on the notion of correlation and causation

[taks]

since the nobel thread is no longer open, i should probably respond in a related thread...

 

However, correlation never implies causality

completely false. it is true the broad statement "correlation does not imply causality" is valid, but to use the word "never" is completely false. my statements regarding the causal relationship between CO2 and temperature based on high correlation are indeed valid, but not solely because of the correlation alone (and i duly noted the third party actor in this). there is a causal relationship, though not a direct cause-effect relationship, which i also pointed out. i also noted that the higher the correlation, the more likely the cause-effect relationship becomes. btw, if "never" were true, communications and radar, indeed nearly all sciences based on signal processing in which we are attempting to detect some event would not work. why? they (signal receivers) use either a matched filter, or its equivalent, the correlator receiver.

 

from the wiki:

To say a "Correlation does not suggest causation" is false: A demonstrably consistent correlation often suggests or increases the probability of some causal relationship (or implies it, in the casual sense of the word).

 

bold and italics mine.

 

in the specific case of the temp vs. CO2 scenario, we have a known causal relationship, namely: the solubility of CO2 in water is reduced with increased temperature. the obvious follow-on to this concept is that as the oceans warm, which is the primary sink for CO2 (as well as source) the oceans release more CO2 into the atmosphere.

 

the wiki offers a follow up example with this conclusion:

 

Determining whether there is an actual cause and effect relationship requires further investigation, even when the relationship between A and B is statistically significant, a large effect size is observed, or a large part of the variance is explained.

 

italics mine again.

 

the "further investigation" in this case is quite simply, the known relationship between the solubility of a gas in a liquid and the temperature of that liquid. in that context, which i was specifically referring to, both necessary (correlation) and sufficient (known relationship between ocean temperature and CO2 solubility) conditions exist. therefore, in this case, correlation does imply causation.

 

that said, the likely culprit, historically speaking, is that the total solar irradiance incident on the planet increases/decreases and likewise causes a change in the atmospheric temperature. the oceans, containing significantly more mass than the atmosphere, warm at a slower rate. as they warm, they release more CO2 into the atmosphere, but this will necessarily lag the increase in atmospheric temperature as well. the cause-effect relationship is clearly there, and i even offered that it was not the increased temperature of the atmosphere that was the cause, but some other forcer, probably the sun. the reason for the change in solar irradiance, as well as other extra-planetary influences, are many, and not limited simply due to a change in the sun itself. indeed, the earth is not in a completely spherical orbit, and it "wobbles" around its axis (the probable cause of ice ages). the sun itself has cycles of low and high activity, the sun spot cycle which is 11 years, and probably others that are not completely understood.

 

taks

Edited October 15, 2007 by taks

[taks]

i should add that the short term correlation between CO2 and temperature is weak at best. this is hardly a surprise since the correlation actually occurs with a very large lag, between 800 and 2000 years. looking at the last 100 years, which is not long enough to measure an 800 year lag, the largest increases in atmospheric CO2 content occurred post WWII, but the temperature of the planet dropped several tenths of a degree © from then till the mid-70s. we are certainly adding to the content through the burning of fossil fuels. however, if CO2 were the/a cause, then the historic record would show CO2 leading, rather than lagging, temperature as well, e.g. as the oceans warmed and CO2 increased, at some point, if CO2 were indeed a cause not an effect, it would begin to lead the temperature. but it does not.

 

taks

[Qwerty the Sir]

completely false. it is true the broad statement "correlation does not imply causality" is valid, but to use the word "never" is completely false. my statements regarding the causal relationship between CO2 and temperature based on high correlation are indeed valid, but not solely because of the correlation alone (and i duly noted the third party actor in this). there is a causal relationship, though not a direct cause-effect relationship, which i also pointed out. i also noted that the higher the correlation, the more likely the cause-effect relationship becomes.

 

It is not completely false taks it is completely true, and in the rest of my statement which was not quoted, I go on to show that.

 

Now, here is that part again:

 

...and even very strong correlations suggesting some sort causality is a very tricky point that is hotly debated (you with your supposed background in statistics must know of the problems with the problems of applying Bayesian analysis and probability values to events with a correlation as the basis on which the casual link is supposed to be inferred (either through a logical necessity [not happening] or through a empirical verification of a casual nexus [problematic, as we only see events in succession and never actually perceive any casual nexus between them; the casual link is formed through our minds]).

 

Again, the reason why correlation can never imply any sort of causation or casual link is because the casual link is not a logical necessity and it is not empirically perceived. Hume and before him medieval occasionalists pointed this out.

 

After understanding this, some people will then go to the claim that correlation never implies a causation, but suggests one, and they use probability. They say, like you that "the higher the correlation, the more likely the cause-effect relationship becomes".

 

However, there really is no justification for that statement. Why is it the case that the higher the correlation, the more likely a cause-effect relationship exists? Since we still do not perceive any sort of causation.

 

When we look about us towards external objects, and consider the operation of causes, we are never able, in a single instance, to discover any power or necessary connexion; any quality, which binds the effect to the cause, and renders the one an infallible consequence of the other. There is required a medium, which may enable the mind to draw such an inference, if indeed it be drawn by reasoning and argument. What that medium is, I must confess, passes my comprehension; and it is incumbent on those to produce it, who assert that it really exists, and is the origin of all our conclusions concerning matter of fact. This question I propose as much for the sake of information, as with an intention of raising difficulties. I cannot find, I cannot imagine any such reasoning. But I keep my mind still open to instruction, if any one will vouchsafe to bestow it upon me.

 

The same problem we have when saying that correlation implies causation occurs when saying that correlation makes it more likely that causation exists.

 

To any purporter of the statement "correlation implies causation" or "correlation likely suggests causation" must show why this is true.

 

btw, if "never" were true, communications and radar, indeed nearly all sciences based on signal processing in which we are attempting to detect some event would not work. why? they (signal receivers) use either a matched filter, or its equivalent, the correlator receiver.

 

Let's take an example of two events which seem to be 100% correlated and which everyone assumes a casual link occurs. Fire burning cotton (this is an example that Hume and Ghazali both used when dealing with causation).

 

In everyday speak, we say that cotton, when placed into fire, is burned by it (the fire causes the burning of cotton). We say this because 100% of the time in the past (at least to my knowledge) when cotton is placed into a fire, the cotton is burned.

 

But we actually are only seeing two events coming one after the other in succession, the cotton is placed into a fire, and the cotton is burned. We never actually perceive any kind of casual link between the two. What inference, then allows us to make the conclusion that the fire caused the burning of the cotton. Certainly, we cannot purport any logical necessity between the two (it involves no contradiction to say that fire does not burn cotton), and we have no empirical basis to form a link...

 

It's just one of those pesky epistemic problems of science and scientific inference that still needs to be resolved...

 

from the wiki:

 

To say a "Correlation does not suggest causation" is false: A demonstrably consistent correlation often suggests or increases the probability of some causal relationship (or implies it, in the casual sense of the word).

 

Someone really needs to change that Wikipedia entry then; they have made an assertion, and yet not shown it. Again, the same question posed to you above can be posed to the Wikipedia article.

 

How exactly does a demonstratively consistent correlation suggest increased probability of a casual relationship?

 

Going on the case with Gore's film, you were quite correct to show that Gore had the relationship mixed up. It makes more sense to point to a graph and say that higher temperature is correlated with higher carbon dioxide levels as the former usually precedes the latter (according to the best data gathered).

 

So Gore would not really even be able to make the jump from the correlation to a cause-effect relationship because he is misreading the data in the first place. However, neither is it correct for you to say that "second, correlation does indeed imply at least some causation, like it or not. that the correlation exists over six hundred and fifty thousand freaking years is clear evidence that there is a cause-effect relationship. unfortunately for the activist world, since temperature precedes CO2 rise, that means temperature is the cause, not the effect. oops."

[taks]

Again, the reason why correlation can never imply any sort of causation or casual link is because the casual link is not a logical necessity and it is not empirically perceived. Hume and before him medieval occasionalists pointed this out.

i already supplied the reason why the word never is false. every definition you can find will say the same thing i said above:

 

To say a "Correlation does not suggest causation" is false: A demonstrably consistent correlation often suggests or increases the probability of some causal relationship (or implies it, in the casual sense of the word).

 

After understanding this, some people will then go to the claim that correlation never implies a causation, but suggests one, and they use probability. They say, like you that "the higher the correlation, the more likely the cause-effect relationship becomes".

 

However, there really is no justification for that statement. Why is it the case that the higher the correlation, the more likely a cause-effect relationship exists? Since we still do not perceive any sort of causation.

again, there needs to be some sufficient mechanism in order to justify using the term implies, which exists in the specific case i pointed out.

 

The same problem we have when saying that correlation implies causation occurs when saying that correlation makes it more likely that causation exists.

 

To any purporter of the statement "correlation implies causation" or "correlation likely suggests causation" must show why this is true.

and i have shown why, in this specific case, the implication is true. there is a known physical relationship between the temperature of the ocean and the amount of CO2 it releases. i.e., i met the condition of sufficiency. this is science, not philosophy.

 

Let's take an example of two events which seem to be 100% correlated and which everyone assumes a casual link occurs. Fire burning cotton (this is an example that Hume and Ghazali both used when dealing with causation).

 

In everyday speak, we say that cotton, when placed into fire, is burned by it (the fire causes the burning of cotton). We say this because 100% of the time in the past (at least to my knowledge) when cotton is placed into a fire, the cotton is burned.

 

But we actually are only seeing two events coming one after the other in succession, the cotton is placed into a fire, and the cotton is burned. We never actually perceive any kind of casual link between the two. What inference, then allows us to make the conclusion that the fire caused the burning of the cotton. Certainly, we cannot purport any logical necessity between the two (it involves no contradiction to say that fire does not burn cotton), and we have no empirical basis to form a link...

what? the sufficiency in this statement is that there is a known physical relationship between the heat of the fire, and the burning of the cotton.

 

Someone really needs to change that Wikipedia entry then; they have made an assertion, and yet not shown it. Again, the same question posed to you above can be posed to the Wikipedia article.

 

How exactly does a demonstratively consistent correlation suggest increased probability of a casual relationship?

now you're getting into philosophical nits, but even you note that the concept is increased probability, which is entirely what the correlation is attempting to show. actually, what correlation shows is increased dependence, though not necessarily cause-effect. wiki makes it clear that causal relationships need another factor in order to be established.

 

in nature, most things being either random or chaotic, correlations typically result from either direct cause-effect e.g. rain makes sidewalks wet, or indirect cause-effect, e.g. the sun causes temperatures to rise in both the atmosphere and the oceans, and the oceans release more CO2 as a result. the atmospheric temperature is thus highly correlated (at a time lag). this does not imply that the atmospheric temperature causes the CO2 rise, rather, it implies that there is likely to be some causal relationship between the two. in this case, the causative agent is the solar irradiance, which directly caused the rise in both atmospheric and oceanic temperatures and thus, increased CO2.

 

Going on the case with Gore's film, you were quite correct to show that Gore had the relationship mixed up. It makes more sense to point to a graph and say that higher temperature is correlated with higher carbon dioxide levels as the former usually precedes the latter (according to the best data gathered).

i actually made that point several times and i noted that the "cause" was likely some other agent, namely the sun.

 

So Gore would not really even be able to make the jump from the correlation to a cause-effect relationship because he is misreading the data in the first place. However, neither is it correct for you to say that "second, correlation does indeed imply at least some causation, like it or not. that the correlation exists over six hundred and fifty thousand freaking years is clear evidence that there is a cause-effect relationship. unfortunately for the activist world, since temperature precedes CO2 rise, that means temperature is the cause, not the effect. oops."

the correlation alone does not establish the causation. i noted in several places that the contributing factor was the known physical relationship between the oceanic temperature and rate of CO2 release (though that is not immediately clear with this quote alone, it must be deduced from the rest of my argument in which i made my assertion more clear). i.e. i met both conditions, necessity, and sufficiency.

 

note, too, i never claimed this was a proven causality. the use of the word "implied" can be debated on a semantic level for sure, and even the wiki entry uses "suggests," which is probably, from a philosophical standpoint, more appropriate.

 

taks

[Qwerty the Sir]

taks, if you are going to dismiss my arguments against casual relationships by saying that "this is science, not philosophy" (this is actually an exploration of the epistemic nature and value of a scientific and statistical maxim), then I'm afraid we'll be at an impasse.

 

You are basically presupposing that correlation implies or suggests causation in each one of your examples. The "known physical relationships" you speak of are nothing more than correlations that have been observed, which you already presuppose are casual relationships...

 

You are saying (to take one example) that we already know that there is a relationship between the heat of fire and the burning of cotton, and that is the reason why there is a correlation between them.

 

This does nothing to actually justify the claim of "To say a "Correlation does not suggest causation" is false: A demonstrably consistent correlation often suggests or increases the probability of some causal relationship (or implies it, in the casual sense of the word)", as it is already presupposed.

 

So when you say:

 

the correlation alone does not establish the causation. i noted in several places that the contributing factor was the known physical relationship between the oceanic temperature and rate of CO2 release (though that is not immediately clear with this quote alone, it must be deduced from the rest of my argument in which i made my assertion more clear). i.e. i met both conditions, necessity, and sufficiency.

 

... you are saying that the "known physical relationship" is what established the casual link, correct? So how was the physical relationship known? Well, it was by the observed correlation between the oceanic temperature and the rate of carbon dioxide release. Now, you are arguing in a circle (which is valid, and fine [for you] if you are already presupposing your premise, however, I am not).

 

You have really not offered any justification for the statement "To say a "Correlation does not suggest causation" is false: A demonstrably consistent correlation often suggests or increases the probability of some causal relationship (or implies it, in the casual sense of the word)", which is what I was asking for from the beginning.

 

You are looking into the issue of correlation and causation from inside it. I am looking into the issue from outside of it. You are already assuming that correlation implies causation; I am not. The impasse lies here.

 

 

--

 

Just a quick note on the carbon dioxide-global warming examples, I think I have been a bit ambivalent on my dealing with global warming with my previous statements (due to other issues that distracted me). I do indeed think that Gore's film [and not the scientific community's view] had overinflated claims and was sensationalist to say the least, and I do believe that public policy should remain as far away from the influence of science as possible, and I am really not interested in the nitty bitty details of scientific data and research (what I'm interested in is the relationship of science and society).

 

With that, you can probably guess I am quite apathetic towards the issue of global warming (aside from how it relates to society and policy).

 

That being said, I do know that you are kinda on shaky ground on a couple of your statements concerning the issue.

 

Firstly, on your statement that "most scientists believe only one thing about GW: the planet has warmed" (it was in reply to a quote that scientists [and hippies] believe humans are a significant factor in the warming trend), it is also the case that most scientists and scientific organizations endorse the view that humans play a significant role or the main role in the warming trend.

 

http://en.wikipedia.org/wiki/Scientific_op..._climate_change (a wiki link for convienence)

 

Secondly, on the fact that temperature's rise before carbon dioxide levels, the majority of scientists who endorse the above view explain that in terms of feedback argument.

 

http://en.wikipedia.org/wiki/Attribution_o...s_CO2_increases (another wiki link for convenience).

[taks]

taks, if you are going to dismiss my arguments against casual relationships by saying that "this is science, not philosophy" (this is actually an exploration of the epistemic nature and value of a scientific and statistical maxim), then I'm afraid we'll be at an impasse.

you're citing hume as a reference, and based on what i read, he seems to conclude that cause-effect is purely a philosophical idea, invented by the mind. in nature, this is fundamentally a flawed premise: cause-effect certainly exists, though we can only test it empirically.

 

You are basically presupposing that correlation implies or suggests causation in each one of your examples. The "known physical relationships" you speak of are nothing more than correlations that have been observed, which you already presuppose are casual relationships...

 

You are saying (to take one example) that we already know that there is a relationship between the heat of fire and the burning of cotton, and that is the reason why there is a correlation between them.

 

This does nothing to actually justify the claim of "To say a "Correlation does not suggest causation" is false: A demonstrably consistent correlation often suggests or increases the probability of some causal relationship (or implies it, in the casual sense of the word)", as it is already presupposed.

here's where your argument becomes scientifically untenable. cause-effect relationships do indeed exist. if we take your exposition at face value, we can never empirically know anything about the world around us.

 

... you are saying that the "known physical relationship" is what established the casual link, correct? So how was the physical relationship known? Well, it was by the observed correlation between the oceanic temperature and the rate of carbon dioxide release. Now, you are arguing in a circle (which is valid, and fine [for you] if you are already presupposing your premise, however, I am not).

the known physical relationship satisfied the sufficiency, and the correlation satisfied the necessity. in science, one must always believe in certain truths, which is not necessarily valid for a philosophical debate. empirically measured relationships are such truths. if we are not allowed to do so, then all of science suffers from the same problem, and nothing can ever be stated with any measure of certainty. heck, even the "law" of gravity would be invalid. this is preposterous.

 

You are looking into the issue of correlation and causation from inside it. I am looking into the issue from outside of it. You are already assuming that correlation implies causation; I am not. The impasse lies here.

you are looking at it through the eyes of a philosopher, not a scientist. again, science is not possible unless at least some truths are accepted as fact through empirical study. proof can never exist in science, or at least rarely, as that is generally attributed to mathematical expositions.

 

Just a quick note on the carbon dioxide-global warming examples, I think I have been a bit ambivalent on my dealing with global warming with my previous statements (due to other issues that distracted me). I do indeed think that Gore's film [and not the scientific community's view] had overinflated claims and was sensationalist to say the least, and I do believe that public policy should remain as far away from the influence of science as possible, and I am really not interested in the nitty bitty details of scientific data and research (what I'm interested in is the relationship of science and society).

though immaterial, i did not infer your position one way or another from any statements you made.

 

With that, you can probably guess I am quite apathetic towards the issue of global warming (aside from how it relates to society and policy).

indeed.

 

That being said, I do know that you are kinda on shaky ground on a couple of your statements concerning the issue.

 

Firstly, on your statement that "most scientists believe only one thing about GW: the planet has warmed" (it was in reply to a quote that scientists [and hippies] believe humans are a significant factor in the warming trend), it is also the case that most scientists and scientific organizations endorse the view that humans play a significant role or the main role in the warming trend.

 

http://en.wikipedia.org/wiki/Scientific_op..._climate_change (a wiki link for convienence)

that's an opinion piece, which really has no relevance in terms of evidence for or against my assertion. clearly nearly everyone agrees that the planet has warmed. even with the latest problems uncovered with the data and methods, there's still at least some warming, though the exact magnitude is debatable. satellite data showed warming from the late 70s till the late 90s, and it has been mostly flat since then (hard to really give a 30 year trend any significance in light of a 4 billion year climate history, however). in order for you to conclude that "most" agree humans play a significant, you'd need a bit of the empirical evidence which you otherwise don't seem to think exists. indeed, even if we ignore that strangeness, it would require polling all scientists with any interest in climate for their opinions. that said, it is a stretch for me to say "only thing" as well, but that's the easiest to uncover because most of the data is irrefutable. there have been plenty of studies, and all of them seem to disagree to what extent any consensus exists OTHER than what we can absolutely test: a warming trend.

 

Secondly, on the fact that temperature's rise before carbon dioxide levels, the majority of scientists who endorse the above view explain that in terms of feedback argument.

 

http://en.wikipedia.org/wiki/Attribution_o...s_CO2_increases (another wiki link for convenience).

oh, i know all about the feedback argument, and it is nonsense. this is a control theory problem. feedback cannot violate causality, no matter how much they hand wave on the idea. the ONLY way for cause to follow effect is through an acausal (or non-causal) system, which is not possible in nature. i can construct an acausal system simulation, for sure, but it is not really acausal because i'd be operating on data that already existed. if nature was suddenly able to develop psychic powers, i suppose it could be done, but i'm not sold on this possibility either (john edwards is a hoot, and a hoax).

 

by definition, feedback causes LAG, not lead. consider the simple difference equation (x is input, y is output):

 

y(n) = y(n-1) + x(n)

 

for a DC input, the phase response is a 90 degree LAG, which increases up to 0 at pi radians/sample. in order to get a phase lead, you'd have to reformulate the equation as

 

y(n) = y(n+1) + x(n)

 

the phase lead starts at 0 degrees and increases up to 90 degrees LEAD at pi radians/sample (pi radians per sample implies an oscillation at exactly half the sample rate, in binary terms it would be +1, -1, +1, -1, etc...)

 

however, at time n, the feedback term is from time n+1, which has not occurred yet.

 

their argument is absolutely bogus.

 

taks

Edited October 15, 2007 by taks

[Gromnir]

logic, more often than not, is a blade w/o a handle. in any event, it feels strange being no more than an observer in one of these threads... almost surreal.

 

HA! Good Fun!

[blue]

Some context for readers:

 

The Nobel Prize that sparked this debate was jointly awarded to Gore and the Intergovernmental Panel on Climate Change. The IPCC reports represent the work of thousands of scientists with scholarship extending back over two decades. By way of example, here's a list of contributors to the 4th assessment report of the first working group.

 

IPCC publications represent an attempt to form a consensus among leading scholars, but some contributors do not endorse IPCC positions -- particularly the statement that it is "90% likely" man is having an impact on world temperatures. Some researchers have publicly disagreed with committee findings.

 

The issue of consensus is contentious. A 2004 study found significant (75%) endorsement (either implicit or explicit) of IPCC positions, but a more recent sampling found that less than half of published scientists endorsed IPCC ideas with only 7% providing unequivocal support. This latest survey has not been peer-reviewed, and remains the source of considerable controversy. Nevertheless, it underscores divisions within the scientific community and suggests that a significant portion of opinion is neutral, neither accepting nor rejecting climate change hypotheses.

 

I'll add that while neutrality is essential to the scientific method, it is frequently absent from debates about climate change. Obsidian forums not excepted.

 

This comment is worth a second glance:

http://forums.obsidianent.com/index.php?sh...mp;#entry748745

 

/offtopic

Edited October 15, 2007 by blue

[taks]

curious if BLUE means "best linear unbiased estimator?"

 

though off-topic, the actual IPCC report was actually only written by a handful of scientists, which is probably why many of the "contributors" do not agree with all the conclusions. this is also the basis of my assertion that most do not agree with all the conclusions of AGW theory, the sole testable component being rising temps.

 

and yes, the oreskes study has been largely discredited since their assumption of implicit endorsement included even those papers that only agreed with part, often only one piece, of the complete theory (or hypothesis, as i prefer, though that is semantic as well).

 

what is on-topic, however, is that most connections between the CO2/rising temperature conundrum is primarily based on correlation, sometimes strong, other times not at all. the concepts of heat retention by CO2 have some scientific basis, though the magnitude of the effect is currently fiercely contested, particularly theories on how much CO2 equates to X rise in temperature. steve mcintyre over at climateaudit has requested a mainstream text that explains why a doubling of CO2 would add 2.5 C to the temperatures without any reasonable response to date.

 

taks

[Colrom]

The thing to do, in these cases is to read the peer reviewed papers - not the web publications and the press releases and all that stuff - the peer reviewed papers.

 

As far as I have been able to tell and also those other folks who have actually done analysis of the data and been reviewed by experts and been given awards, those who sneer at the global warming models and predictions are wrong and the popular mischaracterizations and misinterpretations of the data and the science are flawed in a variety of ways - in some cases willfully wrong.

 

The thing to do is fire somebody who makes too many mistakes - and anybody who makes willful errors.

 

Us Physicists and such need to think smartly or go into other areas of work.

 

Maybe politics?

Edited October 15, 2007 by Colrom

[Tale]

The thing to do, in these cases is to read the peer reviewed papers - not the web publications and the press releases and all that stuff - the peer reviewed papers.

 

As far as I have been able to tell and also those other folks who have actually done analysis of the data and been reviewed by experts and been given awards, those who sneer at the global warming models and predictions are wrong and the popular mischaracterizations and misinterpretations of the data and the science are flawed in a variety of ways - in some cases willfully wrong.

 

The thing to do is fire somebody who makes too many mistakes - and anybody who makes willful errors.

 

Us Physicists and such need to think smartly or go into other areas of work.

 

Maybe politics?

 

But, I read a post by someone on a forum saying that all the peer reviewed journals are in it for the money and everyone who disagrees is too afraid to speak out! The politicians are conspiring to control information!

[Qwerty the Sir]

you're citing hume as a reference, and based on what i read, he seems to conclude that cause-effect is purely a philosophical idea, invented by the mind. in nature, this is fundamentally a flawed premise: cause-effect certainly exists, though we can only test it empirically.

 

Hume says that causality is invented in the mind precisely because nobody has ever shown how it could be anything else but a figment of the imagination. It is not his premise that is flawed, it is the premise he critiques that is flawed. If cause-effect relationships exist, why can we not perceive them empirically. You say we can only test it empirically, but in fact, the only thing we get out of our testing is empirical data of a correlation. We never empirically perceive a casual link.

 

here's where your argument becomes scientifically untenable. cause-effect relationships do indeed exist. if we take your exposition at face value, we can never empirically know anything about the world around us.

 

Again you restate that cause effect relationships exist, but without any satisfactory justification for the assertion. Your second sentence is in the ballpark of correctness though. The conclusion that Hume would make is that we can never have rational justification for our knowledge about "matters of fact" (the "world around us" is included in its domain), as that knowledge is gained through an irrational method.

 

To take the fire burning cotton example, Hume would say that it is wrong to say that the fire causes burning of the cotton, because we can never perceive any casual relationships, only correlations, and there is no valid inference from correlations to casual links.

 

the known physical relationship satisfied the sufficiency, and the correlation satisfied the necessity. in science, one must always believe in certain truths, which is not necessarily valid for a philosophical debate. empirically measured relationships are such truths. if we are not allowed to do so, then all of science suffers from the same problem, and nothing can ever be stated with any measure of certainty. heck, even the "law" of gravity would be invalid. this is preposterous.

 

The known physical relationship was inferred from the correlation in the first place. To use it as if it existed before empirical data of the correlation between two things was known is definitely problematic in the matter of justification.

 

Now, if you are looking at the issue already presupposing causality, then it is not much of a problem, but if you were to try and convince someone of who didn't presuppose it, then you must offer a justification for the principle. Until you do so, it is quite reasonable for them to reject any scientific conclusions or theories based on cause-effect relationships as irrational.

 

Like Hume says:

 

There is required a medium, which may enable the mind to draw such an inference, if indeed it be drawn by reasoning and argument. What that medium is, I must confess, passes my comprehension; and it is incumbent on those to produce it, who assert that it really exists, and is the origin of all our conclusions concerning matter of fact. This question I propose as much for the sake of information, as with an intention of raising difficulties. I cannot find, I cannot imagine any such reasoning. But I keep my mind still open to instruction, if any one will vouchsafe to bestow it upon me.

 

As for the second part of your statement, you are right that all of science (and much more) suffers from the same problem, and that nothing (that is "matter of fact" knowledge) can be known at all, and that scientific theories and laws would become decidedly invalid and irrational to hold.

 

I don't see what is preposterous about it at all however, except for the fact that it is an unpalatable conclusion for some. Like I said before, this is one of the epistemic problems with science, its methods, and its inferences.

 

you are looking at it through the eyes of a philosopher, not a scientist. again, science is not possible unless at least some truths are accepted as fact through empirical study. proof can never exist in science, or at least rarely, as that is generally attributed to mathematical expositions.

 

Ok, and you are looking at it with the eyes of a scientist (or statistician). This is why I said there is an impasse.

 

that's an opinion piece, which really has no relevance in terms of evidence for or against my assertion. clearly nearly everyone agrees that the planet has warmed. even with the latest problems uncovered with the data and methods, there's still at least some warming, though the exact magnitude is debatable. satellite data showed warming from the late 70s till the late 90s, and it has been mostly flat since then (hard to really give a 30 year trend any significance in light of a 4 billion year climate history, however). in order for you to conclude that "most" agree humans play a significant, you'd need a bit of the empirical evidence which you otherwise don't seem to think exists. indeed, even if we ignore that strangeness, it would require polling all scientists with any interest in climate for their opinions. that said, it is a stretch for me to say "only thing" as well, but that's the easiest to uncover because most of the data is irrefutable. there have been plenty of studies, and all of them seem to disagree to what extent any consensus exists OTHER than what we can absolutely test: a warming trend.

 

First let me correct you on attributing to me that I don't think empirical evidence exists. I am not propounding a skepticism of the external world here. Indeed, how can I be accused of holding that position when I affirm that there is indeed empirical evidence for correlations between two things. What I am propounding is that there is no empirical evidence for casual links and causality, as we do not perceive causality. Add to that the fact that there is no valid inference from data of correlation to causal links, and you get the basic gist of my argument: there is no reason to believe cause and effect relations exist.

 

Yes, it is an opinion piece, as is your opinion piece that human don't play a significant part, and natural causes play the most significant part. Now, I don't know about the inclinations of others, but disregarding my skepticism towards science (due again, to epistemic problems with some questionable assumptions it makes), if I were to hold a strong position on the issue, it would be in line with the vast majority of the scientific organizations.

 

Also, I would say that there is quite a bit of empirical evidence against your assertion given how many scientists let those organizations speak for them (I'll try to bring up a count of signatures of all the scientists who have latched on to the manifesto that global warming has a significant or main cause in human activities, and we'll compare it to the scientists who disagree).

 

Frankly though, 15 major scientific organizations propounding the view that human activities play a significant role or play the main role in the current trend of global warming says quite a lot, and at least for me, it more than demonstrates your assertion as incorrect.

 

If you want more evidence, that is more in line with statistical methods, try some of these polls of scientists on the issue:

 

http://www.washingtonpost.com/wp-dyn/conte...7020400953.html

http://www.csm.ornl.gov/PR/NS-10-25-03.html

 

Both links show that there is wide agreement among scientists that human activities play a significant factor in global warming (on the severity of the problem, there is disagreement).

[Colrom]

Here's the executive summary of the UN IPCC report:

 

http://www.aaas.org/news/press_room/climat...h_spm2feb07.pdf

 

There are lots of pretty pictures. Some pictures show CO2 ice sample results. Others show a variety of global temperatures.

 

Model predictions are also given.

 

Some of the scientists involved complained that the report was toned down somewhat under political pressure.

[taks]

peer review is not without its problems. peer review is NOT audit, nor does it constitute any form of an attempt at falsification. there is also a form of confirmation bias that results from reviewers that are not otherwise unbiased: they control the keys to publication, which may result in a sort of filtering of ideas that do not adhere to widely held views.

 

there are plenty of "con" AGW papers out there, btw, they simply don't get press. this isn't necessarily a condemnation of peer review, or science in general, or even of climate science, but more of media expectations of what the public wants to hear. of course, nobody wanted to hear einstein's theory either, until it finally did make it mainstream and oila, an upheaval of quite a few rather cherished opinions that had been otherwise mainstream for a long time (what rarely gets mentioned is that einstein's theories were largely based on a lot of previous work that simply didn't get a lot of press, see lorentz, for example).

 

yes, btw, tale, most journals do have a profit motive which can clearly affect their objectivity, though i do not think it fair to condemn them all as a result. there are also many scientists that are afraid to speak their minds (i personally am in contact with two competent statisticians/signal processing experts that refuse to divulge their names for fear of losing their jobs) because of the political climate surrounding, well, the climate...

 

taks

[taks]

Here's the executive summary of the UN IPCC report:

 

http://www.aaas.org/news/press_room/climat...h_spm2feb07.pdf

 

There are lots of pretty pictures. Some pictures show CO2 ice sample results. Others show a variety of global temperatures.

 

Model predictions are also given.

 

Some of the scientists involved complained that the report was toned down somewhat under political pressure.

yeah, the SPM. written by the same guys that reviewed their own work in the working groups. there were also some that complained the results were exaggerated. note, too, the SPM was written before the body of work was released, which is highly unusual for standard publications.

 

taks

[Colrom]

By the way, the percentage change in the CO2 concentration over the last few centuries or so is much much much greater than the percentage change in the temperature over that same time frame (lets use centigrade or kelvin temperatures). Before that the CO2 concentration was nearly constant. This of course is not suprising since the earth is kinda large. But it does make the whole idea of the temperature change leading the CO2 concentration seem like a strange joke from another universe. Of course they really don't mean that do they? But then why don't they say what they mean and not try to play with the heads of people who arn't scientists.

 

It's not nice to fool with mother nature! or decent people who would just like some straight info! for a change! maybe just this one time!

[Colrom]

Taks said:

 

"yes, btw, tale, most journals do have a profit motive which can clearly affect their objectivity, though i do not think it fair to condemn them all as a result. there are also many scientists that are afraid to speak their minds (i personally am in contact with two competent statisticians/signal processing experts that refuse to divulge their names for fear of losing their jobs) because of the political climate surrounding, well, the climate..."

 

I don't understand.

 

Have they done something - anything?

 

Have they said something to somebody - anything?

 

Let then do something and say something and then they can be afraid.

 

And then it will be too late and you will be able to quote them!

Edited October 15, 2007 by Colrom

[taks]

If cause-effect relationships exist, why can we not perceive them empirically. You say we can only test it empirically, but in fact, the only thing we get out of our testing is empirical data of a correlation. We never empirically perceive a casual link.

i disagree. cause-effect is easily observed, at least at the macro level (quantum causality gets strange).

 

Again you restate that cause effect relationships exist, but without any satisfactory justification for the assertion. Your second sentence is in the ballpark of correctness though. The conclusion that Hume would make is that we can never have rational justification for our knowledge about "matters of fact" (the "world around us" is included in its domain), as that knowledge is gained through an irrational method.

this is one of the "truths" we have to accept as existing. your assertion that i need to provide some satisfactory justification of causal relationships is a philosophical nit.

 

To take the fire burning cotton example, Hume would say that it is wrong to say that the fire causes burning of the cotton, because we can never perceive any casual relationships, only correlations, and there is no valid inference from correlations to casual links.

few, if any, scientists would agree with that. i consider myself a scientist.

 

The known physical relationship was inferred from the correlation in the first place. To use it as if it existed before empirical data of the correlation between two things was known is definitely problematic in the matter of justification.

no, the known physical relationship was known from a different empirical test. it is quite easy to test the solubility of CO2 in a liquid outside of the correlation between ocean temperature and CO2 content. put a coke can in your driveway on a sunny day.

 

As for the second part of your statement, you are right that all of science (and much more) suffers from the same problem, and that nothing (that is "matter of fact" knowledge) can be known at all, and that scientific theories and laws would become decidedly invalid and irrational to hold.

 

I don't see what is preposterous about it at all however, except for the fact that it is an unpalatable conclusion for some. Like I said before, this is one of the epistemic problems with science, its methods, and its inferences.

it is one we have to live with. hume is not an end-all be-all, either. even the wiki points out that his conclusions are flawed. whomever wrote the wiki article clearly disagrees with his notions of causality. so do i. apparently you don't. the differences between scientist and philosopher are clearly irreconcilable.

 

Ok, and you are looking at it with the eyes of a scientist (or statistician). This is why I said there is an impasse.

a bit of both, actually. i'm a signal processing engineer (electrical engineering degrees), which necessarily requires a significant background in statistical analysis methods (though we approach problems differently than pure statisticians). i do research and development for my career (and next degree), which is why i consider myself a scientist as well. really, i'm a theoretical engineer that gets to do practical engineering as well.

 

What I am propounding is that there is no empirical evidence for casual links and causality, as we do not perceive causality. Add to that the fact that there is no valid inference from data of correlation to causal links, and you get the basic gist of my argument: there is no reason to believe cause and effect relations exist.

point noted. i simply disagree with this explanation. most of my peers would agree with my position, though i'm sure most of yours would agree with your position.

 

Yes, it is an opinion piece, as is your opinion piece that human don't play a significant part, and natural causes play the most significant part. Now, I don't know about the inclinations of others, but disregarding my skepticism towards science (due again, to epistemic problems with some questionable assumptions it makes), if I were to hold a strong position on the issue, it would be in line with the vast majority of the scientific organizations.

uh, that's not the point i was trying to make. what i meant is that since it is an opinion piece, its estimation of "consensus" is at the very least biased. they choose not to point out all the significant numbers of people that disagree, or have made significant counter claims. i.e., saying that article is truly representative of "the vast majority of scientific organizations" is disingenuous at best (i'm not accusing you here of being disingenuous, rather, the article is).

 

Also, I would say that there is quite a bit of empirical evidence against your assertion given how many scientists let those organizations speak for them (I'll try to bring up a count of signatures of all the scientists who have latched on to the manifesto that global warming has a significant or main cause in human activities, and we'll compare it to the scientists who disagree).

19,000 signed the OPIM petition.

 

your first link is to a report on the IPCC SPM, which is NOT any indicator of any consensus. it was written by only a few people and many of the "consensus" scientists that are included in the group are slowly coming forward voicing their disagreement with many of the SPM conclusions.

 

your second link is more of the same but for the the 3rd IPCC report. indeed, the 3AR was where the hockey stick got promoted heavily, which has now been shown to be the result of spurious correlations as a result of overfitting (among other problems.

 

Both links show that there is wide agreement among scientists that human activities play a significant factor in global warming (on the severity of the problem, there is disagreement).

no, both links demonstrate that the same authors still hold the same views they did in 2001 as they do in 2007. the consensus view of the IPCC is manufactured. the IPCC, in general, does not conduct its own studies, btw. it is a political body, not a scientific body.

 

taks

Edited October 15, 2007 by taks

[taks]

btw, qwerty, given your obvious apprehension at accepting causal relationships resulting from science, i offer you this challenge. look into how some of these conclusions have been derived by the various organizations that are proffering them. many are related to results from climate models, which don't even correlate well with known climatological history. the now infamous hockey stick is based on an unproven assertion that tree-rings are good proxies for temperatures - a correlation that does not even exist in modern times. in fact, an NAS panel convened on the subject stated quite clearly that the primary culprit of the "blade" of the stick was due to the bristlecone pine proxies, and it recommended that they be avoided for reconstructions (yet still they are touted as viable). phil jones, a UK scientist in control of their temperature data has been quoted as saying (paraphrased) "why should i open my data and methods resulting from 25 years of effort to those whose sole purpose is to find flaws"! wow, if you are truly skeptical of science in the manner you suggest, this should be as appalling to you as it is to me. given the flaws of science that i cannot disagree with, if proper procedures are not followed, doesn't that make it even more flawed? the efforts by a few to hide contrary results are numerous, and in many cases, significant (michael mann actually told congress regarding his hockey stick reconstruction, for which he claimed "significance" for r2 in his original paper, that it would be silly to calculate r2 - he did calculate it, and failed to report it was nearly zero for all but the calibration period of the current century. the actual results were recently published in a paper by wahl and amman).

 

taks

[WITHTEETH]

Tibetan monks pray every morning so the that the sun comes up, thank god for them or else i wouldn't have this great tan! I disagree with Tak again.

 

The correlation between cause and effect is a phenomenon that we can't with-out-a-doubt explain using the deductive reasoning taken far back enough. Humes billiard ball example explains this. Our senses, our tools, may not be able to perceive everything, so whos to say we can explain everything...

 

the correlation between cause and effect are conveniences that help us explain the world around us to the best to our ability. We try and try to find new ways to help us perceive new ways to explain phenomenons IE LHC at CERN.

[taks]

The correlation between cause and effect is a phenomenon that we can't with-out-a-doubt explain using the deductive reasoning taken far back enough.

um, sorry, but i made my points pretty clear and i never said there can be a without-a-doubt link between correlation and causation. in fact, i even remarked that "proof" only exists in the mathematical realm. i've always said "implied," except on the issue of causation itself. the burning cotton example is definite cause and effect, in spite of the lofty notions of esoteric philosophers. it is not simply correlation.

 

Humes billiard ball example explains this. Our senses, our tools, may not be able to perceive everything, so whos to say we can explain everything...

based on what i read of humes, which admittedly isn't much, i fundamentally disagree with him. i'd be willing to bet most hard sciences types such as i am would as well.

 

taks

[Gromnir]

"i'd be willing to bet most hard sciences types such as i am would as well."

 

ain't that the truth.

 

HA! Good Fun!

[WITHTEETH]

The correlation between cause and effect is a phenomenon that we can't with-out-a-doubt explain using the deductive reasoning taken far back enough.

um, sorry, but i made my points pretty clear and i never said there can be a without-a-doubt link between correlation and causation. in fact, i even remarked that "proof" only exists in the mathematical realm. i've always said "implied," except on the issue of causation itself. the burning cotton example is definite cause and effect, in spite of the lofty notions of esoteric philosophers. it is not simply correlation.

 

Humes billiard ball example explains this. Our senses, our tools, may not be able to perceive everything, so whos to say we can explain everything...

based on what i read of humes, which admittedly isn't much, i fundamentally disagree with him. i'd be willing to bet most hard sciences types such as i am would as well.

 

taks

lofty notions of esoteric philosophers? A new low Tak, congrats.

[taks]

no, favoring the ideas of someone that seems to think of causation as an invention of the human mind, over hard science, is a joke. that you would support such a concept is no surprise. science doesn't get done by people that sit around and think about what is going on in the universe, it gets done by people that do something about it. i also read hume's billiard example and it is equally incompatible with the real concepts of the real world. folks like you and qwerty can think about how much meaning he imparts on our lives, but in the end, folks like me are using the concepts of causation to do everything it is you do in life. all technology, all science, indeed all progress, is due to science, not unprovable thought problems.

 

taks

[Qwerty the Sir]

i disagree. cause-effect is easily observed, at least at the macro level (quantum causality gets strange).

 

When I watched a friend light a cigarette last night, I saw the event of of the flame from the lighter touch the cigarette, and then I saw the cigarette light on fire. I could say that these two events are closely correlated, I saw one happen right after the other. However, at no point did I actually perceive any sort of causation. If one thinks about it, this makes sense; how the hell does one actually perceive causation? The only thing I saw was that the lighting of the cigarette occurring together (or right after) with the contact of the flame. I never saw the lighting of a cigarette

 

If you really assert that we can perceive causation empirically (and not just construct it mentally), please provide an example where a cause effect link is perceived between two things (better yet, provide a video showing it).

 

this is one of the "truths" we have to accept as existing. your assertion that i need to provide some satisfactory justification of causal relationships is a philosophical nit.

 

When someone makes a claim about a truth, some sort of rational justification usually needs to be provided to make acceptance of that truth rational.

 

few, if any, scientists would agree with that. i consider myself a scientist.

 

it is one we have to live with. hume is not an end-all be-all, either. even the wiki points out that his conclusions are flawed. whomever wrote the wiki article clearly disagrees with his notions of causality. so do i. apparently you don't. the differences between scientist and philosopher are clearly irreconcilable.

 

OK, although if they are going to disagree with the skepticism against causality, you would think they could provide some justification to make their opinions more solid...

 

Nobody is saying that Hume is the end-all. I didn't even want to bring him up by name in the thread, however, his quote about providing justification was a very precise way of summing up the problem posed and bringing up the question of justification. Add to that the fact that it is hard for me to find online texts of Ghazali and Malebranche (both who put forth less ambiguous and more cogent criticisms against causality earlier than Hume), who also bring in God into the equations (occasionalism) further complicating the situation.

 

no, the known physical relationship was known from a different empirical test. it is quite easy to test the solubility of CO2 in a liquid outside of the correlation between ocean temperature and CO2 content. put a coke can in your driveway on a sunny day.

 

That empirical test would have only given a correlation between things. The known physical relationship was once again (invalidly) inferred from that correlation.