The Problem of Induction

So my science and philosophy-based professors are getting irriatated that i keep pointing out that their scientific evidences are commiting the causation/correlation fallacy. Then my philosophy teacher hit me with this: ALL scientific evidence either commits the causation fallacy or the converse accident fallacy. These are both inductive arguments, both invalid.

examples:

all P that has been observed is Q,
Therefore all P is Q.
(converse accident)

P shares a spatial/temporal correlation with Q,
Therefore P causes Q.
(causation fallacy)


I have yet to find a scientific analysis that does not commit one of these or other fallacies. Is there really no logical ground to stand on in the scientific community? I was hoping to find just one case where logic follows to a scientific conclusion, but cannot seem to.

If one does not exist, this poses an interesting dilema for the scientific community.

Wouldn't that be solved by repeatable experiments? 

If condition A exists, then event B happens, it would of course be wrong to assume that condition A caused event B.  However if you design an experiment to repeatedly reproduce condition A, whilst ensuring no other variables are present, and B always happens as a result, and at the same time a control study where condition A is not present, B does not happen, then couldn't you conclude that it is at least probable that A is the cause of B?

Quantum Mechanics explains your problem. You need to start believing in my friends, the quanta.

Science > Philosophy

Scientific theories are not logically deduced. I thought that was understood. They only make a statement about what current data suggests about the Universe, and we can never be entirely sure that they are valid, no matter how many times we repeat the relevant experiments.

But, if they are self-consistent and they make valid predictions about the Universe, we can say with some certainty that they are either correct, or (as in the case of Newtonian mechanics) they are valid approximations to reality in certain situations.

Scientific theories are not logically deduced. I thought that was understood. They only make a statement about what current data suggests about the Universe, and we can never be entirely sure that they are valid, no matter how many times we repeat the relevant experiments.

Aye, this is what many creationists see as a "chink" in the armor of evolution.  Every evolutionist worth anything will tell you that they cannot say with 100% certainty that evolution is true.  This applies to all theories across all sciences.  This open-minded approach to finding answers that allows for things to be falsifiable is often incorrectly seen as a weak point.  In fact, it is a cornerstone of what makes science what it is and if you ever see anyone arguing the contrary then it is a safe bet that you can discount their argument as nothing more than ignorant regurgitated rhetoric.

I believe it is a valid criticism of science. Obviously, science still has immense value, but it is important that people recognise that it is not an infallible system. Most people here know better, but a lot of people do seem to think that science is synonymous with truth.

Wouldn't that be solved by repeatable experiments? 

If condition A exists, then event B happens, it would of course be wrong to assume that condition A caused event B.  However if you design an experiment to repeatedly reproduce condition A, whilst ensuring no other variables are present, and B always happens as a result, and at the same time a control study where condition A is not present, B does not happen, then couldn't you conclude that it is at least probable that A is the cause of B?

Yes, but these probabilities of truth never actually reach 100% certainty. It can only approach it.

Quote from: Marcus Aurelius on September 17, 2009, 02:10:04 PM

Wouldn't that be solved by repeatable experiments? 

If condition A exists, then event B happens, it would of course be wrong to assume that condition A caused event B.  However if you design an experiment to repeatedly reproduce condition A, whilst ensuring no other variables are present, and B always happens as a result, and at the same time a control study where condition A is not present, B does not happen, then couldn't you conclude that it is at least probable that A is the cause of B?

Yes, but these probabilities of truth never actually reach 100% certainty. It can only approach it.

Absolutely, that is why every theory in science must remain falsifiable.