Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

We do this all the time. This is basic bayesian analysis. We make predictions on the probabilistic increase/decrease in events. When the evens occur at the predicted rates, we can attribute them, in aggregate, to the underlying theoretical causal factors.

If say, climate scientists were predicting a 5% decrease on rainfall over a 10 year period, and the rain completely stopped over that time, we could rightly say the hypothesis was bunk. However, to suggest that probabilistic causality is incompatible with scientific claims, I would say you need to need to re-read your philosophy of science.

Yes, you should always read a "very probably" with any claim of fact for any empirical claim (this is the problem of induction), but yes, you can say that "this weather event is (very probably) the result of climate change," if it fits in nicely with the probabilistic predictive model.



Bayesian analysis does not support absolute claims of cause for single events.

If chance of drought increases 10%, then you can say this drought is 10% likely to be caused by climate change.

If you observe 110 events and expected 100, you can say that this likely supports your theory.

At no point can you say "this drought right here was caused by climate change".


Again, you are getting extremely pedantic about the nature of causality. At that level of rigidity, empiricism becomes solipsism.

What causality means is framework dependent.

If we are making accurate predictions within a consistent framework, it is perfectly reasonable to use a sense of causality within that framework.


You are ignoring the unreasonableness of making definitive claims, independent of relative influence.

If the chance of drought is increased 1%, can you still claim that every one of them is caused by climate change?

At what point do you think someone becomes a liar or misleading? Is there any threshold in your mind?


> You are ignoring the unreasonableness of making definitive claims, independent of relative influence.

I'm making the important point that "definitive claims" are only definitive within the the framework in which they are made.

>If the chance of drought is increased 1%, can you still claim that every one of them is caused by climate change?

Again, we can and we can't. We have a seemingly reliable framework that would suggest that every one of them is cause by climate change. There ought to be builtin error bars, but if the probabilities are falling under the predictions, it's reasonable to say there is a causal link. It's also reasonable to have concerns about noise. The concept of definiteness ultimately just breaks down for all empirical claims. That's not to say that we don't use the term all the time, it's just to say that the term "definitive" is effectively a heuristic for "very probably" in the frame we are using.

>At what point do you think someone becomes a liar or misleading? Is there any threshold in your mind?

Again, I have a lot of sympathy about this point. There is a dance between what may we believe and what must we believe. The claim in the frame of we may believe that this drought was caused by climate change is reasonable. The argument against the claim that we must believe that the drought was caused by climate change is generally problematic and generally unreasonable. We can know in the sense that it's falls out from the prediction, but we can not know in the sense that there are error bars and noise and there will be probabilistic events that fall outside of our model's level of precision.

Right now, predictions about climate change made by institutions with academic reputations have shown to be very accurate. If we have predictions that are shown to be completely inaccurate or involve a nonsensical framework, then yes, i would say "lying or misleading" is reasonable. This is especially true for post-hoc justifications for these inaccurate predictive results. This happens all the time in the world of finance, but it rarely happens in the world of science because most of the time scientists are actually seeking the best model of understanding. There was a bit of push back against Einstein, for example, but generally speaking, the scientific community goes with the frameworks that best explain the phenomenon.




Consider applying for YC's Winter 2026 batch! Applications are open till Nov 10

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: