I think you make some good points. There's "science" in it's pure form, where a scientist that finds results that contradict his beliefs gets excited that this is likely to lead to more advanced and better science, and in contrast there are vested beliefs and interests where (contrary to the pure science ideal) evidence that departs from the incumbent beliefs, or vested interests, is ignored or hidden. The former must battle with the latter. Both sides have advantages: resources and (some) control of the funds allocated (the questions being asked) on the one hand, and the fact that the "truth" can only be hidden for so long (at least where it has practical observable consequences) on the other. In terms of agw, on our current course we'll know in a century who was right and who was wrong.
Further, despite the complexity surrounding climate science, in my view it's not too hard to come to an initial, rebuttable view. We know that CO2 acts as a 'greenhouse' gas, we know raising atmospheric levels of it will lead to a greater atmospheric heat retention, we know human activity has raised atmospheric levels over the last century, and we know mean temperatures have risen abnormally over the last century. So in my view it's reasonable to take a rebuttable position that agw is likely (and to try to do something about it). There may be other factors that come into play: global dimming, algal blooms, changing winds or currents, but after taking that rebuttable position, the onus should be on those that want to show that continuing excess CO2 emission won't continue to cause rising temperatures in the long term. By way of analogy, if I see (or think I see) Bob punch Jim on the nose, and then see Jim act as though he's in pain, I take the position that Bob's punch hurt Jim. Could I be wrong? Sure. There could be alternative explanations (such as play acting, or a pathetic punch from Bob barely affecting Jim, but the stress of the situation triggering one of Jim's frequent migraines) but I need some evidence, and perhaps persuasion, to sway me from my initial, reasonable position.
Incidentally, one of the best pieces of evidence I've heard for the "greenhouse" effect, is that measurements have clearly shown that mean night time temperatures are riding faster than mean daytime temperatures. This is entirely consistent with some sort of insulating effect (cooling is retarded) but not with explanations such as solar activity (which would suggest increased daytime heating, but steady or slightly increased night time cooling). I don't know why this isn't more commonly discussed.
Finally I disagree with the earlier blood-letting example. I think the opposite: if we "trust" the science, then we trust (sufficiently well established) advances in the science - that's the OPPOSITE of trusting superseded science due to some kind of faith that the "old ways" must be right.
