Fallacies for induction?

What is a fallacy anyway?

A fallacy is a heuristic that's known to be bad. Heuristic being simply what we call methods for making judgments.

So, if we assume someone is lying because they're from Sweden, that's a bad heuristic. Swedes can be correct about things (this will shock the Danes out there, no doubt).

Since philosophy worked under a mistaken notion of deductive primacy for about two thousand years, we've got long lists of heuristics that are bad when doing deduction. But given that what we actually rely the most on is induction, this can lead to some confusion.

So, apart from the unfortunate tendency of people to think the problem of induction is a problem for induction (it's actually a problem for deduction, as deduction is completely dependent on induction if it wants to speak of reality), there's another overlooked aspect to the differences between these two forms of reason:

No deductive fallacy can be assumed to be fallacious for induction.

For some, it's simply because evidence != proof. Some examples are argument from ignorance and appeal to probability both of which are prohibitive for deduction, but crucial heuristics for induction - we could not function at all if we did not assume the probable is probable or that things for which no evidence exists are less likely to be true than things for which evidence does exist.

For others it's that the fundamental process of forming an inductive argument is different from the process of forming an deductive argument.

Let's say you observe a number of events (including event x) leading you to conclude the rule y for events like x, is it now fallacious to apply y to x, as x was part of the basis for concluding y?
I don't think so.

So, for each fallacy, it needs to be established independently whether it's relevant for induction:

Distrusting all evidence provided by Swedes is clearly not called for, but it's clearly not a good idea to place great trust in evidence gained through torture or from people who have a vested interest in a given conclusion being reached.

However, induction is a great deal more complex than deduction, so clear-cut rules are hard to come by, and one might establish inductively how to weight various sources of evidence for maximum accuracy : In fact, this is what many sciences do as a matter of course.

Inductive Neo-Kantian Deflationism - A fuzzy robustness.

According to Blackburn, deflationism is the view that there's nothing in general to say about what it means to be true in general.(1)

This means that what Blackburn calls Ramsey's Ladder (which goes from "p", to "it is true that p", to "it is a fact that it is true that p" ... all the way to "it corresponds to the eternal normative order governing the universe that p") is horizontal. Each rung says just the same as the bottom run, namely "p".(2)

In the same text Blackburn says that the correspondence theory of truth is trivially acceptable, namely "p is true because it corresponds with the facts", but this too is a horizontal claim "p corresponds with the facts" also just means "p".

Now, that's something I will agree with in the street man sense of "facts", but what of my own, neo-Kantian sense of "fact"? My stance is that we don't have any facts, because I relegate "fact" to the inaccessible objective reality, and we are all subjects here.(3)

So I am not going to accept that being true means corresponding to facts (although I definitely agree that saying "'p corresponds to the facts' only says 'p'".

In my neo-Kantian epistemology (4) I will not allow any reference to anything mind-independent in the epistemology, hence truth cannot be based on facts.

So, what is truth, if not a correspondence to facts?

My answer is that "truth", the very concept, is an inductive conclusion from the general tendency of the universe towards consistency.

The general tendency of the universe towards consistency also happens to be why induction works so well (no problem, seriously).

So how does that unpack in inductive terms?
I say it goes something like this: we see that some beliefs seem to correspond well with how the world seems to behave. (Seem is my indicator of an inductive conclusion, so the formation of truth would be "we seem to keep seeing some beliefs seeming to correspond well with how the world seems to behave").

This means that if we want to explain to ourselves why we think that p, we should look to the evidence, to the apparent reliability of the belief p. Because evidence is experience, and experience is nicely mind-dependent, and hence it is allowed in my epistemology.

So what does this mean with regards to the really deep questions (the 5Y of reality, if you wish)?
It means, as always when dealing with induction, that at some point we should expect to come up with "it just is". (5)

And that's kind a ground-breaking, or rather ground-establishing when it comes to truth.

[This is the most fleshed-out bit of mind-lint I have so far produced - there may be problems with it, there may be unanswered questions and gotchas, but by golly, looking for them will be fun]

(1) Truth, edited by Simon Blackburn and Keith Simmons, Oxford readings in philosophy, Oxford University Press, 1999, page 3
(2) ibid, p.6
(3) if you are an object, you may continue reading, but be aware that the text will be puzzling.
(4) which Kant would no doubt be rather upset about, so perhaps it's not a good name, but here we are...
(5) As I read it, this is where Yudkowski suggests recursive justification hits rock bottom  http://lesswrong.com/lw/s0/where_recursive_justification_hits_bottom/

Three worlds for the NeoKantians, under the sky:

In which I attempt to clean up some hopelessly tangled terms.

We humans have three distinct worlds that we track beliefs about.

World 0) There is (we must assume) the objective world, which we can't access, but which we (must) presume forms the causal basis for our sensory input.

World 1) There is our subjective world, which we form inductively from our sensory input. It is not "mere opinion", but rather each person's entire world. It is the polar opposite of solipsism to acknowledge that we are confined to this level, we do not create this world on a whim, nor with any kind of freedom: It is our map of how the world seems to behave. There are as many World 1s as there are subjects.

World 2) There is the social (quasi-real) world, in which we negotiate with other subjects all manner of things, which become very real indeed, but not in the ontological sense. Morality, Truth, Math, etc. they are all born at this level, NOT at the objective level, hence there can never be objective morality or objective truth. These are, however, concepts that largely function as if real, and which we are forced to acknowledge as real because we are social beings. The only part of these concepts that can be said to be objectively existing is the conditions in whatever parts of the objective world make up the basis for what we (in world 2 and 3) call our brains.

As can be imagined, a lot of what people think is objectively true is just World2 stuff. The English Language, for instance, only really exists in the World1 instantiations, the idiolects. But abstracting from and negotiating with other speakers (that's how we learn the language in the first place), we project an "underlying" English Language. That, however is the polar opposite of objective - it's an abstraction, and hence it is non-real.