Gap Junctions, Postage Stamps
Kairos.
vendredi 21 mars 2025
vendredi 15 novembre 2024
dimanche 29 septembre 2024
Today I want to talk about consilience versus rumors.
For example, compare two claims - both of which might be true or false.
1) Haitian immigrants have been eating their neighbors' dogs and cats in Springfield, Ohio.
2) A Big Bang occurred at the start of the universe.
Both of these claims may be maintained as beliefs by the frequency of bumping into the claim. That is, multiple sources seem to make the claim, and these sources might be independent. When we hear something repeatedly, especially from different sources, we often, perhaps gradually, begin to vibe with the belief.
I do not know for a fact, personally, whether either of these happened.
So how can we proceed beyond admitting what we don't know for sure?
First, ask yourself why you believe - or don't believe - each of these two claims. If you can, put a number to each one, a percentage probability.
Here are my spur-of-the-moment numbers:
1: 25%
2: 67%
I don't claim any correctness. It's just how I'm feeling at the moment.
Ok, so we have numbers, but we haven't really justified them, so they might as well be random digits.
I'm sure some of you think my number for 1 is high, and my number for 2 is low. I know you're thinking that.
The point is I have no first-hand evidence about 1, so I have to contend with multiple professional fact-checkers versus the former president and an actually reasonably articulate mob on Twitter who say they have seen "dozens of videos" that "prove" claim 1. Many were initially skeptical but became convinced. I do not want to dive through such videos (some are gruesome). If it matters either way, I'd rather wait for the wisdom of crowds and the wisdom of experts to reach a consensus. This is a case where I really do not want to do the research myself. It's more about curiosity than about proving a point; I don't think the validity of claim 1 makes a significant difference, but there may always be some value in accurate verification or falsification.
Generally, I would consider someone describing all fact-checkers as in cahoots to be an addle-minded paranoiac. But amid claims that everyone has become too polarized (which I agree about) and that leftists have our heads in the sand about immigration and the truth of what Trump says (well, on the immigration, that might apply to me), I want to be cautious.
Simply ridiculing the idea that Haitians are eating cats and dogs provides zero evidence. People who are trying to dismiss the claim on the basis that it sounds absurd - sounds just like the usual, hateful tripe out of Trump, only a bit worse this time - are not making solid evidence-based statements. They're mainly persuading others and themselves.
I do not find it implausible that Haitian immigrants have eaten pets in this country, given that this is done in Haiti. And if it could well have happened somewhere in this country, it could have happened in Springfield, Ohio.
Now, I've looked at the original post that got JD Vance's attention and then Trump's and then, regrettably, everyone else's. That post was traced back to a woman who admitted she'd heard a rumor and was speculating, and has since felt bad and made an apology. Still, to be thorough, just because it's no more verifiable than a rumor doesn't mean it *didn't* happen. It just means that we essentially have to acquit for lack of further evidence.
There are multiple other leads with varying degrees of believability, including verified incidents that involved pet eating (by a white person) but not Haitians. We must admit that if a white person can eat a pet, a Haitian might have. I can't rule that out, however wacky it sounds or hatefully motivated the debate moment was.
The fact checks tend to amount to "there have been no credible claims," meaning that nothing has been proven - we can't prosecute the case further, in a metaphorical sense - any reports there may have been, we have to acquit based on lack of solid evidence.
Ok. Phew. We're getting to my point now.
For many, the *number* of *seemingly plausible* stories or leads will amount to *definite proof*. Even though each one may, on inspection, fail to amount to solid evidence, taken as a constellation of data points, they will strongly suggest that something's going on. "Where there's smoke, there's fire," many people like to say.
This does resemble how science works.
However, to really appreciate how science works, we need to tease the details apart a little more.
Many scientific theories have essentially been "proven" by what's called "independent, converging lines of evidence." A simpler way to say that is "consilience."
When people encounter many reports of cat-snacking on Twitter and feel that therefore claim 1 must be correct, they probably feel that they are responding to "independent, converging lines of evidence." Multiple stories from different people who don't know each other (independent lines) are consistent with the same overall picture (converging evidence).
For this extremely effective method to work, though, we need to be extremely careful about what we call "independent" and what we call "evidence."
For example, if most of the stories are coming up by way of Trump supporters, they may come from different people, but those people do happen to share a motivation, and these days, someone might even have deep-faked the video.
Similarly, if we watch video evidence that someone was eating a pet, how sure are we that it is occurring in the time and location asserted? If someone in the video looks Haitian... do we know they are?
If we watch 100 videos of people who seem Haitian appearing to eat pets, any of our assumptions could be off about any or all of them.
For any piece of evidence that cannot be confirmed, it is a mistake to assume that it brings a fractional, cumulative amount of truth. It might bring none at all. Each one might bring none. Evidence might not be accumulating while rumors and reports and "proofs" accumulate.
If after consideration we deem that each of 100 videos is 1/20 likely to be real, this does not mean that, after watching all of them, we are (100)(1/20)=5, or in other words completely certain 5 times over, that claim 1 is true. Not how it works!
When each story is no more fully confirmable than a rumor - and, yes, even video evidence can match this description, given that we do not know what part of space and time, if any, the video depicts - then we cannot say that the many rumors add up to proof. Many rumors can easily add up to nothing, no matter how many there are.
I understand this mistake in thinking, because I know that thinking in a technical way is tricky and error-prone. But it's better if we're all aware of that.
Again, I don't actually know about the truth of claim 1, and my 25% guesstimate is just an easy number to toss out there. It includes (for example) the possibility that just one Haitian immigrant ate just one pet in Springfield Ohio at some point this century, it was never reported, and the rumors we are hearing are separate and false, but this thing still happened in the last 24 years (I'm making up that time limit to stay specific). It might have! Just saying. (Arguably if it happened just once, "They're eating the cats and the dogs..." is false, as it implies many incidents, but we'll let that go and be as generous as possible to the claim.)
As for the Big Bag, sometimes laypeople - and lately, increasingly scientists, I think - will observe that we don't have any definite proof that the Big Bang happened, even though we have many pieces of evidence that range from hinting to strongly suggestive.
The evidence for the Big Bang takes many forms (or so I'm led to believe, having seen a few). These arguments can allegedly be made independently of each other (and they at least seem independent to me as well).
It's important to remember that science does not actually prove any claims directly. All it can do is vigorously fail to disprove claims.
We cannot prove that no (0) pets were eaten by Haitians in Springfield. Proving a negative is famously nigh impossible. The best we can do here, along the lines of proving nothing happened, is vigorously look into it, as well as we can, yet despite our earnest and thorough efforts continue failing to disprove that nothing happened (double negative, eep). That's the closest we'll ever get to proof: we looked really hard and didn't find any fire behind the supposed smoke.
The same goes for science and the Big Bang. We probably can't indisputably prove it happened. But we can vigorously fail to disprove it for a long while, via many approaches, until we're kind of exhausted and we shrug and accept that the Big Bang does seem to have happened.
If the claim of pet eating had been accurate, we might have expected any of these:
- The person who made the meme gave details about when and where the pet was eaten, and by whom.
- JD Vance defended the claim as definitely true.
- Springfield police and other city officials such as the mayor acknowledged it.
- Fact checkers reported it happened or the jury was still out.
- Reliable news agencies reported on the story as a fact.
- The Ohio governor, who is from Springfield, said it did happen or might have happened.
These are likely results of a real incident along these lines that draws attention.
Yet none of these things are happening (though I can't vouch for all news stations with Fox News at large).
We look for the smoking gun behind the smoke, and we keep coming up blank.
Now, maybe some of the reports and video evidence is real.
Once again, I insist I do not ultimately know. But I can talk a bit about how evidence combines.
Multiple independent, converging lines of evidence - if this is taken very seriously - provide the closest science gets to proof.
But what we humans often do informally when we hear multiple reports is not that. When we aren't super careful, our belief can amount to nothing more than magical thinking.
Multiple doubtful pieces of evidence do not automatically add up to credibility. Yet our minds often want to believe they do, for reasons that have to do with how we evolved and the very real power of what's called Bayesian statistics, the statistics of uncertain contingencies based on networks of evidence.
This is the best explanation I can give at the moment of an extremely important aspect of thinking about the "facts" you're given.
Remember: rumors versus consilience. They're not the same. What you're trying to find is actual consilience. That means different, unconnected sources using different methods - as many and as different as possible - pointing, perhaps uncertainly but with thorough credibility as to what they do have, to the same conclusion.
vendredi 20 septembre 2024
(It's something I do not usually find difficult, but it has taken practice.)
Sometimes we need to intervene and save someone from a stupid idea.
Sometimes we need to get in the way of a behavior so it doesn't happen, or can't continue.
I would like this more precise, more effective mindset to spread.