I think your post is basically on the right track, but it needs a bit of refinement.

Our minimal assumptions are that the world is consistent (so logic is useful) and at least partially structured (induction works, some detectable laws apply). These assumptions appear to underlie everything we know, from the proofs of mathematics, to the fact that rocks roll downhill. Why should long division always give us the same result? This, too, is induction.

Thus, it seems that there is no knowledge that is not predictive in some way. It follows that knowledge is justified by evidence of its corresponding prediction. That’s where causality comes in. Induction assumes that there are detectable, causal connections between events.

Apply this criteria to our knowledge of the meaning of propositions. What you’ll find is that propositions that are free of prediction have infinite uncertainty in their meaning.

There is a correlation between milk consumption and low cancer rates. But there is no inductive law here, only many levels of secondary correlations (higher milk consumption in the 1st world, which has better health care and therefore lower cancer rates)

Collecting data to create laws about the universe is called induction. Without induction I cannot really collect data even, because for all I know what I just observed ceased to be the case the minute I stopped seeing it. Without induction the only thing I can know is what I am experiencing this second. Furthermore, Occam’s Razor and induction are strongly tied to each other; Occam’s Razor cannot be proved through deductive logic alone, as there is nothing contradictary about not believing it.

It would be contradictory for mass to not attract mass as it would violate the symmetry of space and time. One could construct a framework where this is not the case, but you would begin to involve a large number of hypothesis that would force an Occum’s Razor choice between this new framework and the old one. This is primarly why new phenmenon is so exicting as it leads to new frameworks!

Carl said what I was going to say. Your response to Carl is: “For example I am certain about the applicability of Newton’s laws, even into the future, but I don’t have knowledge of their future correctness.”

I’m not sure what this means. My point is that all knowledge is either assumed to be true (unfalsifiable), true by definition (deductive logic, unfalsifiable), or true given the truth of induction (induction is an unfalsifiable assumption). So by your definition I would say nothing is knowledge.

Abyss:

How are Newtonian laws deduced? If there is no induction necessary, and only deduction, they must follow from the law of non-contradiction in some way. How would it be contradictary for mass not to attract mass? Doesn’t a mathemetical framework require real-world data to become meaningful? How can we determine whether something is a fundamental symetry in the universe without making observations, and using induction to formalize those observations?

On top of that, I would say that if you really really want to doubt the certainty of mathematical logic, there’s nothing anyone can do to stop that person (though I suggest that it’s probably better not to).

Carl said what I was going to say.

Your response to Carl is:

I’m not sure what this means. My point is that all knowledge is either assumed to be true (unfalsifiable), true by definition (deductive logic, unfalsifiable), or true given the truth of induction (induction is an unfalsifiable assumption). So by your definition I would say nothing is knowledge.

Abyss:

How are Newtonian laws deduced? If there is no induction necessary, and only deduction, they must follow from the law of non-contradiction in some way. How would it be contradictary for mass not to attract mass? Doesn’t a mathemetical framework require real-world data to become meaningful? How can we determine whether something is a fundamental symetry in the universe without making observations, and using induction to formalize those observations?