[polynox]

I trust that when I drive over a bridge it won't collapse.

But I cannot say I know that the bridge won't collapse.

Knowledge is a true and justified belief. But trust is not a justification - quite the opposite. If you had a justification, you wouldn't need trust.

Trust whomever you want, on whatever basis you want, but that's fundamentally not reason or rationality, it's just an argument from authority [1]. No statement becomes more true just because someone said it, regardless of who they are.

I'm not saying it's easy or even possible to do in practice. Science is insanely complex. But let's not confuse trust for rationality, not least of which because it brings to the fore the very real need for institutions that are worthy of our trust.

[1] https://en.wikipedia.org/wiki/Argument_from_authority

[skissane]

> Trust whomever you want, on whatever basis you want, but that's fundamentally not reason or rationality, it's just an argument from authority [1]. No statement becomes more true just because someone said it, regardless of who they are.

A statement can become more likely to be true just because someone said it, depending on the nature of the statement and the person making it. For example, my brother is an oncologist. If I had some health concern about cancer, I would certainly be asking for his opinion. His opinions on that topic are not guaranteed to be correct, but as an oncologist, they are significantly more likely to be correct than that of the average person.

One problem with the frequent popular invocation of claimed "fallacies", such as "argument from authority", is many of them are only strictly speaking fallacious when used as purely deductive arguments, but real world human reasoning isn't purely deductive, it involves a great deal of induction and abduction as well. There is nothing inherently wrong with an "argument from authority" as an inductive or abductive argument.

> Knowledge is a true and justified belief. But trust is not a justification - quite the opposite. If you had a justification, you wouldn't need trust.

One can estimate the conditional probability P(proposition X is true|agent A says that X is true) based on observations of what kinds of things agent A (or other significantly similar agents) has said in the past, and how many of them turned out to be true or false. You can then use that conditional probability estimate to inform your own decision as to how much credence to give to proposition X. That is both trust and justification.

[autoexec]

> One problem with the frequent popular invocation of claimed "fallacies", such as "argument from authority", is many of them are only strictly speaking fallacious when used as purely deductive arguments,

There's a fallacy for that: https://en.wikipedia.org/wiki/Argument_from_fallacy

[polynox]

I was not aware of abduction before today, thank you!

I think we may agree in practice if not in theory. I did have the Hume-an problem of induction explicitly on my mind. It would be quite fair to say my epistemology is quite conservative.

[jasonhansel]

Isn't your citing Wikipedia, in itself, an appeal to authority? Why should we accept Wikipedia as an authority on which arguments are fallacious and which are rational?

(Also, that wiki page mentions that many authors don't see it as a fallacy.)

[polynox]

I'm just citing it for the context of the argument both for and against appeal to authority being fallacious and irrational, not that Wikipedia is the final arbiter of what's rational or not. It's more to say that my position is not novel than anything else. Indeed, to be consistent I would need to say that I don't care how many authors do or don't see it as a fallacy :)

[thatcat]

You could do some theoretical calculations about bridges yourself, inspect for cracks, and prove from first principles that the bridge is in fact safe; it's not some complicated biological process.