The simplest explanation is usually the correct explanation
Guest essay by Eric Worrall
Professor Keven Trenberth once campaigned for the scientific world to accept the alarmist view of climate change as the “null hypothesis”, the baseline theory against which all other theories must be measured.
http://wattsupwiththat.com/2011/11/03/trenberth-null-and-void/
The reason Trenberth faced an uphill battle to have his view accepted, and ultimately failed, is that the simplest explanation of contemporary climate change does not involve Anthropogenic CO2.
As Professor Phil Jones of the CRU once admitted in an interview with the BBC, the instrumental record contains periods of warming which are statistically indistinguishable from the 1990s warming – periods of warming which cannot have been driven by anthropogenic CO2, because they occurred before humans had made a significant changes to global CO2 levels.
Between 1860 and 1880, the world warmed for 21 years, at a similar rate to the 24 year period of warming which occurred between 1975 and 1998. There was simply not enough anthropogenic CO2 in the atmosphere to have driven the 1860s warming, so it must have been driven by natural variation.
So how does Occam’s Razor apply to this observation?
According to the definition in Wikipedia, the principle of Occam’s Razor states “that among competing hypotheses, the one with the fewest assumptions should be selected. Other, more complicated solutions may ultimately prove correct, but—in the absence of certainty—the fewer assumptions that are made, the better.”
From Wikipedia, the reason why Occam’s razor is important:
“To understand why, consider that, for each accepted explanation of a phenomenon, there is always an infinite number of possible, more complex, and ultimately incorrect alternatives. This is so because one can always burden failing explanations with ad hoc hypothesis. Ad hoc hypotheses are justifications that prevent theories from being falsified. Even other empirical criteria like consilience can never truly eliminate such explanations as competition. Each true explanation, then, may have had many alternatives that were simpler and false, but also an infinite number of alternatives that were more complex and false. However, if an alternate ad hoc hypothesis were indeed justifiable, its implicit conclusions would be empirically verifiable. On a commonly accepted repeatability principle, these alternate theories have never been observed and continue to not be observed. In addition, we do not say an explanation is true if it has not withstood this principle.
Put another way, any new, and even more complex theory can still possibly be true. For example: If an individual makes supernatural claims that Leprechauns were responsible for breaking a vase, the simpler explanation would be that he is mistaken, but ongoing ad hoc justifications (e.g. “And, that’s not me on film, they tampered with that too”) successfully prevent outright falsification. This endless supply of elaborate competing explanations, called saving hypotheses, cannot be ruled out—but by using Occam’s Razor.”
Source: http://en.wikipedia.org/wiki/Occam’s_razor
In other words, if we reject the principle of Occam’s Razor, we open the door to accepting theories of arbitrary, ultimately infinite complexity. A theory created by researchers who do not accept the principle of Occam’s Razor cannot be falsified, because the theory can always be tweaked in arbitrary ways to avoid falsification.
So why does applying the principle of Occam’s Razor force us to reject the theory that anthropogenic CO2 is the main driver of contemporary climate change? The reason is that nature has produced periods of warming similar to the recent warming, without any significant contribution from Anthropogenic CO2.
So we have two competing hypothesis for what is driving contemporary climate change:-
1. Observed natural variation, which has produced periods of warming statistically indistinguishable from the warming which ended in 1998.
2. Observed natural variation + an unproven assumption that Anthropogenic CO2 is now the main driver of Climate Change.
Clearly the second hypothesis fails the test of Occam’s Razor. In the absence of compelling evidence that anthropogenic CO2 has overridden natural variation, we have to accept hypothesis 1 – that observed climate change is the result of natural variation.
The climate is not hotter than it was in the past, periods such as the Holocene Optimum, or looking further back, the Eemian Interglacial. The warming which ended in 1998 was not faster, or of significantly longer duration, than similar natural warmings which occurred in the recent past.
Nothing about the current climate is outside the bounds of climatic conditions which could reasonably be produced by natural variation – therefore, according to the rules of science, we have to reject hypothesis which unnecessarily embrace additional unproven assumptions, unless or until such assumptions can be tested and verified, in a way which falsifies the theory that natural variation is still in the driver’s seat.
LGL, it’s not Richard’s fault it remains unexplained, go speak to the those who have blown the billions of public money doing activism instead of science.
Science advances by observation, analysis, hypothesis. Because of the last three decades been largely wasted, we are still at step 2 on this this one. That does not count as a rebuttal of the observation of an obvious pattern, that requires explanation.
All we have for our money so far is : temperature has been generally rising for 300 years. CO2 has been significant for 60 years. Therefore, most of the warming of the latter half of the 20th c. is due to human emissions.
That also is just a description, not an explanation imo.
It’s so pathetic it defies parody.
Temperatures in February 1878 were exactly the same as February 2014.
Tom says:”Occam’s Razor is often used incorrectly by those who do not understand it or by those who know better and are trying to justify something that isn’t true. ”
Indeed.
Cyclical leprechauns. That is all.
Bill Illis says:
March 22, 2014 at 6:24 am
Temperatures in February 1878 were exactly the same as February 2014.
Assuming you aren’t using satellite data (grin), again using the latest month and going back as far as possible to show a flat trend, it appears there is some more research to be done.
Have our CO2 emissions helped get us back to 1878 temps?
Just wondering.
The Warmists/Alarmists have made Occam’s Razor into Gillette’s Razor with 5 blades and flex head. It’ll do anything you want…… ;^D
You can look at the long term temperature in relation to the PDO by mapping linear PDO phases against the data.
http://tinyurl.com/kzmzd8y
As is pretty obvious the trend flips for each change in the phase of the PDO. This is another way of looking at the ~60 year cycle. Many will argue the residual warming is due mainly to biased adjustments. Probably somewhat true, but there’s another explanation as well.
The PDO releases heat unevenly during its two distinct phases. During its positive phase it is far more likely to release energy to the atmosphere than during its negative phase. The heat itself comes from solar variations over time. The warming won’t necessarily line up with the Sun’s cycles because the heat is buffered in the oceans and released more strongly during +PDO phases.
The Solar + PDO view explains the warming and cooling cycles since 1850. The warming of the early 20th century combined a +PDO with a more active Sun. Even though the Sun was still quite active in the 1950s the PDO has already changed modes. The extra energy wasn’t released until the PDO became positive in the 1970-2000s.
What we should see in the next 20 years with the -PDO and weak Sun is cooling.
Reblogged this on Sierra Foothill Commentary and commented:
The problem is that our political leaders in Sacramento do not believe in Occam’s Razor, only in the myth of anthropogenic warming.
lgl says:
March 22, 2014 at 5:43 am
“Richard: It wouldn’t, but natural variation is just a description, not an explanation imo. You have to explain how.”
Hmm. Well as that is just one of the whole series of natural variability, then, unless the others can also be explained, I am not sure how and why you would single out just that one.
Or can you explain the 2, 3, 4, 12 and 60 years ones?
Occam was a 14th c Christian scientist, an era populated by geniuses who built modern math and sciences, yet we are told this was a ‘dark age’, even though we still use ‘laws’ developed during this time [see impetus, average velocity, gravity etc]. I would say that today is the ‘dark age’ when something as insidious, stupid and non-scientific as globlaoneywarming is ‘consensus’. Reminds me of abiogenesis, a ‘law’ which was ‘consensus’ for 2100 years and still believed in by evolutionists…..whoops.
Tom in Florida says:
March 22, 2014 at 6:24 am
“Occam’s Razor is often used incorrectly by those who do not understand it or by those who know better and are trying to justify something that isn’t true. Incorrect usage of the principle immediately sets off my sceptic alarm.”
So there is acknowledged natural variability in 2, 3, 5, 12 and ~60 year ‘cycles’ but the greater than 75 years rise is definitely down to CO2?
What would you conclude is the Occam’s Razor statement for that?
michel says:
March 22, 2014 at 1:52 am
What’s really needed is to establish the distribution of warming episodes and absolute temps, and then see if we are one or two sds north of the mean. Or if its not a normal distribution, where we are. Anyone know any study like that?
+++++++++++++++++++++++++++++++++++++++++++++++++++++
Well, gee Michel, how far back in time do you want to do this analysis? 35 years (the satellite record era)? 135 years (the thermometer era)? Longer than 135 years – the cherry-picked tree-ring/Weegie-board era?
If warming/cooling episodes are approximately 30 years long and you assume moderately useful statistical analysis generally requires a minimum of 30 data points, you’re dealing with a minimum of roughly 900-1000 years of cherry-picked tree-ring/Weegie-board data.
The 3 biggest problems I see with the analysis are:
(1) the last 1000 years of existing data and its accuracy are highly suspect;
(2) most “climate scientists” (some of whom even have degrees in science) could care less about actual data;
(3) at no point in time has anybody comprehensively defined the underlying atmospheric physics in a way that can be tested with real, actual data.
All the above not withstanding, I’m sure somebody could punch you up a model that would spit this out in no time at all.
Occam’s Razor is not necessarily the shortest way to the truth.
It is much more pragmatic.
It means that If competing hypotheses are equally valid you should use the simplest one because it is the simplest.
Using the simplest hypothesis is tidier….and is probably, but not necessarily, more correct
RichardLH says:
March 22, 2014 at 6:58 am
re: Tom in Florida says:
March 22, 2014 at 6:24 am
“Occam’s Razor is often used incorrectly by those who do not understand it or by those who know better and are trying to justify something that isn’t true. Incorrect usage of the principle immediately sets off my sceptic alarm.”
So there is acknowledged natural variability in 2, 3, 5, 12 and ~60 year ‘cycles’ but the greater than 75 years rise is definitely down to CO2?
What would you conclude is the Occam’s Razor statement for that?
========================================================================
I wasn’t making a comment on the article but rather a general comment about the incorrect use of Occam’s Razor, to wit: “The simplest explanation is usually the correct explanation” at the start of the article. Just one of my many pet peeves.
lgl says: March 22, 2014 at 2:39 am
Fine, but what is the ‘simplest explanation’ for the 1880-1998 rise?
It is: We don’t know.
The null hypothesis is that anthropogenic C02, comprising an negligible .0017% of the atmosphere, has a negligible effect on the earth’s climate. Extraordinary claims require extraordinary evidence, especially when the proposed mitigation is de-industrialization (ever notice that progressives are anti-progress?) and a massive lowering of civilization’s standard of living. The claim that the flattening of global temperature for the past decade and a half is due to “natural variation” (as Trenberth remarked, a fancy term for “we don’t know”) is actually a counter-argument to the AWG hypothesis. A natural forcing strong enough to temporarily reduce the trend in place since at least the 1850s is also strong enough to explain the enhanced trend during the warming scare of 80’s and early 90’s. Science must refute this simple, natural explanation with proof of human attribution without relying on climate models which have proved themselves to be without predictive power.
lgl says:
March 22, 2014 at 2:39 am
“Fine, but what is the ‘simplest explanation’ for the 1880-1998 rise?”
Solar insolation and ocean dynamics. The ocean is a massive heat storage reservoir, an obvious fact that seems to have been lost on climatologists until recently recognized while floundering about in search of the “missing heat”. As such, the amount of energy contained in the ocean is sensitive not just to the TSI, but to asymmetry in the TSI. If more heat is added in the warming part of the cycle than is released in the cooling part, the next cycle starts from a higher baseline. Taking this integrative effect in to account and adding in the observed AMO/PDO variations accounts for 94% of the observational record. (see this plot and it’s derivation here.
Of course Occam’s Razor cuts both ways. Thousands of independent scientific researchers, multiple lines of physical evidence past and present, and simple physics understood for a century and a half vs. a relative handful of cranks egged on by ideologues and funded by a few companies with strong vested financial interests.
Watch that blade, lads. It’s sharp.
Occam’s Razor isn’t an absolute rule of reason, because sometimes the actual explanation is, in fact, the more complex of two hypotheses. That is, nature itself is not obligated to be “simple”.
The correct (mathematical) way to approach this is from the point of view of Bayesian statistics. The advantage of a simple hypothesis is that it has fewer Bayesian priors. This means that it is (usually) more probable because the Bayesian priors occur in conjunction — if every assumption underlying a hypothesis comes with a probability of being wrong and only one assumption has to be wrong to make the hypothesis untrue, then fewer assumptions (all with an equal chance of being wrong) is better than more assumptions.
However, all assumptions do not have an equal chance of being wrong. A hypothesis that depended on, say, the correctness of Newton’s Law of gravitation for simple macroscopic objects in non-relativistic context and one other 50-50 assumption is not automatically less likely than a competing hypothesis that depends on only one 50-50 assumption, because the law of gravitation is probably correct out in the 0.9999….. range.
Bayesian scientific reasoning (as explained in Jaynes’ classic work Probability Theory, the Logic of Science) deals with Occam’s razor not in verbal terms involving leprechauns, but in terms of using incoming data to recompute the prior probabilities. Even if we truly believed strongly in gravity as a prior probability close to unity, evidence is capable of causing us to reduce that probability as experimental data inconsistent with the theory accumulates, and eventually falsify (in the sense of “make very very improbable”) not just the hypothesis in question, but its prior assumptions.
The classic example of this is you are given an electronic coin flip program that uses a perfect (by assumption) random number generator to produce random 0’s and 1’s with some unknown probability. You use that generator to drive a game, perform a computation, or just do some counting that requires 0’s and 1’s to occur randomly with equal probability. The game/counting is your hypothesis — “If the generator produces 0’s and 1’s with equal probability, then in the long run my counts will be distributed according to a binomial distribution with p = 0.5”. The assumption is clearly evident.
Bayesian reasoning gives you a way of starting with a prior assumption of probability of p = 0.5 and effectively weight it with the strength of your belief that it is true, and then gradually modify it as you generate a string of 0’s and 1’s in such a way that it converges on the frequentist definition of probability in the limit of infinite trials but differs in the beginning where one has a strong bias towards an unbiased coin flip, as it were. Maximum entropy rocks and posterior probability computation rocks as well.
A second objection to the assertion that no-CO_2 hypotheses are “simpler” than CO_2 hypotheses is that a data with a known, clearly evident degree of complexity has to be explained by a hypothesis at least sufficient of producing that degree of complexity. It obviously cannot be explained by a theory that cannot produce that degree of complexity, but there is nothing like a uniqueness theorem or minimum complexity theorem for general scientific hypotheses.
For example, if one’s data exhibit curvature, one cannot use a hypothesis that only produces a linear response to explain it. One has to have a theory at LEAST complex enough to produce/predict data with curvature, a quadratic theory or higher. But one cannot easily specify that a quadratic theory is better than an exponential theory in spite of the fact that quadratic specifies only three terms in a Taylor series where an exponential specifies an infinite number of terms. Lots of physical phenomena are exponential. Or sinusoidal, ditto. So Occam’s Razor cannot be naively applied.
In the particular case of climate science, nobody (sane) would claim that CO_2 concentration is likely to have zero effect on the climate, because a lot of physics with prior probability (determined by independent theory and experiment) very, very close to unity suggests that it will. The problem is, in fact, the opposite one to the one the top article suggests. CO_2 increase should produce a locally linear (globally logarithmic) response in terms of average temperature from some very simple physics. This is the simplest possible a priori assumption, one that ignores everything else BUT CO_2 and its direct effect on the climate. This is why, in fact, climate scientists hold so adamantly to their conclusions about anthropogenic global warming.
However, the climate is not simple, and not linear. The big question is, how well does the available data support the (very strongly held, initially) prior assumption of a simple linear independent effect for CO_2?
That’s what the data above pertain to. They suggest that the explanation for the 1983-1998 temperature increase is not CO_2 not because it isn’t a simple explanation, but because as the data clearly indicate that the problem has more complexity, that the climate is a function of more than one thing, and that the other things are capable of producing effects as large as the one attributed to CO_2 in the particular set of years indicated in circumstances where CO_2 is not a possible explanatory factor.
That is, we need a more complex theory, not a simpler one!
A vastly more complex theory, because we know (again, based on sound, strongly believed prior assumptions) that the climate is determined by the solution to a dynamical system we can conceptually define at the microscopic level but that is almost incomprehensibly complex at the macroscopic level — a set of coupled Navier-Stokes systems driven by a highly erratic set of forcings and with multiple coupled energy reservoirs with an array of pertinent time constants. Frankly, the notion that the system is linearizable in a single parameter, or a single parameter plus simple feedbacks, is and always has been absurd. We actually know the physics here, and it is not at all simple! It is arguably the most complex problem we’ve ever tried to solve. Occam’s Razor is the exact opposite of the thing needed, but our attempts to date to solve the physics problem in GCMs have met with empirical failure at almost every level it is possible to fail at. The best that can be said of them is that they produce something that looks like climate evolution, while having almost zero predictive value. It isn’t even clear that the existing models aren’t divergent where the climate is empirically enormously stable (or rather, broadly bistable or multistable with two primary attractors, a warm phase interglacial and cold phase glacial in the current orbital-geological configuration).
So sorry, I appreciate what you are trying to say and agree that the data call into question the assertion that CO_2 is solely or even mostly responsible for the single 15 year long stretch of warming visible in the latter half of the 20th century, but rejecting the CO_2 hypothesis isn’t a matter of Occam’s Razor, it is a matter of the data suggesting that the solution is a lot more complex than CO_2 alone can explain.
rgb
Richard
Or can you explain the 2, 3, 4, 12 and 60 years ones?
The ~3 and ~60 is ENSO, more or less trade winds, upwelling of cold water and more or less clouds in the tropics. I haven’t heard about a 12 years cycle, where do you have that from?
As already stated, Occam’s razor is a statistical generality. It does not apply to any specific problem. Occam’s razor will not reveal the location of lost aircraft. That does not mean that it is useless, such as the behavior of criminals being analyzed by detectives. Engineering uses “simpler” solutions to problems all the time.
As for “natural variations” being the cause of changes in atmospheric CO2, there are some people who know that the atmosphere is of biological origin. The atmosphere has evolved, and continues to evolve as a part of the global biogeochemical carbon cycle.
Atmospheric CO2 follows biological respiration at all time scales, with some lags due to pool sizes, ocean turnover, etc.
On longer time scales, abiotic factors also apply, such as plate tectonic changes to ocean currents and milankovich oscillations of global orbit. There may be some solar magnetic influences on biology along, etc.
Fossil fuel burning has increased the global carbon cycle by about 3 percent, which is beneficial to the ecosystems and has zero effect on Earth’s temperature.
Erric
Good effort. Please refine.
Define your terms.
“Climate change is a significant and lasting change in the statistical distribution of weather patterns over periods ranging from decades to millions of years” Wikipedia
“Climate change” is being used as an equivocation for “majority anthropogenic driven global warming”.
“Climate change” was politically redefined by the UNFCCC as
State simple hypotheses:
As stated, the hypothesis begs the question:
2. Observed natural variation + an unproven assumption that Anthropogenic CO2 is now the main driver of Climate Change.
Rephrase as:
1. Natural variation is the main driver of climate change with minor anthropogenic affects.
2. Anthropogenic CO2 is the main driver of climate change on minor natural variation.
Recommend referring to Einstein’s Razor
rgbatduke says, March 22, 2014 at 7:42 am:
“In the particular case of climate science, nobody (sane) would claim that CO_2 concentration is likely to have zero effect on the climate, because a lot of physics with prior probability (determined by independent theory and experiment) very, very close to unity suggests that it will.”
Wow. That’s quite an (implicit) accusation! So you actually have to be ‘insane’ to claim that ‘CO_2 concentration is likely to have zero effect on the climate’?!
Magma says:
March 22, 2014 at 7:34 am
Of course Occam’s Razor cuts both ways. Thousands of independent scientific researchers, multiple lines of physical evidence past and present, and simple physics understood for a century and a half.
Perhaps you’d be kind enough to point me to the evidence which clearly demonstrates the attribution of the observed warming to human activity and does not rely on computer models to do so. I’d be most interested in seeing it.
GregK says:
March 22, 2014 at 7:07 am
Occam’s Razor is not necessarily the shortest way to the truth.
It is much more pragmatic.
It means that If competing hypotheses are equally valid you should use the simplest one because it is the simplest.
Using the simplest hypothesis is tidier….and is probably, but not necessarily, more correct
In science There is no Truth, only a (hopefully) ever improving model of reality. Occam’s Razor is itself a model of reality based on an empirical observation of the scientific process. History has shown that the overturning of scientific consensus often results in a new theory that is simpler and more elegant than the old. This observation is the motivating impetus for the belief that the current Standard Model in physics is incomplete and will eventually be overturned by a more elegant “theory of everything”.
Like the scientific bedrock philosophy of falsification, Occam’s Razor says more about incorrectness than it does about correctness. We rule out unnecessary complexity because it makes our experimentation simpler. IF some facet of that complexity becomes an absolute requirement to explain a thing, we accept it tentatively, awaiting that single instance that falsifies it forever. A long period of non-falsification increases our confidence.
Geologic history “proves” that the Earth undergoes very large natural variation. Major periods of natural cooling and warming prove that Earth is at a very low risk of a Venusian-style warm tipping point. Facts that rule over all else: a biosphere that eats carbon dioxide; water vapor that coalesces into clouds that reflect sunlight cause rain; ice cores that show higher atmospheric CO2 concentrations after climatic warming, not before. If one says these things are not true or that they are overwhelmed by human emissions of CO2 you need some very compelling evidence. Correlates that do not imply causation are not compelling. Already falsified computer models are not compelling. The burden of proof, is, as Carl Sagen once said, enormous.