Twice now I have made you a promise and then stopped short of keeping it. I have brought you up to a question and then said: not yet, hold it, we will come back. The first night I opened the deeper question under the whole deepfake panic — how much of what is in your head you ever took in through your own eyes — and I said we would come back and do it properly, with your own beliefs set out in front of you, and that the count would surprise you. Then I did not do it. Last night I gave you Thomas Reid and set him against Hume: for Reid, belief comes first and is withdrawn only for cause; for Hume, belief has to be earned by evidence, and the default is no. I ended by handing you one small question to take away — how many of the things I know did I actually check for myself — and I told you, again, not to answer it yet.
Tonight you answer it.
I am finished surveying. We have two major philosophers, and they give two completely different descriptions of you, and I told you last time that no third philosopher is going to arrive and settle the disagreement. So we will not wait for one. We are going to do the only thing left that could settle which of the two descriptions is true of you: stop talking about the human knower in general and examine the one knower you can inspect from the inside. You. Your own beliefs.
So tonight is not a lecture. Tonight we take an inventory. I am going to ask you to take the things you know — the important ones, the ones your life rests on — and look, one at a time, at where each of them actually came from. And this is the only fair way to settle it. You cannot settle it by choosing the theory you prefer. Of course you prefer the careful one; it is flattering; it makes you the responsible adult. But the question is not which description you would rather were true of you. The question is which one is true of you. And the only way to find out is to count.
This is Philosophy for Us — philosophy for everyone, no degree required. This is the fourth of eight, on how you know anything you did not see for yourself, in a world where the recordings have started to lie. And tonight, for the first time, I am going to stop describing the problem and ask you to find it in your own beliefs. Have a belief ready — one you are sure of. We start with the easy ones.
Let me set the rules, because an inventory is only worth anything if it is fair, and I want this one to be fair to the careful checker in you — the one already preparing an objection.
Here is where we are. Two nights ago I gave you Hume at full strength: belief should be earned, piece by piece, by evidence; the default is no; trust is something you extend only once the evidence is in. Last night I gave you the other major figure, Reid, at equal strength: the default is yes; belief comes first and is shut off only for cause; the trusting child is not something we grow out of but the thing that got us to a working mind at all. And I told you the two do not combine — that one of them is the foundation and the other is a story resting on top — and that I would not tell you which. I still will not. What I am going to do is put the question to the only evidence that can settle it, which is the record of your own life.
The test is this, and I will state it as precisely as I can, because everything tonight depends on getting it exactly right. Name one belief. Not any belief — a real one, an important one, something that actually carries weight in your life, something you would be genuinely shaken to find was false. And the one condition is this: it has to be something you know entirely for yourself, with no one's word in it. No testimony. Not something you read, not something you were taught, not something you absorbed from the people around you, not something you saw on a screen that someone else put there. Something you, personally, with your own senses and your own reasoning, established and could establish again, from the beginning, with no other person's report anywhere in the chain. One. That is all I am asking for.
Before you answer, let me set aside two easy answers, because they are the cheap ones and I do not want to spend the hour on them. The first is the bare sensation. I know I am in pain right now. I know I am seeing red. Yes — those you have firsthand; no one told you your headache hurts. But they are not load-bearing in the sense we mean. No one's headache is the foundation of their picture of the world. "I am in some pain" is not a fact about the world that you act on, argue about, build a life on; it is a private sensation. I am not asking for the sensations. I am asking for the large, standing facts about the world that your life leans on.
The second easy answer is the one the careful checker reaches for first, and it is the one to watch, because it will come back all night in more sophisticated forms. It goes: but I do verify things for myself — I saw it. I watched the footage. I read the study. I looked it up. Keep that one — we take it apart later. For now just notice its shape. You saw the footage — on a screen, which someone else recorded, and edited, and sent. You read the study — which someone else ran, and wrote up, in a language of statistics you are trusting them to have used honestly. You looked it up — and got someone else's words, ranked for you by a machine someone else built. Every one of those is testimony in the form of firsthand sight. I am not saying they are worthless. I am saying they are not what the test asked for. The test asked for something with no one's word in it. So keep looking.
I am going to give you a moment to do this — not as a trick, but because the whole episode fails if you do not actually try. Pick your belief. The most solid, most important thing you are sure you know. Then we are going to take it and trace it back to where it came from. Have you got one? Hold on to it. Let us begin.
Start with the easiest one — the one you would have laughed at me for asking about an hour ago. Your birthday. The date you were born. You know it the way you know your own name; it is as certain as anything you know; it is barely a belief at all, just a fact about you.
So take it and ask where it came from. You were there, of course — no one is more present at a birth than the person being born — but you remember none of it, and you know that, so that is not where the knowledge lives. The knowledge is a date, a specific day and month and year, and that date reached you in exactly one way: someone told you. Your mother said it. It was written on a certificate, by a clerk, who copied it from a form, filled in by a tired person in a hospital at the time, or near the time, or from someone's memory of the time. Every one of you has simply taken their word for the single most intimate fact there is about you — the day you began — and you have never verified it — could not. There is no experiment you can run on yourself this evening that returns your birthday. The source of that certain-seeming fact is the same as all the rest: someone told you, and they were told. Your birthday is testimony. It always was.
And if your birthday is, the same holds, and faster. Who your parents are. You presumably believe the people who raised you are biologically yours — or you believe, just as firmly, that they are not, that you were adopted, that there is a story there. Either way: how do you know? You did not run the test. You took their word. Millions of people are walking around right now perfectly certain of a parentage that a cheek swab would overturn, and the only reason you are confident you are not one of them is that no one has given you a reason to doubt — which is not the same as having checked. The most basic fact about where you come from, who your people are, your biological origin: testimony. You took it as a child and never went back to it.
Now go outward, away from yourself, to the large, solid facts about the world — surely those are firmer. The Earth has a molten core. You know that; everyone knows it; it is in every schoolbook. Have you been? Has anyone you have ever met been? No one has. No one has ever seen the core of the Earth and come back to report it; the deepest hole human beings have drilled does not get a fraction of the way through the crust. Everything you believe about the inside of this planet you are standing on, you believe on the word of people who also never went, who inferred it from instruments you have never operated, reading signals you could not interpret, and wrote it down, and it passed to a teacher, who passed it to you. You are as certain of the molten core as you are of your own hand, and you have had as much firsthand contact with it as you have had with the far side of the moon. None.
Pick a country you have never been to. Madagascar. You believe Madagascar exists — you would stake your house on it; you would think someone was unwell to doubt it. On what? You have never been. You have seen it on maps other people drew, in photographs other people took, in the reports of travellers and the statements of atlases. If everyone who claimed Madagascar exists were somehow in on a hoax, you personally would have nothing to catch them with. Your certainty about an entire island nation of thirty million people rests, from end to end, on trust. Now think of a war — one happening right now, somewhere you are not, that you have opinions about, strong ones, opinions you would argue over dinner and lose a friend over. Have you seen it? You have seen footage. Selected, framed, sent. You have read reports, by people you are trusting, relaying people they are trusting. Your entire moral relationship to a war — your sense of who is wronged and who is doing the wrong — is built entirely out of secondhand report, and you have been treating it the whole time as if you knew it firsthand.
I could continue all night and it would never stop being true, and that is the point: it never stops being true. The money in your bank account, which is a number on a screen maintained by an institution you trust to be telling you the truth. That the medicine in the bottle is the medicine on the label, which you have swallowed, on trust, into your own body, more times than you can count. That smoking causes cancer — did you run the study? That the food you ate today was not poison — you trusted a chain of strangers a thousand people long. Every one of them has the same source: someone told you, and they were told.
So now I am going to make you do the part I cannot do for you, because if I only list mine it stays my problem, not yours. Take the belief you got ready at the start. The most important, most certain thing you brought to this. Maybe it is something about your own life, your own history. Maybe it is something large about the world — a scientific fact, a historical one, a fact about a place or a people. Hold it up. And do to it what I just did to the birthday and the core and the island: trace it back, step by step, and find the point where you, personally, with no one's word in it, established it for yourself. Not where you could have. Where you did. Take your time. I will wait. … And what you are finding, if you are honest, if you did not reach for a bare sensation, is that there is no such point. The trail runs back and back and never reaches ground that is your own. It is testimony at every step.
Now the careful one in you is not going to accept this, and good — the inventory is not fair unless the checker gets to make his defence. He has two strong replies, the ones you are reaching for right now, and we will see whether they hold.
The first is the most natural thing to say: all right, I have not actually checked these — but I could. I could verify my birthday; I could find the certificate. I could have the test done and confirm my parentage. The molten core, the island, the war — the evidence is out there in principle, and I could go and get it. I have not bothered, because life is short, but the point stands: these are things I am capable of verifying for myself, and that is what makes me a checker. The checking is available to me; I am simply choosing not to spend the time on it.
That reply gives a great deal away even as it tries to save you. Could is not did. The self-image we are testing was that you are someone who knows things by checking them — and now, pressed, the most you can say is that you are someone who knows things he has not checked but believes he theoretically could have. That is a very different thing. That is not a checker. That is a person who has approved the entire account without reading it and is now insisting he would have caught any errors if he had looked — which is what every person who has ever been defrauded also believed. And it is worse than that, because the "could" is mostly empty too. Could you really verify the molten core? You cannot drill there; no one can. What you could do is go and read more people — better ones, more careful ones, but more people, more testimony, more words from more strangers. You cannot reach the core. You can only reach a more reputable account of the core. "I could check it myself" turns out, almost everywhere it matters, to mean "I could find a source I trust more." Which is not checking. It is choosing better testimony. Mark that; there is a more sophisticated version of it, and it is the question we take up next time.
The second reply is the deeper one, and it is the one the deepfake panic is really built on, so it matters most. It goes: but some of it I did see. With my own eyes. I saw the footage of the war. I watched it happen. That is not someone's word — that is my own perception, the one thing that is mine. And here is the serious consequence the whole argument has been leading to. You did not see the war. You saw a screen. You saw a lit rectangle in your hand, showing you an image that a camera somewhere captured, that a person aimed, that an editor cut, that a platform chose to put in front of you rather than the thousand others it could have. Your own eyes, your one firsthand instrument, were pointed the whole time not at the world but at other people's reports of the world. The seeing was real. What you were seeing was testimony. And this is where the deepfakes have their real effect, because for as long as there have been photographs and film, that screen was a reasonably reliable stand-in for the world — reliable enough that you could forget it was a report at all and feel as though you were seeing the thing itself. The forgery does not damage your eyes. It removes the last case in which you could still tell yourself you were looking at the world directly, when you were only ever looking at the record someone handed you.
So put it together, and look hard at what it is, because we have reached something worth not hurrying past. You came in holding a picture of yourself. The picture was: I am a checker. I grant belief after I verify; trust is the cheap thing I fall back on when I am too lazy or too rushed to do the real work. That picture is the standpoint — the position you were viewing everything from without realising you had taken a position at all. And we did not argue against it. We tried to act on it — to live by it, for one episode, to find a single load-bearing belief the picture said should be everywhere — and there was none. You did this once before, at a lower level: back in Mind and Machine, you caught yourself treating your own mind as the measure of every mind, the inside you happened to have as the standard every other inside had to meet, and you never noticed you were doing it until you tried to defend it. Here it is again, one level up. You have been treating your own eyes as the measure of all knowledge — real knowing is what I see for myself — and it turns out you have relied the whole time on what other people saw, and were told, and passed along.
I am not going to soften what that is. Discovering it is disorienting, and it should be. It is what it feels like to discover that the beliefs you built your life on rest on no verification of your own. You are not the checker. You never were. There is no record you keep, no verification you run, that grounds the great mass of what you know — there is a large inherited structure of things people told you, resting on things they were told, and you in the middle of it, having taken almost all of it on someone's word, including the things you hold most strongly. That is what the inventory found. Not that some of your beliefs are secondhand. That almost all of them are, and that the picture of yourself as the one who checks was a story you were telling, not something you were doing.
And here is the thing I am most careful not to tell you, because it is the whole matter and I will not give it away tonight. I am not going to tell you whether this is a disaster or a relief. Hume could stand here, once the checker self-image has collapsed, and say: yes, exactly — this is the charge against you: you have believed thousands of things you had no right to believe, carelessly, your whole life, and the deepfakes are merely the point at which the problem becomes impossible to ignore; the response is not to feel better about it, but to begin, at last, doing the checking you always owed. And Reid could stand in the same place and say the opposite: of course you found no firsthand foundation — there was never going to be one — and that is not your failure, it is the human condition, the default condition you were always in, and the only mistake was ever believing the checker story in the first place. Both replies remain open. The inventory did not choose between them. It did something narrower and stranger, and I think more important: it took away the one thing you were sure of before either of them spoke — that you were the kind of creature the checker story described. You are not. Now the disagreement between Hume and Reid is no longer about some general human knower. It is about you, and the strange position you turn out to be in.
What tonight did was this, and then I am going to leave you with it rather than resolve it for you.
We stopped surveying. We had two descriptions of the human knower and no way to choose between them from the outside, so we turned inward and ran the one test that could decide it on the single subject you have access to — and the test did not settle it in Hume's favour or Reid's. It did something to you first, before either of them gets a turn. It removed the picture of yourself as a checker. You came in certain you were a person who knows things by verifying them and falls back on trust only when he is lazy. You cannot say that now. You have looked, and the checking is not there — not under your birthday, not under your own parentage, not under the molten core or the island or the war you would argue about, not even under the footage you watched with your own eyes. It was testimony at every step, and you, the checker, were a character in a story you had been telling, not something you have ever actually been.
You want me to make that all right — to turn round now and say: but this is fine, this is good even, trust is a wonderful thing, we are all in it together, relax. I am not going to. Not because it is false — I genuinely do not yet know whether it is false — but because saying it tonight would be asserting a conclusion I have not earned, and I am not going to start now. Maybe losing the checker is a liberation. Maybe it is a catastrophe. Reid would call it the first, Hume the second, and neither of them has won. I cannot settle that for you tonight, because I cannot settle it at all yet, and pretending I can is exactly the comfort I told you at the start I would not sell.
So do not resolve the disorientation too quickly. It is the honest place to be tonight. You are someone whose knowledge rests almost entirely on other people's word, and you have just found out that this is so.
But I will not leave you with nothing to do, because there is one reply left, and you reached for it tonight, and I let you, deliberately, so we would have it ready. When you ran out of beliefs to name, you said: all right, I cannot check every claim myself, I grant that — but I do not have to. I can do something more intelligent. I can check my sources. I can work out which people, which institutions, which reports are reliable, test them once, and then trust what they tell me — and that way the trust is not blind, it is earned, it rests on reliability I have confirmed, and I get to go on being a careful person after all, one level up. That is the checker's last defence. It is the most reasonable thing anyone says in this whole argument. It is what every sensible person believes they are doing. And next time we are going to examine it as we examined everything else tonight — not taking my word that it works, and not taking my word that it fails, but actually looking, and finding out whether there is anything solid underneath it. I am not going to tell you tonight which way that comes out. I will only say that it is the last defence, the one the careful person is most confident of, and that it has not yet had the hard look the others have just had. We take it up next time.
For now, take one thing with you, the way you took the last one. The next time you hear yourself think it — I could check that if I wanted to; I take no one's word — stop, and ask the question the inventory taught you. Not could I. Have I? And when the answer comes back, as it almost always will, no — then you are not, in that moment, the careful checker you pictured. You are doing the thing you have done your whole life and never looked at. You are trusting. The only question left, the one I am not answering yet, is whether that is the disaster Hume says it is or the ground Reid says it is — and you cannot answer it honestly until we have examined the last defence and found whether anything holds it up. So keep the question open. Do not let anyone close it for you, including me.
Thanks for listening. I'll see you next time.