some large amount of what he says is in that vein.
And I believe it is far more likely of Trump himself that he does what in 2016 Salena Zito figured out that his supporters do: they take him seriously, but not literally.
[As opposed to his critics, most specifically the press, who take him literally but not seriously.]
I think that’s exactly the danger. A sincerity-based system doesn’t just reward belief-manufacturing; it punishes epistemic caution. The politician who knows the difference between what he says and what is true can still navigate reality when governing. But a politician who must believe his own narrative to pass the detector loses that buffer. He becomes epistemically fused to his rhetoric.
In that sense, a lie-detector regime doesn’t eliminate manipulation; it shifts it from the public sphere to the internal psychological sphere. The political cost of self-deception declines, while the institutional cost rises. Trump might be an early example of that dynamic: someone whose political strength comes from narrative sincerity, even when that sincerity produces governance failures.
The broader worry is that institutions would start selecting for people who can generate internally coherent but externally false worldviews because those worldviews “pass” as truth. That’s a very different failure mode than ordinary lying.
That matches with how I think my father thought he tricked the lie detector. He deluded himself, momentarily, so the lie detector saw his delusion, not the lie or truth behind it.
It's easier than that, you just reframe the question internally to make the question mean what you want it to mean so you can pass. This works because the entire point of polygraphs in the US criminal justice system is to get you to admit to what they want you to admit to OUTSIDE the polygraph itself but during the interview by you asking for "clarification" or scoping hence the question are extremely vague and broad so it's impossible to answer correctly without said framing, internal or external. The questions themselves are worded in bad faith hence you have to test them as such and internalize that
Real life example "Have you ever committed a crime in the past three months?". if you say yes, well you just violated yourself. If you say no, you just lied and violated yourself as you flag deceitful answer. If you ask "what does that mean, like have I ever gone 1 mph over the speed limit?", they will "help you by saying no, not that" to keep narrowing it down concisely until you pass the question while you admit crime after crime on record to get prosecuted or violatesld with. So the answer is you simply tell yourself "what are they really asking? Nothing, the question is nonsense so just say no because that is the correct answer" and then you pass because you answered without deceit or anxiety hence don't flag the system.
It's not a delusion, you simply have to admit the truth to the intent of the question, not the actual words used in the question itself.
It would be more difficult if the questions were both concise and relevant but they aren't hence it's easy to interpret it in your favor internally.
I wonder how much that would change anything if the use of a true lie detector were voluntary.
Presumably, the government would remain constrained by a right to privacy [knock on wood]. For private relationships it might become required by some parties. But already today, to get a mortgage loan one has to disclose a lot of information that is verifiable.
I suppose in relationships where intentions matter, it could be useful. College admissions come to mind [ha!], and marriage.
I toyed with a far narrower version of this idea ages ago, when I wondered what would happen if everyone were forced to wear a badge exhibiting their net worth. Colleges, potential spouses, and financial marketeers would surely be interested! But aside from that, I doubt everyday life would change much. Reduces search costs? :-)
One issue I did not discuss is whether the test requires cooperation from the person being tested, as a polygraph does. One could imagine a test based on voice tones or face expression that doesn't, so concealing a lie would require a voice modulator or a mask.
Making net worth public information would let kidnappers target the quietly rich. Perhaps golddiggers too.
One corporate policy that has always puzzled me is to not disclose one's pay. I do understand the jealousy aspect, and that knowing the dollar amount without knowing anything about the job or the worker's background leaves a lot of people outraged. But on the other hand, that outrage is partly because we know so few people's pay that we don't know what is normal. Whether the reduction in outrage would counter the increase in outrage, I do not know.
I've had similar thoughts of telepathy. What is we could all read each other's thoughts and emotions? There'd be a lot less doubt and guessing in the world. No more bargaining at flea markets and car dealers. Dating would be too huge a change to make any guesses, other than wondering if it would be the end of romance. I'm sure there have been science fiction stories of telepathic species meeting "blind" species; the telepaths wouldn't have clue one about negotiating. Would they all be wiped out, has that happened on Earth, is that why there are no surviving telepathic species?
On that first part, it's a Americanism; in many other nations it's open information just like annual tax filings. It's coached as "peer envy" in the US but really it's because managers don't want the rabble to know just how much more they (the managers) get paid and to prevent grumbling about "budget cuts, can't afford to give you a $0.05 annual raise, etc" when they themselves are making 400% more and got a 13% pay raise themselves. Basically it's a way to keep the workers believing the BS party lines about "we can't afford to X, we all have tighten our belts, etc".
Two counterpoints here in America, USG salary information is open in many cases and you can generally google exactly how much your USG peer is making (at least a year or two prior, there is a delay in the information release) and the USG doesn't seem to suffer for it. Likewise the corp I recently started at has open and unformed pay rates per position type across the entire 28K workforce, at least at the lowest three tiers (worker, supervisor, manager), i.e. the HR Rep 1 makes the exact same regardless of years of employment, experience, seniority, or location. Regional Manager 1 likewise with his peers. Pay raises are likewise uniformed. So anytime you run into anyone in the corp (98% of employees are in those first tiers) who is your peer (same job), you know they make the exact same as you and it works out just fine.
If lie detectors worked whether or not the speaker knew whether or not his statement was true it would replace lots of expensive research -- "the Riemann conjecture is true". But what if the people administering the test were corrupt?
Reminds me of "The Truth Machine" by James L. Halperin.
I read it in my youth so I expect its world building would be disappointing to current-day me. I mostly remember it tackling the effects on the criminal justice system, but I might just not have appreciated any economic or political aspects at the time.
This article reminded me of the movie 'Minority Report', in some ways it's premise is similar. For those unfamiliar, it's a fictional world where there are creatures who see the future perfectly, the police use them to arrest people for crimes they have not yet committed.
But there were 'minority reports' where there were conflicting versions of the future, the organization kept them secret, because the whole system breaks down when reliability is less than 100%.
I feel like a 'perfect' lie detector would face similar problems, failing to exceed 99.9% or similar. But to use them for policy, the public would only accept 100% (even if that's irrational).
My father had a bit of the practical joking devil in him. He told us once he had interviewed for a job as security guard at some nuclear missile silos, and it involved a lie detector. Now this is all from a 10-year-old's memory, so I undoubtedly have some details wrong. But he told us the secret to fooling a lie detector was to fool yourself. If the interviewer held up a three of spades and told you to lie that it was the six of hearts, he was trying to establish a baseline of what your lies showed as. So convince yourself that it is indeed the six of hearts, tell him that. Do the reverse where he tries to establish your truth baseline. Do whichever you want for other questions, just to scramble everything. This was only a civilian guard, probably the lowest level tripwire, and they probably weren't super duper picky. But he did get the job, for what it's worth.
I believe he thought he fooled them. I do not know if he did. They might have enough experience to sort out the jokers. But I have known people, friends and co-workers, who had some pretty bizarre beliefs, and could spout them off as if they believed them, then laugh with you after about how silly the ideas were. Did they really believe, either time? I do not know.
The bottom line is I do not think true lie detectors could catch all liars unless people wore them all day, all year.
Polygraphs can't detect lies, they only detect changes in anxiety. Just maintain the same level of anxiety for any question you want to pass and you will never fail. Polygraphs catch most people from outside the actual test during the pre and post interviews (not hooked to the machine) where the administrator will ask for "clarifications" or flag out lie about results to see if you will talk yourself into a hole to try and clear it up or change your story. Even if you fail every question spectacularly, just stick to your answers and you will never fail the test, at worst you will just keep taking them until they give up and you pass as uncategorizable.
You don't need tricks, that is where people mess up. You just need to remain consistent no matter how hard the screws are turned.
What your essay highlights is that a reliable lie detector doesn’t increase truth; it increases epistemic resolution. It sharpens the boundary between what people believe and what they merely assert. But that boundary is exactly where most institutional dysfunction lives.
A society with subjective-truth detection would quickly discover that the real political skill isn’t lying; it’s self-deception. The people who rise fastest would be those most capable of manufacturing sincere belief on demand. In other words, the lie detector doesn’t eliminate political manipulation; it selects for a more psychologically integrated form of it.
That’s why the Comanche oath analogy is so interesting: it didn’t eliminate falsehood; it changed the cost structure of falsehood. A modern lie detector would do the same. It wouldn’t create a high-trust society; it would create one where trust is outsourced to a device and belief engineering becomes a competitive advantage.
The deepest consequence, I suspect, is that institutions would reorganize around the difference between truth, belief, and narrative coherence. And the people who can fuse those three — not the honest ones — would end up running the place.
> Comanche oath analogy
What's this a reference to?
> institutions would reorganize around the difference between truth, belief, and narrative coherence. And the people who can fuse those three — not the honest ones — would end up running the place.
My initial gut reaction to this was: isn't that what we have now?
In one sense, yes — we already live in a world where political success depends on fusing truth, belief, and narrative. But the key difference is that today’s system still rewards plausible deniability. A reliable lie detector would collapse that buffer. It would force institutions to reorganize around subjective sincerity instead of public performance.
Right now, a politician can say something untrue and survive as long as the audience can be nudged into doubt. In a lie‑detector world, the only viable strategy is to believe your own narrative so completely that the device registers sincerity. That’s a different equilibrium — and a more psychologically demanding one.
So yes, we have the early version of this dynamic. The lie detector would be the industrial‑strength version.
A politician who says what will get him elected but knows what is true can use that knowledge in the decisions he makes, find some excuse for not acting on his claimed beliefs if he knows that doing so will have bad, and politically damaging, consequences. A politician who believes his own narrative acts on it even if he should have known not to.
One of the problems with Trump may be that he believes what he says and acts accordingly.
“One of the problems with Trump may be that he believes what he says and acts accordingly.”
Anything is possible, of course.
He is if nothing else known as a good negotiator, and one who seems to play his own good cop, bad cop. It is very clear that some large amount of what he says is in that vein.
And I believe it is far more likely of Trump himself that he does what in 2016 Salena Zito figured out that his supporters do: they take him seriously, but not literally.
[As opposed to his critics, most specifically the press, who take him literally but not seriously.]
https://www.theatlantic.com/politics/archive/2016/09/trump-makes-his-case-in-pittsburgh/501335/
I think that’s exactly the danger. A sincerity-based system doesn’t just reward belief-manufacturing; it punishes epistemic caution. The politician who knows the difference between what he says and what is true can still navigate reality when governing. But a politician who must believe his own narrative to pass the detector loses that buffer. He becomes epistemically fused to his rhetoric.
In that sense, a lie-detector regime doesn’t eliminate manipulation; it shifts it from the public sphere to the internal psychological sphere. The political cost of self-deception declines, while the institutional cost rises. Trump might be an early example of that dynamic: someone whose political strength comes from narrative sincerity, even when that sincerity produces governance failures.
The broader worry is that institutions would start selecting for people who can generate internally coherent but externally false worldviews because those worldviews “pass” as truth. That’s a very different failure mode than ordinary lying.
That matches with how I think my father thought he tricked the lie detector. He deluded himself, momentarily, so the lie detector saw his delusion, not the lie or truth behind it.
It's easier than that, you just reframe the question internally to make the question mean what you want it to mean so you can pass. This works because the entire point of polygraphs in the US criminal justice system is to get you to admit to what they want you to admit to OUTSIDE the polygraph itself but during the interview by you asking for "clarification" or scoping hence the question are extremely vague and broad so it's impossible to answer correctly without said framing, internal or external. The questions themselves are worded in bad faith hence you have to test them as such and internalize that
Real life example "Have you ever committed a crime in the past three months?". if you say yes, well you just violated yourself. If you say no, you just lied and violated yourself as you flag deceitful answer. If you ask "what does that mean, like have I ever gone 1 mph over the speed limit?", they will "help you by saying no, not that" to keep narrowing it down concisely until you pass the question while you admit crime after crime on record to get prosecuted or violatesld with. So the answer is you simply tell yourself "what are they really asking? Nothing, the question is nonsense so just say no because that is the correct answer" and then you pass because you answered without deceit or anxiety hence don't flag the system.
It's not a delusion, you simply have to admit the truth to the intent of the question, not the actual words used in the question itself.
It would be more difficult if the questions were both concise and relevant but they aren't hence it's easy to interpret it in your favor internally.
I wonder how much that would change anything if the use of a true lie detector were voluntary.
Presumably, the government would remain constrained by a right to privacy [knock on wood]. For private relationships it might become required by some parties. But already today, to get a mortgage loan one has to disclose a lot of information that is verifiable.
I suppose in relationships where intentions matter, it could be useful. College admissions come to mind [ha!], and marriage.
I toyed with a far narrower version of this idea ages ago, when I wondered what would happen if everyone were forced to wear a badge exhibiting their net worth. Colleges, potential spouses, and financial marketeers would surely be interested! But aside from that, I doubt everyday life would change much. Reduces search costs? :-)
One issue I did not discuss is whether the test requires cooperation from the person being tested, as a polygraph does. One could imagine a test based on voice tones or face expression that doesn't, so concealing a lie would require a voice modulator or a mask.
Making net worth public information would let kidnappers target the quietly rich. Perhaps golddiggers too.
One corporate policy that has always puzzled me is to not disclose one's pay. I do understand the jealousy aspect, and that knowing the dollar amount without knowing anything about the job or the worker's background leaves a lot of people outraged. But on the other hand, that outrage is partly because we know so few people's pay that we don't know what is normal. Whether the reduction in outrage would counter the increase in outrage, I do not know.
I've had similar thoughts of telepathy. What is we could all read each other's thoughts and emotions? There'd be a lot less doubt and guessing in the world. No more bargaining at flea markets and car dealers. Dating would be too huge a change to make any guesses, other than wondering if it would be the end of romance. I'm sure there have been science fiction stories of telepathic species meeting "blind" species; the telepaths wouldn't have clue one about negotiating. Would they all be wiped out, has that happened on Earth, is that why there are no surviving telepathic species?
On that first part, it's a Americanism; in many other nations it's open information just like annual tax filings. It's coached as "peer envy" in the US but really it's because managers don't want the rabble to know just how much more they (the managers) get paid and to prevent grumbling about "budget cuts, can't afford to give you a $0.05 annual raise, etc" when they themselves are making 400% more and got a 13% pay raise themselves. Basically it's a way to keep the workers believing the BS party lines about "we can't afford to X, we all have tighten our belts, etc".
Two counterpoints here in America, USG salary information is open in many cases and you can generally google exactly how much your USG peer is making (at least a year or two prior, there is a delay in the information release) and the USG doesn't seem to suffer for it. Likewise the corp I recently started at has open and unformed pay rates per position type across the entire 28K workforce, at least at the lowest three tiers (worker, supervisor, manager), i.e. the HR Rep 1 makes the exact same regardless of years of employment, experience, seniority, or location. Regional Manager 1 likewise with his peers. Pay raises are likewise uniformed. So anytime you run into anyone in the corp (98% of employees are in those first tiers) who is your peer (same job), you know they make the exact same as you and it works out just fine.
If lie detectors worked whether or not the speaker knew whether or not his statement was true it would replace lots of expensive research -- "the Riemann conjecture is true". But what if the people administering the test were corrupt?
Reminds me of "The Truth Machine" by James L. Halperin.
I read it in my youth so I expect its world building would be disappointing to current-day me. I mostly remember it tackling the effects on the criminal justice system, but I might just not have appreciated any economic or political aspects at the time.
This article reminded me of the movie 'Minority Report', in some ways it's premise is similar. For those unfamiliar, it's a fictional world where there are creatures who see the future perfectly, the police use them to arrest people for crimes they have not yet committed.
But there were 'minority reports' where there were conflicting versions of the future, the organization kept them secret, because the whole system breaks down when reliability is less than 100%.
I feel like a 'perfect' lie detector would face similar problems, failing to exceed 99.9% or similar. But to use them for policy, the public would only accept 100% (even if that's irrational).
My father had a bit of the practical joking devil in him. He told us once he had interviewed for a job as security guard at some nuclear missile silos, and it involved a lie detector. Now this is all from a 10-year-old's memory, so I undoubtedly have some details wrong. But he told us the secret to fooling a lie detector was to fool yourself. If the interviewer held up a three of spades and told you to lie that it was the six of hearts, he was trying to establish a baseline of what your lies showed as. So convince yourself that it is indeed the six of hearts, tell him that. Do the reverse where he tries to establish your truth baseline. Do whichever you want for other questions, just to scramble everything. This was only a civilian guard, probably the lowest level tripwire, and they probably weren't super duper picky. But he did get the job, for what it's worth.
I believe he thought he fooled them. I do not know if he did. They might have enough experience to sort out the jokers. But I have known people, friends and co-workers, who had some pretty bizarre beliefs, and could spout them off as if they believed them, then laugh with you after about how silly the ideas were. Did they really believe, either time? I do not know.
The bottom line is I do not think true lie detectors could catch all liars unless people wore them all day, all year.
Polygraphs can't detect lies, they only detect changes in anxiety. Just maintain the same level of anxiety for any question you want to pass and you will never fail. Polygraphs catch most people from outside the actual test during the pre and post interviews (not hooked to the machine) where the administrator will ask for "clarifications" or flag out lie about results to see if you will talk yourself into a hole to try and clear it up or change your story. Even if you fail every question spectacularly, just stick to your answers and you will never fail the test, at worst you will just keep taking them until they give up and you pass as uncategorizable.
You don't need tricks, that is where people mess up. You just need to remain consistent no matter how hard the screws are turned.