On the Success of Scientific Theories

[JuliaL]

3) If past experience (potentially non-existent if history is bunk) is consistent with a proposition, that proposition is true.

I'll go out on a limb and choose 3) today.
Science is successful because science has been successful.

I think that is the Texas Sharpshooter Fallacy if you aren't careful.
https://en.wikipedia.org/wiki/Texas_shar...er_fallacy

What if victory conditions morph along with science? We probe a little farther and that gives direction for further probing. If we knew right now where it all was going to end up, wouldn't we be done?
I try to enjoy the journey and not worry too much about the destination.

Maybe you need to say that the theory has been successful - not true , because it successfully predicted new experimental results in the past? Then you apply inductive reasoning to say that this theory will probably be successful in the future too?

This is what I was trying to say is the limit to our knowledge of truth.

---

Perhaps this points to a deeper problem in that the binary values true and false tend to gut anything we can say about actual theories because accuracy of predictions (demonstration) is a gradated property, not a binary one. Perhaps that points to the No Miracles argument being phrased in terms of "more successful theories.... than less successful theories," and then the litmus of success is demonstrability as Ben suggests.

That sounds like a more practical approach IMO.

This too is in line with my private definition of truth: A proposition is true in direct proportion to which it serves to accurately predict future events. I only use true/false as a binary value in boolean operations. There are confidence intervals everywhere, including in parameters used for calculation in nominally binary chips & neuron depolarization potentials.

[Alex K]

This too is in line with my private definition of truth: A proposition is true in direct proportion to which it serves to accurately predict future events.

So in your philosophy, there cannot possibly be true statements about singular events in the past?

I only use true/false as a binary value in boolean operations. There are confidence intervals everywhere, including in parameters used for calculation in nominally binary chips & neuron depolarization potentials.

I strongly agree

[FallentoReason]

I notice that you haven't defined what it means for a theory to be false.

The *definition* of a theory should give you the answer

A scientific theory is one that can be falsified, amongst other things.

[Pizza]

Falsifiable in principle or in practice?

[FallentoReason]

Falsifiable in principle or in practice?

I'd say in principle, where the hypothetical situation involves an observable entity that can be (i) seen not just a one-off event and (ii) that doesn't depend on your expectations i.e. it's an entity that is robust/objective etc.