A Precedent for Failure of Consensus

Nobody can be expert on everything, and when one is inexpert on a topic at hand, one must judge both the capabilities and the honesty of those presenting as experts.

This is the less sinister (but no less disastrous – see Lou Grinzo’s latest jeremiad) aspect of the world’s intransigence in dealing with the carbon problem.

People don’t believe us because 1) our competencies are not especially visible to them 2) our message seems improbable 3) what we say reminds them of superstitious, excessive, innumerate apocalyptic green radicalism (and is embraced by green radicals) and 4) what we recommend is spending on specific expensive projects that would be otherwise unnecessary or premature. This is even before the professional liars come in to muddy the waters.

Against this we have, only, a genuine scientific consensus to show. And that showing can easily be weakened by a concerted effort by fringe characters, as we have seen.

But never mind that. Most will concede that the field, as represented by IPCC, rightly or wrongly is solidly behind a range of opinion that, if correct, makes current policy toward carbon emissions flatly insane.

So how should a skeptical person who does not have the inclination to look into the details evaluate the consensus?

One approach to this is to ask whether scientific consensus opinions still fail spectacularly (in the way phlogiston or the luminiferous ether did)?

To this, one can rejoin that we are not discussing the validity of a hypothesis. The “global warming hypothesis” is not in the same category as phlogiston or luminiferous ether. We are not asking about the mechanism of an observed phenomenon. We are not asking whether some process is or is not a part of nature. Rather, we are asking about the amplitude, the extent or the amount of a generally agreed phenomenon. This quantity is often boiled down to the equilibrium sensitivity of the earth’s mean surface temperature to a doubling of CO2; the infamous “sensitivity”. It is not a perfect boiling down of the problem but it is a very useful one.

The IPCC-reported consensus has S in the range 2.0 to 4.5, with a best bet of 2.7 or so. This is enough to yield expectations of enormously disruptive changes in the world’s climate in the lifetimes of people now living.

Some of the so-called skeptics argue instead for something in the range of 1.0 to 1.5, failing to realize that such a sensitivity doesn’t, once you parse it through the whole problem, cut us very much slack. Really you need a number as small as 0.5, a quarter of the consensus low-end, less than a fifth of the best estimate, before you can make a strong case for a delayed response. And what’s more, you need it to be that small with high confidence.

So what are the chances that the consensus is so badly wrong, by a factor of 5, in estimating an important quantity that everyone agrees must have some reasonably well-defined numerical value?

It seems that to address this, one might want to look through the history of science for cases where a crucial quantity had an accepted range that turned out to be badly wrong. I’ve recently come across one, and the story is instructive.

In the fascinating if peculiarly titled pop science book The Feathered Onion by Clive Trotman, the age of the earth plays a major role. Trotman begins his argument, then, by giving a history of estimates of the age of the earth, which have now converged at 4.6 billion years. He observes that nevertheless, for decades, a consensus of 0.1 billion to 0.2 billion prevailed as a firm scientific consensus. This was well into the period of modern science (from about 1860 to 1900). This is an error of a factor of more than twenty at the least. How could such a crucial number be so widely held for so long at a value that was so badly wrong?

At least two key reasons are available.

One factor is the commonly remarked-upon herd mentality. The scientific community is necessarily heirarchical, and approval from proven past masters is a key to success. And so, as it happened, the greatly esteemed and powerful Lord Kelvin had made an argument that the sun could not possibly be much over 0.1 billion years old.

Challenging Kelvin in his day was not something an aspiring scientist would do without not only solid convictions but plenty of courage to go along with it. This may have slightly distorted conversation on this particular matter, but on the whole scientific renown serves a number of useful purposes in the progress of science. Kelvin had been a major contributor to science in many ways, and his reasoning on the question was regarded as sound: there was no way known to 19th century physics a body the size of the sun could muster the energy to burn much longer than the hundred million years Kelvin favored, though the geological evidence was enough to convince him that the earth was ancient compared to traditional estimates.

The second factor, though, is that intuition is influenced by such tradition.

The discussions of the age of the earth only became plausible after Darwin. Prior to that, the consensus was that set by biblical scholars, of a few thousand years. (Any orthodox Jew, to this day, will happily and without hesitation give the exact age of the earth according to the rabbinical reckoning, which is a bit less than 5773 years at the time I write this. Christian biblical scholars have come up with numbers of the same order, if a bit higher.) And any discussion of the matter happened in a broader social context wherein a world only a few thousand years old was taken for granted.

So talk of a hundred million years was considered shocking and outre in its day. Scientists as much as anyone else would be fighting their own intuitive sense of a 6000 year old world. A hundred million seemed too much to ask for already. Fancy talk about billions of years!

Scientists, being human, will err on the side of intuition, and having done so, will cluster together until new evidence changes the picture. Of course, we are amenable to new evidence. That’s the good news. But we can be blind to it if we are comfortable with the herd, and especially if the new evidence is even further from the intuitive sense of things.

And we are still fighting an intuitive sense based on our historical context, that the world is much larger than humanity. That’s totally wrong now, at least insofar as its surface processes go. The earth’s surface is a wholly owned subsidiary of the human enterprise. But most of us are still awestruck by the size and, well, awesomeness, of the world.

That is, our consensus estimate of 2.7 for climate sensitivity is a battle between the evidence at hand and our own innate cultural intuitions, to which as scientists we are not immune.

Not to mention our desperate wish that the whole problem would magically go away, along with McIntyre, Morano, Cuccinelli, Horner, Michaels, Singer, all following Joe Bast into obscurity and perhaps unemployment. Leaving us scientists with the pretty set of questions and the peaceful, modestly comfortable working conditions that pulled us into science in the first place. Somewhere over the rainbow.

Our opponents think we are motivated to overstate the problem, but that is one of the most irritating of their misunderstandings. We are really intensely motivated to have it go away.

So we have our own wishful thinking, as well as the general run of the mill wishful thinking, and our own perceptual biases, as well as the run of the mill perceptual biases.

And this leaves me wondering if we haven’t actually underestimated the climate problem in some way. Now, let me be clear, I’m not actually arguing for high S. In fact, I suspect there are other ways we might have underestimated the problem, and for those of you in America or the UK, the events of the last few weeks should give you a clear idea of what I mean by that.

But the opposition are really arguing (whether they realize it or not) that our skills and trustworthiness are such that our estimate of the key quantity could be fivefold in error. The prospect that they are right about that terrifies me, because if we are fivefold wrong, it’s as likely, no, perhaps more likely, to be an underestimate than an overestimate.

Just like the 19th century scientists who couldn’t say “well, 100 million years is really a lower bound, eh wot?” we can’t even contemplate an equilibrium sensitivity of 13 or 14C, comparably wrong.

Fortunately, we actually do have a better handle on it than that. It is bad, but it isn’t that bad.

But it’s at least as likely to be that bad as it is to be as sunny as the naysayers believe. Perhaps more so.

My point is this. Just as uncertainty is not our friend, the herd mentality is not our friend either. The herd will always split the difference between the evidence and what they want to believe. Consequently, we may be worse off that we think.

Image: William Thomson, Lord Kelvin, striking a rather Pierrehumbertesque pose


  1. > what we recommend is spending on specific expensive projects

    I've attempted to argue elsewhere that this is a mistake. You shouldn't be recommending specific solutions; not just because you have no expertise in that area, but also because people react badly to it.

  2. Mitigation is necessary, and will involve specific expensive projects, but it's not necessary to name them and Michael didn't.

    That said, one might call halting all coal combustion (in the presumed absence of workable capture and storage technology) a project. If so, scientists shouldn't feel constrained from pointing out the bleedin' obvious.

    I would agree that it's a different matter when we start talking about projects that might not be necessary, or that are part of a range of possible project choices pertaining to a given policy goal.

  3. Hasn't climate science already gone through its "age of the Earth" stage in (a) acceptance that it isn't always in some steady state and (b) that additional CO2 is warming the system an appreciable amount? It's been a while since I read Spencer Weart's stuff, but I seem to remember that there was quite a lot of kick-back against those ideas for a while.

    Okay, so the number's not as tightly constrained, but it's a younger discipline, and a different sort of quantity. It may be that the history of the Hubble Constant (or parameter) may be a good comparison to the history of S?

  4. While in general I think we do disagree about how much deference to give to credentialed expertise, I don't think this is a good case in point for you to raise it.

    There is a perception that "climate science" is making specific expensive recommendations, whereas it is the larger community of people who take the results of climate science seriously. The skeptics out there are not alone in having trouble making the distinction, and to some extent the formulation of IPCC blurs it.

    I don't think climate scientists are excluded from having opinions about which technologies are most promising or which policies are most likely to enhance the selection of viable options. And here perhaps we have some disagreement. But even if we all scrupulously avoided such conversation we would be perceived as being part of a concerted effort to expend resources on certain technologies, in particular energy-related ones, and it is that perception that is my point.

  5. It's important to note that Lord Kelvin did not just pull his figures out of a hat; his hypothesis was based on scrupulous calculations employing thermodynamics employing the best available assumptions about the evolution of the solar system then available. Naturally there was much he and all of his contemporaries did not know, so much so that his conclusions were wildly inaccurate.

    Others performed independent calculations of Earth's age employing other lines of speculation, producing results consistent with Lord Kelvin's numbers. Necessarily these workers also had no notion of nuclear physics and a number of other important "details" thus leaving their results equally incomplete and wrong.

    All of these folks were employing all the best tools contemporaneously available; it does not really appear that argumentation around the matter hinged on silverback gorilla political dynamics. However if we want to entertain the "herd" idea, Huxley was not silent on the matter, though his arguments were basically handwaving against detailed and calculated lines of reasoning; it seems reasonable to suppose lesser primates in the hierarchy would feel emboldened to speak up with Huxley taking the lead.

  6. Speaking of a different consensus, see here. I've always assumed that the self-conscious tacticians on the denialist side would break away as soon as public opinion began to shift as a consequence of gathering climate change impacts, their idea being to engineer as soft a landing as possible for their funders and more broadly for the Republican Party, but is this conference truly such a first crack in the denial-o-sphere?

    Note also the just-published poll results appearing to show a shift in public sentiment, and it was conducted before the recent heat wave and IIRC at least some of the fires (so is perhaps somewhat a response to the March event?). Presumably outfits like AEI, having the wherewithal for their own polling and focus groups, are ahead of the curve on such things.

    If true, this would make AEI the worst sort of opportunists, but I suppose we knew that already.

    I'm still digesting this news and will have some more thoughts, but IMHO it's worth a separate post.

  7. I hope we all understand that the Charney climate sensitivity involves only the so-called fast feedbacks. Including everything is called the Earth System Climate Sensitivity and that is roughly esitmated to be at least 4.5 degrees Celcius by using some Pliocene data.

    On another matterr, James Hutton is considered to be the father of modern geology, well before Darwin learned his geology from reading Charles Lyell's books, those being heavily influenced by Hutton's work. The geologists at that time did not have a means of determining just how very slow geological processes are, but generally came to the concensus that at least millions of years were required. NOt shabby alhough too small by a factor of a thousand or so.

  8. Further to David A. Benson's comment, the age of the earth was already under discussion before Hutton. The article is wrong in stating that "The discussions of the age of the earth only became plausible after Darwin. Prior to that, the consensus was that set by biblical scholars, of a few thousand years."

    In 1779 Saussure wrote that it was universally accepted (among savants) that the earth had undergone a "long succession of ages", around then Werner's lecture notes suggested that the period might be a million years, and in 1793 Lavoisier suggested a timescale running into millions of years. This was in line with Kant's 1755 suggestion that the cosmos had developed over "a series of millions of years and centuries". Hutton's proposals for an indefinitely old earth, continually recycling, went beyond these proposals.

    By the time of Lyell an ancient Earth was well accepted in geology, with his "catastrophist" opponents proposing human civilisation within a Biblical time frame but preceded by multiple geological ages extending over millions of years. The thermal calculations by Thomson (Lord Kelvin) conflicted with the ideas of geologists at the time, suggesting a shorter time frame than they expected.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.