> To answer that question in the title, I assume in the first instance
> that logic is a posteriori or world-dependent. I begin by arguing for a
> particular distinction between logic and informal description:
>
> A (formal) logic always reflects the worldy objects and behaviours that
> our informal descriptions have traditionally presented to us. That is,
> logic is a posteriori - world-dependent. A distinction here is that in a
> formal logic the multiplicity of contexts found in informal descriptions
> is reduced to a general case. This allows logic to be represented
> through syntax alone.
>

> However, neither a formal symbolism nor the employment of general cases
> can bring us the hoped-for elimination of informal vaguary that a formal
> logic has always seemed to promise us. Indeed, the claim of logic
> 'clearing up' informal description fails to recognise the distinctions
> between formal logic and informal descriptions that I argue for here.
>
> For example, if entailment is to be employed in an a posteriori logic
> then we must recognise entailment as an informal description sourced
> from worldy events. An applied logic then reduces the contexts of
> entailment to a general case, and presents it formally. I argued that in
> the case of modus ponens such a presentation was not accurately given to
> us by logic - this in spite of a formal representation and/or the
> simplification of contexts to a general case. We had to return to the
> informal description to inform logic correctly.
>
> Hence (if the above example is sufficient for my purposes) without
> either formal description or generalisation available as guides on our
> journey toward clarity, we might yet decide to follow logic through
> informal pathways and fall by the wayside because of it; with some
> exceptions, of course: where logic facilitates bulk operations, or,
> where a symbolism can facilitate 'speed reading'...
>
> But if it is a 'conceptual' clarity that we seek to promote in logic
> then this cannot be guaranteed by either formalism or by generalising
> particular cases; we are left to wonder precisely how such clarity could
> be achieved. But then such a means, if it existed, would make logic a
> priori and not a posteriori. So until such a means is given a name and a
> description, let us leave it the way I started out and say that logic is
> a posteriori - world dependent.