6 ms·

So when you can't subject something to empirical observation... such as the original creation of the sun, where exactly does one gain a sure confidence in the e

by kirse 13y ago

So when you can't subject something to empirical observation... such as the original creation of the sun, where exactly does one gain a sure confidence in the explanation for it's original cause? Let's look at a few simple cause--(produces)-->outcome examples: - When a chair moves in a room (outcome) and you didn't observe the chair moving, can you with 100% confidence identify why/what moved the chair (cause)? - When your wallet goes missing from your desk (outcome) and you didn't observe its disappearance, can you 100% identify who/what moved it (cause)? - When you walk up to your house with a broken window (outcome) having never seen the window being broken, can you 100% with confidence identify the what broke the window (cause)? I'm not saying that science is using faith to establish confidence in observable OUTCOMES. Science is using faith when it cannot observe CAUSES and proceeds to make claims about those CAUSES. Faith has to do with attributing confidence to that which cannot be observed or subject to empiricism. For example, the original cause of the sun's existence. To be clear, when I use the word "observe", I mean in the sense of using empiricism and the senses to gather data about reality.

dragonwriter 13y ago
> Interesting, so when you can't subject something to the empirical observation... such as the original creation of the sun, where exactly does one gain a sure confidence in the explanation for it's original cause? Causality, as such, is not directly observable. Correlation is, and causation is inferred from patterns of correlation. When done scientifically, the specific method to go from observed correlation to conclusions about causation is to form a hypothesis, predict future observations based on that hypothesis that could be falsified to disprove it, and then go and test those predictions. Eventually, doing this enough, and you'll end up with falsifiable models that have been rigorously tested from which you can make inferences about what must have caused events where you cannot observe the correlated event that you would infer to be the cause that are as strong as the inferences of causation you can make with events where you observe both the putative cause and the effect for which you infer it is the cause. > If you haven't observed the cause, you have no rational reason to establish 100% confidence about what it was. If you have observed the supposed cause -- and its supposed effect -- you have no rational reason to assert a causal link with actual 100% certainty. OTOH, with sufficient related observations, prediction, and testing those predictions, you can construct a model of causality that lets you get arbitrarily close to 100% certainty in stating what the cause of an observed effect is, whether or not you can observe the event that is the inferred cause. > This is really simple rational thinking It is simply incorrect, and seems grounded in both a mistake about what science does and the mistaken belief that causality is directly observable. > I don't get how science arrives at any level of confidence about causes that it can't (nor has ever) observed. The same way it arrives at any level of confidence about causes that it can observe (observe the event, that is, not the fact that it is a "cause", which is never observable, only inferable.) > I have no problem with confidence in what it can observe, I take issue when it can't observe and attempts to take what it can observe and confidently extrapolate that into the past. The whole thing science does is produce and refine methods of inferring things that are not observed from things that are.
- kirse 13y ago
  Causality, as such, is not directly observable. The whole thing science does is produce and refine methods of inferring things that are not observed from things that are. This is awful, awful science. Science works best when the entire process (from cause to outcome) can be subject to empiricism. I think we've reached an impasse here, because clearly we differ on whether or not unobservable events can be reasoned about with any level of confidence. You seem to state that inference and models that produce predictions will lead to a "arbitrarily close to 100%" level of confidence, and that's where I completely, utterly disagree... especially when we're trying to talk "billions of years ago". Even in the recent case of a broken window, only your neighbor who saw the event could have a decent level of confidence in what caused it. Millions of tests, models, and inferences would only produce an educated guess as to what happened, all of which would be obliterated by a neighbor simply explaining what they saw. As I said before, I am strictly empirical when it comes to science and am willing to call my personal claims about unseen events as "faith" when I am trying to establish truth about them.
  - dragonwriter 13y ago
    > Science works best when the entire process (from cause to outcome) can be subject to empiricism. Events can be observed. Causation is not an event, its a relationship between events which cannot be observed, only inferred. Science is entirely about developing, testing, and refining rules for inferring things that are not observed from things that are observed. > Even in the recent case of a broken window, only your neighbor who saw the event could have a decent level of confidence in what caused it. I disagree. Its quite possible, in many cases, to observe how it broke, what other objects are in proximity to the broken glass, and other things that were going on in the area and have a "decent confidence" in how it was broken without observing the actual event of breaking. > Millions of tests, models, and inferences would only produce an educated guess as to what happened To the extent that this is true, causality is always an educated guess even when you observe the event directly, and the educated guess, with or without observation of the event, can have any arbitrary degree of certainty. You seem to just be repeating the belief in the myth that causation is directly observable and that science is about reporting observed causation, and not developing predictive rules. That's just plain wrong. Empirical observation -- which never includes causation -- is the input of science, but it is no the whole of the process. If it was, science would never get beyond describing events that occur (without their causes) and observed correlations. That's not what science is.
zAy0LfpBZLC8mAC 13y ago
dragonwriter has already answered this pretty well, I think, I just want to provide one more challenge for your categorization: So when you can't subject something to empirical observation... such as the original creation of the you yourself, where exactly does one gain a sure confidence in the explanation for your original cause? Well, leaving aside the "sure confidence" (science never gets you 100 % certainty on anything ever): You correlate that for all of the other similarly shaped moistened collections of organic molecules for which you know anything of where they originated from, it was their parents. That's why you probably are pretty sure that you were born by your mother. If you want to know the same thing about our sun, you look at any number of the 100 billion or so stars in our galaxy which are in all kinds of different stages of their lives to correlate how stars commonly tend to come about - and your best guess is, again, that the sun probably happened similarly to other stars with similar observable properties. BTW, I think this is a good demonstration of your misunderstanding of how science determines probable causation: Have you ever been to a magic show? Your model of causality would mean that you would not be fooled by any of the tricks (if you could observe actual causation, you would notice the lack of causation between the misdirection and the effect). The fact that you (probably) are fooled is evidence that you do actually use the principles described by dragonwriter, you just aren't quite aware of it. Or in short: If you are interested in this, you really should read a bit about the philosophy of science. People have spent a great deal of effort on figuring out this stuff and all the fallacies that you haven't quite thought through yet.
- kirse 13y ago
  you really should read a bit about the philosophy of science Sounds like you're quite well-educated on this topic. Can you list a few recent philosophy of science books that you've read? I'm also interested in hearing your personal thoughts on each book you recommend, because I will not accept a simple copy/pasted list of coursework from Wikipedia. I want to hear about material you've personally read and your thoughts on each one, seeing as how you're thrusting these expectations on me. It's not like what I am saying is totally new here. I'm plenty well read on the topic, and most of what I've said is similar in vein to Hume [1] (and others') work on the problem of induction [2]. For such a well-educated "philosophy of science" man as yourself, I'm surprised you've missed that obvious connection. [1] http://en.wikipedia.org/wiki/David_Hume#Induction http://en.wikipedia.org/wiki/David_Hume#Induction [2] http://en.wikipedia.org/wiki/Problem_of_induction http://en.wikipedia.org/wiki/Problem_of_induction

by kirse 13y ago

dragonwriter 13y ago

> Interesting, so when you can't subject something to the empirical observation... such as the original creation of the sun, where exactly does one gain a sure confidence in the explanation for it's original cause? Causality, as such, is not directly observable. Correlation is, and causation is inferred from patterns of correlation. When done scientifically, the specific method to go from observed correlation to conclusions about causation is to form a hypothesis, predict future observations based on that hypothesis that could be falsified to disprove it, and then go and test those predictions. Eventually, doing this enough, and you'll end up with falsifiable models that have been rigorously tested from which you can make inferences about what must have caused events where you cannot observe the correlated event that you would infer to be the cause that are as strong as the inferences of causation you can make with events where you observe both the putative cause and the effect for which you infer it is the cause. > If you haven't observed the cause, you have no rational reason to establish 100% confidence about what it was. If you have observed the supposed cause -- and its supposed effect -- you have no rational reason to assert a causal link with actual 100% certainty. OTOH, with sufficient related observations, prediction, and testing those predictions, you can construct a model of causality that lets you get arbitrarily close to 100% certainty in stating what the cause of an observed effect is, whether or not you can observe the event that is the inferred cause. > This is really simple rational thinking It is simply incorrect, and seems grounded in both a mistake about what science does and the mistaken belief that causality is directly observable. > I don't get how science arrives at any level of confidence about causes that it can't (nor has ever) observed. The same way it arrives at any level of confidence about causes that it can observe (observe the event, that is, not the fact that it is a "cause", which is never observable, only inferable.) > I have no problem with confidence in what it can observe, I take issue when it can't observe and attempts to take what it can observe and confidently extrapolate that into the past. The whole thing science does is produce and refine methods of inferring things that are not observed from things that are.

kirse 13y ago
Causality, as such, is not directly observable. The whole thing science does is produce and refine methods of inferring things that are not observed from things that are. This is awful, awful science. Science works best when the entire process (from cause to outcome) can be subject to empiricism. I think we've reached an impasse here, because clearly we differ on whether or not unobservable events can be reasoned about with any level of confidence. You seem to state that inference and models that produce predictions will lead to a "arbitrarily close to 100%" level of confidence, and that's where I completely, utterly disagree... especially when we're trying to talk "billions of years ago". Even in the recent case of a broken window, only your neighbor who saw the event could have a decent level of confidence in what caused it. Millions of tests, models, and inferences would only produce an educated guess as to what happened, all of which would be obliterated by a neighbor simply explaining what they saw. As I said before, I am strictly empirical when it comes to science and am willing to call my personal claims about unseen events as "faith" when I am trying to establish truth about them.
- dragonwriter 13y ago
  > Science works best when the entire process (from cause to outcome) can be subject to empiricism. Events can be observed. Causation is not an event, its a relationship between events which cannot be observed, only inferred. Science is entirely about developing, testing, and refining rules for inferring things that are not observed from things that are observed. > Even in the recent case of a broken window, only your neighbor who saw the event could have a decent level of confidence in what caused it. I disagree. Its quite possible, in many cases, to observe how it broke, what other objects are in proximity to the broken glass, and other things that were going on in the area and have a "decent confidence" in how it was broken without observing the actual event of breaking. > Millions of tests, models, and inferences would only produce an educated guess as to what happened To the extent that this is true, causality is always an educated guess even when you observe the event directly, and the educated guess, with or without observation of the event, can have any arbitrary degree of certainty. You seem to just be repeating the belief in the myth that causation is directly observable and that science is about reporting observed causation, and not developing predictive rules. That's just plain wrong. Empirical observation -- which never includes causation -- is the input of science, but it is no the whole of the process. If it was, science would never get beyond describing events that occur (without their causes) and observed correlations. That's not what science is.

zAy0LfpBZLC8mAC 13y ago

dragonwriter has already answered this pretty well, I think, I just want to provide one more challenge for your categorization: So when you can't subject something to empirical observation... such as the original creation of the you yourself, where exactly does one gain a sure confidence in the explanation for your original cause? Well, leaving aside the "sure confidence" (science never gets you 100 % certainty on anything ever): You correlate that for all of the other similarly shaped moistened collections of organic molecules for which you know anything of where they originated from, it was their parents. That's why you probably are pretty sure that you were born by your mother. If you want to know the same thing about our sun, you look at any number of the 100 billion or so stars in our galaxy which are in all kinds of different stages of their lives to correlate how stars commonly tend to come about - and your best guess is, again, that the sun probably happened similarly to other stars with similar observable properties. BTW, I think this is a good demonstration of your misunderstanding of how science determines probable causation: Have you ever been to a magic show? Your model of causality would mean that you would not be fooled by any of the tricks (if you could observe actual causation, you would notice the lack of causation between the misdirection and the effect). The fact that you (probably) are fooled is evidence that you do actually use the principles described by dragonwriter, you just aren't quite aware of it. Or in short: If you are interested in this, you really should read a bit about the philosophy of science. People have spent a great deal of effort on figuring out this stuff and all the fallacies that you haven't quite thought through yet.

kirse 13y ago
you really should read a bit about the philosophy of science Sounds like you're quite well-educated on this topic. Can you list a few recent philosophy of science books that you've read? I'm also interested in hearing your personal thoughts on each book you recommend, because I will not accept a simple copy/pasted list of coursework from Wikipedia. I want to hear about material you've personally read and your thoughts on each one, seeing as how you're thrusting these expectations on me. It's not like what I am saying is totally new here. I'm plenty well read on the topic, and most of what I've said is similar in vein to Hume [1] (and others') work on the problem of induction [2]. For such a well-educated "philosophy of science" man as yourself, I'm surprised you've missed that obvious connection. [1] http://en.wikipedia.org/wiki/David_Hume#Induction http://en.wikipedia.org/wiki/David_Hume#Induction [2] http://en.wikipedia.org/wiki/Problem_of_induction http://en.wikipedia.org/wiki/Problem_of_induction