I think he meant that the process of medical diagnosis is quite different from the process of experimenting in medical science. Medical diagnoses are far less certain than the experiments in medical science are, the process of medical diagnosis doesn't follow the scientific method at all.miniboes wrote:So is your argument here that if the insights of a science are used in practice, it's therefore not a science?
But I don't see how medical diagnosis can be said to be to medical science what computer science is to actual sciences. Theoretical computer science is basically mathematics. It's hard to tell the difference between the two in many cases. It can be argued, as I did, that it uses a slightly different method (it's more likely to build upon conjectures that are considered almost certain, such as the P=/=NP, than the actual mathematics is), but that's about it. Of course, the applied computer science is more empirical. When I was making a compiler for my own simple programming language, I relied much on trial and error. I wouldn't be able to make a formal proof that it translates from my own programming language to assembly correctly (nor am I aware of the details of assembly language even remotely enough to formally define it), but it appears as though it does work. And, for some things in computer science, such as where it touches linguistics, like the speech recognition software, it is probably impossible to make formal proofs. You also probably can't make any kind of formal proof that my PacMan game works. I even couldn't explain right now how some parts of it work, for some of the bug-fixes I've made, there appears to be no sane explanation of how they work. And the same is true for some of the ways I made the compiler compatible with older browsers.
We've had a relatively long thread about how Red and BrimstoneSalad think science works, and why I think that that conception is wrong. Basically, I think that assertion being falsifiable (possible to evaluate in multiple ways) is enough for it to be scientific. BrimstoneSalad thinks that an assertion being possible to evaluate is not nearly enough, but that it also needs to be generated using some of the right methods. That is, he thinks there is some kind of a fundamental difference between a description and what he calls prediction. And he thinks that sciences function in some kind of a hierarchy, that something that is commonly accepted in physics is almost always more certain than something that is commonly accepted in, for instance, linguistics, no matter how baseless the basic tenets of modern physics seem. That is, that sciences use different methods, and that natural sciences use all the methods social sciences use (almost the opposite of what Auguste Comte claimed, but, as usual in philosophy, you can take some claim, take a contrary claim, and you can't tell which one is true ny looking at the world), and some more which somehow make them more certain. And he insists that assertions, in order to be considered scientific, need to have quantifiable certainty. I think that's the mentality at the core of many pseudosciences (from "I have a geometrical proof that the Earth is flat, so I don't have to listen to what people from softer sciences have to say!" to "Let's apply the methods I think are used in natural sciences to history."), that it doesn't actually help sciences become more rigorous, and that the premise that the certainty of the theories in linguistics can't be quantified is actually false (at the end of the thread, I presented a way the certainty of the linguistic theory I came up with can indeed be quantified).