9 ms·

This is a common misinterpretation of Occam's Razor, IMHO. Simple explanations are not more likely to be true. What Occam's Razor says is that if you have 2 t

This is a common misinterpretation of Occam's Razor, IMHO. Simple explanations are not more likely to be true. What Occam's Razor says is that if you have 2 theories that fit the same data, you should pick the simplest one. Why? Because it is simpler, not because it is more likely to be true. It is important to remember that when we are dealing with scientific theories, we are not really all that concerned with truth. Are Newton's laws "true"? Well, they work reliably given certain constraints and they are definitely more simple than other theories, so it would be stupid not to use them. However, there are data to show that Newton's laws don't hold in other situations. Newton's laws are not "truth" -- they are useful. Another thing to keep in mind is that a scientific theory must make predictions which are falsifiable. In other words, it must not only fit data in the past, it must continue to fit data into the future. If your idea has no predictions, then it isn't a scientific theory to begin with and Occam's Razor doesn't apply. In this case, the author has a theory which (presumably -- I haven't checked) fits the data in the past. Even if we had a simpler theory that also fits the data, we couldn't invoke Occam's Razor, because neither theory makes predictions about the future. The event is already finished and will not happen again. With Occam's Razor, I'm saying, "Well, we could imagine it to be like X or like Y and no matter how you slice it, it's going to work the same. I will pick X because it is simpler". But with this event, it makes a huge difference which "theory" I pick, so I can't use Occam's Razor. Even if I could pick two equivalent theories -- let's say I had a theory that the plane crashed into a mountain and I had another theory that said the plane did a loop-de-loop and then crashed into the mountain. Maybe both ideas fit the data in the past and I could go and verify that the plane ended up crashing into the mountain. Both theories fit the data in the past and fit data in the future. I'll pick the simplest one because, hey, it really doesn't matter if the plane did a loop-de-loop before it crashed. Now, maybe some people really care about the loop-de-loop. If they can't find any data that supports the loop-de-loop but doesn't support the straight-on-crash, then we are still going to reject the loop-de-loop theory. Why needlessly complicate things when there is no data to differentiate the two theories? But this does not mean that the plane did not do a loop-de-loop. There is no justification for saying that at all. We don't have any data to support one position or the other. We are just accepting the simple answer because it makes our lives easy. Where most people get this wrong is: - They try to apply it to non-scientific theories. - They try to apply it to a single theory rather than a set of equivalent theories. - They try to apply it to a set of non-equivalent theories. - They believe that the result is "true". That last one is pernicious because if you believe that something is true just because it is the simplest explanation that fits your data, you have stepped from science to religion.

crimsonalucard 12y ago
In reference to your paragraph on scientific theory: In actuality nothing in science or in reality can be proven to be true. In science it is only possible to disprove or falsify theories. You can actually extract this concept from a famous statement in science and statistics: "Correlation does not imply causation" This basically means nothing can be proven, we can only correlate things. If you negate that statement it turns into: "No Correlation does imply non-causation" Which means by establishing that two things do not correlate we've disproved or falsified causation.
- Retra 12y ago
  If P is true, its negation is false. >In actuality nothing in science or in reality can be proven to be true. If this were the case, the word "true" would have no meaning. We prove things in science all the time. We proved that you can cross a river without getting wet, for instance. We proved that humans can travel to the moon. The examples are endless, really.
  - crimsonalucard 12y ago
    >If P is true, its negation is false. Not exactly sure what you're referring to here. Still let me give you some detail on how I negated that statement: I'm essentially negating phrases within the sentence using the distributive property from algebra: ~(<Correlation> <does not imply> <causation>) = (<No Correlation> <does imply> <no causation>) If the positive statement is valid the negated statement is also valid. Negating +2 into -2 doesn't make the -2 any less true then the +2, in fact it's not even applicable. You should note that the negated statement (No Correlation does imply no causation) is a fundamental pillar in science experiments. It's part of the reason why researchers always convert their hypothesis to null hypothesis and then try to disprove it. ---------------------------------------------------------- >If this were the case, the word "true" would have no meaning. We prove things in science all the time. We proved that you can cross a river without getting wet, for instance. We proved that humans can travel to the moon. The examples are endless, really. I want to emphasize that what I'm talking about isn't some theory I made up. This is established scientific philosophy within academia. It's actually one of the harder things to understand because it flies in the face of our natural intuition. None of your examples have been "proven." What is happening is that you have formed a very reliable conclusion based on intuition, you haven't gathered proof. Lets take a look at your river example... One way in attempt to gather proof that crossing a river makes you wet is to observe 500 people crossing the river. If you observe that 500 people cross a river and all 500 got wet, what you have established is not a proof but a correlation that crossing the river is correlated to getting wet. Intuitively we assume that this is a proof, but from a rigorous and more philosophical standpoint not only have we not proven anything, it is actually fundamentally impossible to establish a "proof." Try to wrap your head around that. Let me give you an example that's easier to understand and more inline with our intuitive understanding of the world. The following link is a research study on the correlation between birthrate and the population of storks: http://web.stanford.edu/class/hrp259/2007/regression/storke.pdf http://web.stanford.edu/class/hrp259/2007/regression/storke.... They arrive at the conclusion that baby's are delivered by storks because of the positive correlation. Our intuition tells us that this conclusion is improbable despite the correlation. Note that correlation, intuition and proof are completely separate things. It is however possible to disprove things. If I establish that there is no correlation between cancer rates and heights then it is by logic impossible for height to cause higher cancer rates. What I have done is disproven height as a causative factor for cancer. It is fundamentally impossible to prove anything. It is only possible to establish correlations. Correlations can only be used to falsify things or demonstrate the possibility of causation. This is not some weird thing I made up, this is established knowledge. I don't blame people for voting my OP down; Many researchers don't consider or even learn about the implications that the scientific method has on reality as we know it.

crimsonalucard 12y ago

In reference to your paragraph on scientific theory: In actuality nothing in science or in reality can be proven to be true. In science it is only possible to disprove or falsify theories. You can actually extract this concept from a famous statement in science and statistics: "Correlation does not imply causation" This basically means nothing can be proven, we can only correlate things. If you negate that statement it turns into: "No Correlation does imply non-causation" Which means by establishing that two things do not correlate we've disproved or falsified causation.

Retra 12y ago
If P is true, its negation is false. >In actuality nothing in science or in reality can be proven to be true. If this were the case, the word "true" would have no meaning. We prove things in science all the time. We proved that you can cross a river without getting wet, for instance. We proved that humans can travel to the moon. The examples are endless, really.
- crimsonalucard 12y ago
  >If P is true, its negation is false. Not exactly sure what you're referring to here. Still let me give you some detail on how I negated that statement: I'm essentially negating phrases within the sentence using the distributive property from algebra: ~(<Correlation> <does not imply> <causation>) = (<No Correlation> <does imply> <no causation>) If the positive statement is valid the negated statement is also valid. Negating +2 into -2 doesn't make the -2 any less true then the +2, in fact it's not even applicable. You should note that the negated statement (No Correlation does imply no causation) is a fundamental pillar in science experiments. It's part of the reason why researchers always convert their hypothesis to null hypothesis and then try to disprove it. ---------------------------------------------------------- >If this were the case, the word "true" would have no meaning. We prove things in science all the time. We proved that you can cross a river without getting wet, for instance. We proved that humans can travel to the moon. The examples are endless, really. I want to emphasize that what I'm talking about isn't some theory I made up. This is established scientific philosophy within academia. It's actually one of the harder things to understand because it flies in the face of our natural intuition. None of your examples have been "proven." What is happening is that you have formed a very reliable conclusion based on intuition, you haven't gathered proof. Lets take a look at your river example... One way in attempt to gather proof that crossing a river makes you wet is to observe 500 people crossing the river. If you observe that 500 people cross a river and all 500 got wet, what you have established is not a proof but a correlation that crossing the river is correlated to getting wet. Intuitively we assume that this is a proof, but from a rigorous and more philosophical standpoint not only have we not proven anything, it is actually fundamentally impossible to establish a "proof." Try to wrap your head around that. Let me give you an example that's easier to understand and more inline with our intuitive understanding of the world. The following link is a research study on the correlation between birthrate and the population of storks: http://web.stanford.edu/class/hrp259/2007/regression/storke.pdf http://web.stanford.edu/class/hrp259/2007/regression/storke.... They arrive at the conclusion that baby's are delivered by storks because of the positive correlation. Our intuition tells us that this conclusion is improbable despite the correlation. Note that correlation, intuition and proof are completely separate things. It is however possible to disprove things. If I establish that there is no correlation between cancer rates and heights then it is by logic impossible for height to cause higher cancer rates. What I have done is disproven height as a causative factor for cancer. It is fundamentally impossible to prove anything. It is only possible to establish correlations. Correlations can only be used to falsify things or demonstrate the possibility of causation. This is not some weird thing I made up, this is established knowledge. I don't blame people for voting my OP down; Many researchers don't consider or even learn about the implications that the scientific method has on reality as we know it.