let's say that to say a theory or assertion is true is to say that it allows us accurately to make predictions about the future, or that it is true or true enough as long as it is not contradicted by observations. so i've begun the hawking, and he doesn't claim to generate a theory of truth. he can't, in some way, because then among other things he'd be squarely in a realm in which no particular observation or calculation could help him. he'd be in philosophy, in short. however there is a theory of truth nevertheless (and no observation or calculation can help him): "A model is a good model if it: (4) Makes detailed predictions about future observations that can disprove or falsify the model if they are not borne out." so the model makes prediction p at time t1. at t2 an observation is made that's incompatible with p (but cf quine: in fact no observation is incompatible with any theory). however, to say that it's true that that observation has actually been made at t2 is to predict that this claim, p2 (such-and-such an observation has been made), will not be falsified at some t3, which is to say that that claim, p3, will not be falsified at some t4. and so on. sometimes a view entails an infinite or vicious regress. this view entails an infinite or vicious progress, and that every assertion (even those about the past) is an infinite assertion about an infinite future. i believe that in this case no model can actually be falsified.
such a theory of truth cannot of course be attributed to hawking etc.; c.s. peirce is a plausible candidate.
anyway, here's one reason that calculations and observations can't help you generate a good theory of truth: because all calculations and observations must be deployed within a conception of truth; if the equation is not true, then generating a theory of truth from it (supposing such a thing makes sense) just takes you further afield. indeed no theory of truth can be established by any evidence. for one thing the question of the truth of the theory of truth itself arises, and a theory of truth (for example, the pragmatic theory) might be true according to itself (it helps, or whatever). that does not establish that it's true, any more than writing in purple ink 'a sentence is true if and only if it is written in purple ink' establishes that a sentence is true if and only if it is written in purple ink. ( hawking retreats to the 'goodness' of a model, probably precisely to avoid the thornbush of truth. but either he actually does mean truth or at any rate something just as problematic.) in all honesty i think theories of truth are articles of faith, or rather they are inchoate expressions of commitment to a certain sort of world. you are not going to be able to do without such commitments.