Nature of Truth

[jack]

Scott Alexander made the point that even if two concepts don't have a clearly defined boundary, they can still have a clearly defined difference. Eg. there's no official number of pebbles that make a heap, or height that makes a building "tall", but most people would agree that two pebbles are not a heap, and 50 pebbles are, and that a bungalow isn't tall and that a skyscraper is.

Some concepts do come with a clearly defined boundary, and those are often useful concepts. But many concepts are useful with a clear difference even without a clear boundary.

It occurred to me the same might be said about truth. We may not have an absolute notion of what makes a statement true, but we can still say that some concepts are closer to it (eg. mathematical proofs), some are clearly in that direction (eg. well tested scientific theories, things you've lots of experience of), and some are less close, but still better than nothing.

Comments

[gerald_duck]

My immediate instinct is that true is true, is true. But that I'd be happy to see what you've said about truth instead said about our degree of certainty that something is true.

[ptc24]

Things to be careful of: truth vs knowledge, with sensible[1] definitions of the latter starting with "justified true belief" and then arguments about Gettier cases start. A stopped clock is right twice a day; by my definitions, truth can be had from the clock twice a day, but no knowledge. I think that practically everyone these days would accept grades of knowledge.

The area for degrees of truth, I think, come from vagueness - from precisely the conceptual boundaries you can think of. Name various mountains or other pieces of elevated terrain from Everest down and you can find ones which it's hard to say whether they're very tall, whether they're even mountains as opposed to hills, or whether they're even mountains as opposed to being mere tops of "larger" mountains. So "Mount Foobar is a very tall mountain", or "there are 15 mountains in the XYZ regions" might be the sorts of statements where truth can be a graded property.

Asimov's essay the relativity of wrong sort-of takes the flip side here, and brings up examples such as "the Earth is a sphere" or "the Earth is an oblate spheroid" which could be construed as having interesting truth-values.

The popular term "half-truth" doesn't seem to be quite what we're looking for, but there's something interesting there. At school I sometimes said "I'm not a regular church-goer" when a more honest me might have said "I'm an atheist" - the former is literally true before you take pragmatics - notably Gricean implicature - into account but once you do, it imparts false believes.

[1] Non-sensible definitions include "sociological" definitions, including "what's commonly believed". There's a general heading of "epistemic" theories of truth which I have little time for, at best I think they conflate truth with knowledge and at worst indulge in all sorts of postmodernist nonsense. Also note the existence of non-propositional knowledge, such as all the knowledge of how to move my fingers to type this.

[nameandnature]

I remember Simon Blackburn's book Truth having some interesting stuff to say about this, but I don't remember enough of it to say anything more here :-) Time to put it in the rucksack and read it at lunch...