> Not really. Plenty of theories ended up being revised when humanity's understanding of the world improved. They still had real world applications: even though they were imperfect, they were good enough for a wide variety of uses.

Those imperfect theories are applied today, and still work in lots of contexts. They are complete, but subsets. We can make predictions, given certain initial conditions, with accuracy as close to 100% as you require, the only issue is measurement. Comparing psychology to physics in this manner is pretty embarrassing, and shows a lack of understanding of physics.

Applying imperfect physical theories causes bombs to explode where they shouldn't, nuclear plants to react when they shouldn't. We don't build fusion reactors today because we still need to determine the safety of isolating the plasma.

> We have no choice but to apply these theories. The alternative would be to do nothing until they are perfected. That would be negligence. Antidepressants do not have a 100% success rate. Nothing in medicine has. That doesn't mean they shouldn't be used.

I didn't say they needed to have a 100% success rate. You just need to find the mechanism. I'd be happy with a 0.1% success rate if our determination of the mechanism implies that there is no cure once it reaches that stage.

Let's do a study--I think drinking alcohol has cured my depression. It doesn't have a 100% success rate, but it would be negligent to tell people that they shouldn't get plastered if they're depressed.

edit: Don't drink alcohol to cure your depression. I suspect the mechanism is related to the gut flora. Alcohol kills the gut flora.