Elliot asking actually_thinking:
It seems so. I’m writing this for my own benefit, I think I’m learning whether actually_thinking answers or not. I’m writing in order to explain, but perhaps mostly to explain to myself. I suspect theory and observation is not the best topic to discuss, perhaps fallibilism would be best. But this has been one of my top interests of CR, so it was very fun to write, felt like only 30 minutes passed while it actually took 2 hours and 30 minutes (from reading the posts I’m responding to)!
You seem to think that just because some theory comes before the observation then that necessarily means that the observation has to become corrupted. I don’t think Popper believes this because Popper doesn’t say we can’t use observation in knowledge creation or reasoning.
I couldn’t find quotes for this (I only looked at Elliot’s recommendations, C&R and Objective Knowledge), so someone please tell me if I’m wrong or missing something important, but this is my understanding of Popper on theory before observations:
Let’s start by explaining a contradicting alternative:
Some inductivists (maybe not @actually_thinking) think that truth becomes manifest through an reading of the open book of nature. Make an observation and true knowledge will be created in your mind. But obviously you cannot only make a single observation, in that case everyone would agree about almost everything. You need to make multiple observations of the same phenomena. This is also needed because the knowledge we’re expecting to read is a pattern. By making many repeated observations, in the correct way (which I think many inductivists disagree about what the correct way is), the pattern becomes manifest.
It goes something like:
Observe phenomena → repeat in some “correct way” → pattern becomes manifest → the pattern is certain knowledge because it came directly from nature
I looked at Deutsch’s diagram in FoR which talks about doing more observations to further justify your theory. I focused on the pattern becoming manifest.
The problem with this is that the pattern never becomes manifest. Because there are always logically infinitely many patterns that are compatible with the list of sense data that we have recorded at any point of time.
Simplified and generalized example: we observe [A, B, C, A, B, C], and so we say “Aha! the pattern is manifest! [A, B, C] will repeat indefinitely.” But then we observe “D” and say “Ok… quite close, it’s [A, B, C, A, B, C, D] and that will repeat indefinitely”. However we again find that unexpected patterns seem to keep appearing. We look back and find that there was nothing logically stopping D from being a part of the pattern, and E or F could also have followed instead. We can attach any random symbol and make a new pattern repeating whatever sequence has come up until now including the new symbol.
But we have plenty of examples of people making an observation, then being intrigued by it, then making more observations, then making a generalization that explains the observations. So how does that happen?
The answer is that you start by making a theory of what the pattern is. Then you observe again and ask “does my theory of a pattern match the observation?” If the observation contradicts the pattern you scrap it and think of a new pattern. When you find a pattern that fits you say:
This works, but I cannot guarantee future observations won’t contradict it. None of the other patterns I thought of were compatible with the observations so I will hold this one to describe the true pattern the best. We wait until we can think of a new pattern that fits the same observations as well as other observations that contradict our current pattern.
So it goes something like:
Be intrigued by something → create a theory of what the pattern is → make observations and see if they fit the pattern you theorized → create new theory for a pattern if the observations contradict → find a pattern that fits, the pattern never became manifest, you theorized first then found one that fits → look for observations that contradict your current pattern such that you can improve upon it.
Notice also that in the final step we do targeted searches for observations that could contradict our pattern. This means we have to theorize about how our pattern could be wrong. Often we have a new pattern which we suspect is better than the current pattern, and we think the new pattern will fit the observations we will see whereas the current pattern will not.
So again we do a whole lot of theorizing before we make observations. We have to do this because there are infinite possibilities for patterns to match the observations. And there are infinite things to focus on while making observations. And there are infinitely many tests we could make that could contradict our theories. In order to not drown in possibilities we make theories about which patterns could fit, we theorize about what aspects to focus on which we think can make up a pattern, we theorize about which tests can contradict our theories.
We can do an experiment to illustrate:
You go into a classroom and ask your students “observe!”.
The students sit still for some time, waiting for truth to become manifest.
No truth is revealed.
Then they finally ask “observe what?”
We conclude you have to look for something in order to find it. If you’re not counting on it happening by accident.
How is it that you can notice things that are totally unexpected and you certainly did not make a theory of to predict? Something like a dragon appearing outside of your window. The answer is you have background theories of things that are expected. A dragon outside of your window is not one of those. When you look outside your window and see a dragon it does not match anything you would expect to see. It contradicts your background theories so you have to make a new theory of whatever is in front of you. You quickly theorize about what the thing could be and identify it as a dragon.
Atlas Shrugged:
The task of his senses is to give him the evidence of existence, but the task of identifying it belongs to his reason; his senses tell him only that something is, but what it is must be learned by his mind.
I don’t know whether Rand said anything about theorizing before observations. This quote however shows that she certainly did not believe that truth became manifest simply from observation. After you get your sense data you have to interpret it through theory in order to understand it.
I read a bit in On the Sources of Knowledge and of Ignorance from C&R, maybe 10% at most.