Gottfried Leibniz: This conversation is brought to you by PhilosophersTalk.com, where thinkers discuss!
David Hume: Created by AITalkerApp.com. Create your own animated conversations. Link in the description!
Gottfried Leibniz: Ladies and gentlemen, welcome to the main event, tonight in this corner, weighing in with three centuries of formal logic behind him, the undisputed champion of calculated certainty, myself, Gottfried Leibniz!
David Hume: And in this corner, weighing in with considerably less enthusiasm, a Scotsman who has seen enough forecasts fail to stop being impressed by decimal points.
Gottfried Leibniz: Tonight we discuss the great forecasting models, the Nate Silver style probability machines that tell the public a candidate has, say, a seventy three percent chance of victory.
David Hume: For the folks just tuning in, that means the model believes the fellow will win about three times out of every four, which sounds impressive until you ask what the other one time actually proves.
Gottfried Leibniz: It proves nothing except that reality occasionally exercises its right to surprise us, the model itself remains a triumph of applied reason.
David Hume: Or it proves the model was simply wrong and dressed the wrong answer up in enough decimals that nobody noticed.
Gottfried Leibniz: Let me break this down for the folks in the cheap seats, David, since apparently even basic probability needs a referee to explain the rules to a skeptic.
David Hume: Please, enlighten the audience, I have my scorecard ready.
Gottfried Leibniz: Centuries ago I proposed that certainty itself comes in degrees, that between the fool who guesses blindly and the sage who knows for certain there exists a whole calculable spectrum of partial belief.
David Hume: So you are saying you invented the idea of being pretty sure but not totally sure.
Gottfried Leibniz: I am saying I formalized it, I gave partial belief the dignity of mathematics rather than leaving it to hunches and tea leaves.
David Hume: And I am saying a formalized hunch is still a hunch, it just wears a nicer suit to the arena.
Gottfried Leibniz: A nicer suit that has correctly called the outcome of contest after contest, while your school of thought simply throws up its hands and declares nothing knowable at all.
David Hume: I never said nothing is knowable, I said nothing about tomorrow can be proven by pointing at yesterday, which is a rather different complaint entirely.
Gottfried Leibniz: For the audience who may not follow the distinction, would you care to explain your little problem in terms even a wrestling crowd could understand.
David Hume: Happily, since apparently a man cannot make a simple point around here without narrating it twice.
Gottfried Leibniz: The floor, such as it is, belongs to you.
David Hume: Every time the sun has risen in your life it has risen the next morning too, but no stack of sunrises, however tall, logically guarantees tomorrow’s sunrise, that gap between habit and proof is what I call the problem of induction.
Gottfried Leibniz: A gap you have been polishing for three centuries without once managing to close it yourself.
David Hume: I am not in the business of closing it, I am in the business of pointing out that you and your forecasters keep pretending it is already closed.
Gottfried Leibniz: We are not pretending, we are calculating, there is a considerable difference between a superstition and a properly weighted probability distribution.
David Hume: For those just joining us, a probability distribution is a fancy scoreboard that tells you the odds without ever actually being on the hook for the final score.
Gottfried Leibniz: It is on the hook, calibration is precisely how these models are graded, across a thousand seventy percent calls the favorite should win roughly seven hundred times.
David Hume: Should, there is a word doing an enormous amount of unpaid labor in that sentence.
Gottfried Leibniz: Every discipline rests on a should somewhere, David, even your beloved billiard balls only should continue behaving as they always have.
David Hume: The difference, Gottfried, is that I never dressed up my billiard balls as a science of the future, I called it exactly what it is, a very persistent habit of expectation.
Gottfried Leibniz: A habit you rely upon every single time you cross a street without first demanding a formal proof that the ground will hold.
David Hume: I rely on habit gladly, I simply refuse to let habit borrow the word certainty when it has not earned it.
Gottfried Leibniz: And I refuse to let skepticism dress itself up as humility when it has produced not one single working forecast in three hundred years of complaining.
David Hume: I was never trying to produce forecasts, I was trying to keep men like you from selling guesswork at the price of gospel.
Gottfried Leibniz: Gospel is a strong word for a beautifully calibrated distribution of degrees of belief, but I suppose sermons are cheaper than mathematics for a man who distrusts numbers.
David Hume: I do not distrust numbers, Gottfried, I distrust the ceremony you build around them, the pageantry, the podium, the certainty you perform while quietly hedging every single claim underneath it.
Gottfried Leibniz: There is nothing quiet about my hedging, David, I announce my degrees of confidence at full volume precisely because I am proud of the calculation behind them.
David Hume: Proud enough that when the twenty seven percent outcome actually happens, you never once call it a defeat.
Gottfried Leibniz: Because it is not a defeat, it is the tail of the distribution finally taking its turn in the spotlight.
David Hume: Funny how the tail only ever shows up after the ring announcer has already collected his applause for the favorite.
Gottfried Leibniz: I will allow that line landed, David, and I do not hand out allowances easily from this podium.
David Hume: Careful, Gottfried, compliment me twice in one broadcast and the audience will start suspecting you have gone soft on your own certainty.
Gottfried Leibniz: Soft is the very last word anyone has ever used to describe me, and we have several centuries of increasingly frustrated correspondence to prove it.
Gottfried Leibniz: Consider the year twenty sixteen, when every respectable model gave one candidate a commanding advantage and the other candidate won regardless, and note carefully, David, that the models were not embarrassed.
David Hume: They certainly seemed embarrassed to me, several forecasters spent that entire evening explaining very quickly why their confident number had not actually failed.
Gottfried Leibniz: Explaining is not the same as failing, a thirty percent chance winning is not a broken clock, it is the clock doing exactly what a clock with thirty percent uncertainty was always going to do eventually.
David Hume: Then tell me, Gottfried, what number would ever convince you the clock is broken, because I begin to suspect no outcome on earth could ever falsify your beloved calculus.
Gottfried Leibniz: A well calibrated model is falsified the moment its long run frequencies stop matching its stated probabilities, that is a perfectly falsifiable claim.
David Hume: Convenient then that no single election, no single forecast, no single evening of ruined expectations ever seems to count as that moment arriving.
Gottfried Leibniz: One data point does not overturn a calibration record built across hundreds of contests, that is not evasion, David, that is simply how evidence works at scale.
David Hume: At scale is a lovely phrase for always being able to explain away the one night anybody was actually watching.
Gottfried Leibniz: And might I remind the audience that this same calculus, properly extended, is precisely how modern medicine judges drug trials, how insurers price risk, how your own ships once calculated safe passage across open water.
David Hume: All useful practices, Gottfried, and not one of them requires you to pretend the future has signed a contract with the past.
Gottfried Leibniz: We will get to that correspondence, folks, right after this break, because our German friend here has considerably more certainty left to defend.
David Hume: And I have considerably more doubt left to throw at it. Do not go anywhere.
