Changing your mind

The grave of Isaac Newton. Newton made himself vulnerable, wrote up his mechanics and optics very clearly. So that it could be falsified and was falsified. While he never repented his errors in public on facebook, his erroneous work did science a great service.

Being a scientist is hard. No, not for the mathematics. Being a scientist is nothing for wimps because you have to make yourself vulnerable, write up your ideas so clearly that someone else can show them to be wrong. That is the background of Karl Popper's requirement that scientific hypothesis should be falsifiable.

Being shown to be wrong is not nice. That could be one of the reasons why science is a very recent invention, it does not come to humans easily.

Being shown wrong is central, however, to scientific progress.

Being trivially wrong, of the kind displayed on WUWT every day, does not help science and is very bad for your scientific reputation.

Being wrong in an interesting way is scientific progress. The moment someone shows why you were interestingly wrong, we understand the problem a little better. If you can make your colleagues think and understand things better, that is good. Like Albert Einstein who was wrong about quantum mechanics until he died, but his challenges were interested and enhance his reputation.

Because it is hard to make yourself vulnerable, because it is good for science when you do so, because it is good for science when others show you wrong, the scientific community has a culture that encourages people to be bold and allows them to be wrong in an interesting way. If scientists would receive a hostile fit every time someone shows them wrong, that would not encourage them much to contribute to scientific understanding.

language

It starts with the language. Instead of saying someone was wrong, scientists talk about "progressive insight". Actually, it is a bit more than just language, often scientists really could not have known that they were wrong because some experiment or observation had not been made yet. Often this observation was inspired by their ideas. Often "progressive insight" also an argument that was overlooked and could theoretically have been thought up before. You could call that being wrong, but "progressive insight" is much more helpful.

Such a change of language is apparently also important in eduction. A high school teacher wrote:

One of the biggest challenges is convincing teenagers that they can be wrong about anything, let alone that there's value in it, ... Therefore, a few weeks ago I assigned my 90 or so Freshmen the task of keeping a 9-week Surprise Journal. ...

I've noticed something, well, surprising. In the class culture, acknowledgement that you are mistaken about something has become dubbed a "moment of surprise" (followed by a student scrambling to retrieve their Journal to record it). As this is much more value neutral than "I screwed up," the atmosphere surrounding the topic is less stressful than in previous years (I suspect -- based on anecdotal evidence -- that they have a history of being punished for mistakes and so are justifiably skittish about the whole topic). That by itself makes me inclined to judge this experiment a success, and has already added richness to our subsequent activities and discussions...

[[Nassim Taleb]], who wrote The Black Swan, mentioned that it is easier for Arabs to admit they do not know something, which is much better than pretending you do, because they express this by saying: only God knows. Much more neutral. (It is highly recommended for anyone in the climate "debate" to read one of his books on new and unpredictable risks and human tendencies to be blind to chance.)

This amicable scientific culture is also the reason that scientific articles are clear on the work performed, but very subtle in pointing our problems with previous studies, to the extent that sometimes only an expert can even notice that a previous study was criticized.

A less subtle example of trying to be amicable would be the article that made me "famous". A study comparing the skill of almost all homogenization method used to remove non-climatic changes. In the tables you can find the numbers showing which method performed how well. In the text we only recommended the application of the five best methods and did not write much about the less good ones. The reader is smart enough to make that connection himself.

At conferences I do explicitly state that some methods should better not be used any more. In private conversions I may be even more blunt. I feel this is a good general rule in life: praise people in writing, so that they can read it over and over again, criticize people in person, so that is can be forgotten again and you can judge whether the message had the right impact and is well understood. Easier said than done; the other way around is easier.

The bar between a stupid error and an interesting one goes up during the process. In a brainstorm with your close colleagues, you should be able to mention any idea, a stupid idea may inspire someone else to find a better one. At an internal talk about a new idea, it is okay if some peers or seniors notice the problem. If all peers or too many juniors see the problem, your reputation may suffer.

Similarly, if all people at a workshop see the problem, your reputation takes a hit, but if only one sees it. If someone from another field, who has knowledge that you do not, that should not be a problem; that is what conferences are for, to get feedback from people with other backgrounds. Only few people heard it, it will normally not be sufficiently important to talk about that with people who were not present and the mistake will be quickly forgotten. Like it should, if you want to encourage scientists taking risks.

There is a limit to this amicable culture. Making trivial mistakes hurts your reputation. Making again them after being been shown wrong hurts your reputation even more. Being regularly trivially wrong is where the fun stops. There is no excuse for doing bad science. Quality matters and do not let any mitigation skeptics tell you that they are attacked for the political implications of their work, it is the quality. I am very critical of homogenization and doubt studies on changes in extremes from daily station data, but I have good arguments for it and being skeptical certainly did not hurt me, on the contrary.

The difference between writing text, fixed for eternity, open for critique forever, and spoken language is also the reason why I am not too enthusiastic about tweeting conferences. On the other hand, to reduce travel it would be a nice if the talks of conferences were broadcasted over the internet. To reduce the tension between these two requirements maybe we could make something like [[snapchat]] for scientific talks. You can only look at it once (in the week of the conference) and then the video is deleted forever.

The above is the ideal case. Scientists are also humans. Especially when people keep on making similar wrong claims, when you can no longer speak about mistakes, but really about problems, also scientists can become less friendly. One case where I could understand this well was during a conference about long range dependence. This term is defined in such a vague way that you can not show that a signal does not have it. Thus a colleague asked whether a PhD student working on this had read Karl Popper? His professor answered: "First of all you have to believe in long range dependence."

For scientific progress it is not necessary for scientists to publicly pronounce that the denounce their old work. It is more than sufficient that they reference the new work (which without critique is interpreted as signalling that the reference is worth reading), that they adopt the new ideas, use the new methods and again build on them. The situation is also normally not that clear, scientific progress is a continual "negotiation" process while evidence is gathered. Once the dust settles and the evidence is clear, it would be trivial and silly to ask for a denouncement.



Calls for walks to walks to Canossa only hinder scientific progress.

politics

In politics the situation is different. If politically helpful, the opponent changing his or her mind is put in a bad light and is called flip-flopping. If someone once made a mistake, very unhelpful calls for public retractions, repentance and walks to [[Canossa]] are made. Anything to humiliate the political opponent. Politics does not have to be this way, when I was young politicians in The Netherlands still made an effort to understand and convince each other. As sovereigns we should encourage this and reward it with our votes.

A funny example in the political climate "debate" is that the mitigation skeptic FoxGoose thought the could smear BBD by revealing the BBD had once changed his mind, that BBD was once a mitigation skeptic. He had not counted on the reaction of more scientifically minded people, who praised BBD for his strength to admit having been wrong and to leave the building of lies.

I discovered that I was being lied to. This simply by comparing the “sceptic” narrative with the standard version. Unlike my fellow “sceptics” I was still just barely sceptical enough (though sunk in denial) to check both versions. Once I realised what was going on, that was the end of BBD the lukewarmer (NB: I was never so far gone as to deny the basic physics, only to pretend that S [the climate sensitivity] was very low). All horribly embarrassing now, of course, but you live and learn. Or at least, some of us do. ...

Always check. Fail to do this in business and you will end up bankrupt and in the courts. I failed to check, at least initially, and made a colossal prat out of myself. Oh, and never underestimate the power of denial (aka ‘wishful thinking’). It’s brought down better people than me. ...

There wasn’t a single, defining eureka moment, just a growing sense of unease because nothing seemed to add up. ... Once I eventually started to compare WUWT [Watts Up With That] with RC [RealClimate] and SkS [Skeptical Science], that was it, really.

Mitigation skeptics sometimes state something along the lines: I would believe climate scientists if only they would denounce the hockey stick of Michael Mann as a hoax. Firstly, I simply do not believe this. The mitigation skeptics form a political movement. When mitigation skeptics do not even tell their peers that it is nonsense to deny that the CO2 increases are man made, they signal that politics is more important to them than science. For the worst transgressions at WUWT you will sometimes see people complaining that that will make their movement look bad. Such a strategic argument is probably the strongest argument to make at WUWT, but the right reason to reject nonsense is because it is nonsense.

Secondly, the call to denounce past science also ignores that methods to compute past temperatures have progressed, which is an implicit sign that past methods could naturally be improved. An explicit denouncement would not change the science. To keep on making scientific progress, we should keep the scientific culture of mild indirect criticism alive. The disgusting political attack on science in the USA are part of their cultural wars and should be solved by the Americans. That is not reason to change science. Maybe the Americans could start by simply talking to each other and noticing how much they have in common. Stopping to consume hateful radio, TV and blogs will make the quality your life a lot better.

Let me close with a beautiful analogy by Tom Curtis:

We do not consider the Wright brothers efforts as shoddy because their engines were under powered, their planes flimsy, and their controls rudimentary. To do so would be to forget where they stand in the history of aviation – to apply standards to pioneers that develop on the basis of mature knowledge and experience in the field. Everybody including Michael Mann is certain that, with hindsight, there are things MBH98 could have done better – but we only have that hindsight because they did it first. So, the proper gloss is not “shoddy”, but pioneering.

Tweet

Related reading

Are debatable scientific questions debatable?
On the difference between scientific disputes and political debates.

Falsifiable and falsification in science
Scientific theories need to be falsifiable. Falsification is not trivial, however, and a perceived discrepancy a reason for further study to get a better understanding, not for a rejection of everything. That is politics.

How climatology treats sceptics
I have been very critical of homogenization. I had good arguments and was royally rewarded for my skepticism. The skeptics that cry persecution may want to have a second look at the quality of their "arguments".

Scientific meetings. The freedom to tweet and the freedom not to be tweeted

Stop all harassment of all scientists now

Climatology is a mature field


* Top photo of Newtons grave from Wikimedia is in the public domain.

Five reasons scientists do not like the consensus on climate change

There is a consensus among climate scientists that the Earth is warming, that this is mainly because of us and that it will thus continue if we do nothing. While any mainstream scientist will be able to confirm the existence of this consensus from experience, explicitly communicating this is uncomfortable to some of them. Especially in the clear way The Consensus Project does. I also feel this disease, so let me try to explain why.

1. Fuzzy definition

One reason is that the consensus is hard to define. To the above informal statement I could have added, that greenhouse gasses warm the Earth's surface, that CO2 is a greenhouse gas, that the increase in the atmospheric CO2 concentration is mainly due to human causes, and so on. That would not have changed much and also the fraction of scientists supporting this new definition would be about the same.

You could probably also add some consequences, such as sea level rise or stronger precipitation, without much changes. However, if you would start to quantify and ask about a certain range for the climate sensitivity or add some consequences that are harder to predict, such as more drought, stronger extreme precipitation, the consensus will likely become smaller, especially as more and more scientists will feel unable to answer with confidence.

Whether there is a consensus on X or not is a question about humans. Such social science questions will always be more fuzzy as questions in the natural sciences. I guess we will just have to live with that. Just because concepts are a bit fuzzy, does not mean that it does not make sense to talk about them. If you think some aspect of this fuzziness creates problems, you can do the research to show this.

2. Scientific culture

By defining a consensus and by quantifying its support, you create two groups of scientists, mainstream and fringe. This does not fit to the culture in the scientific community to keep communication channels open to all scientists and not to exclude anyone.

Naturally, also in science, as a human enterprise, you have coalitions, but we do our best to diffuse them and even in the worst case, there are normally people on speaking terms with multiple coalitions.

However, also without its quantification, the consensus exists. Thus communicating it does not make that much difference. The best antidote is for scientists to do their best to keep the lines of communication open. A colleague of mine who does great work on the homogenization thinks global warming is a NATO conspiracy. My previous boss was a climate "sceptic". Both nice people and being scientists they are able to talk about their dissent in a friendlier tone as WUWT and Co.

3. Evidence

Many people, and maybe also some scientists, may confuse consensus with evidence. For a scientist referring to a consensus is not an option in his own area of expertise. Saying "everyone believes this" is not a scientific argument.

Consensus does provide some guidance and signal credibility, especially on topics where it is easily possible to test an idea. If I had a new idea and it would require an exceptionally high or low amount of future sea level rise, I would probably not worry too much as there is not much consensus yet on these predictions and I would read this literature and see if it is possible to make matters fit somehow. If my new idea would require the greenhouse effect to be wrong, I would first try to find the error in my idea, given the strong consensus, the straight forward physics and clear experimental confirmation it would be very surprising if the greenhouse theory would be wrong.

For scientists or interested people knowing there is a consensus is not enough. Fortunately, in the climate sciences the evidence is summarised every well in the IPCC reports.

The weight of the evidence clearly matters: The consensus in the nutritional sciences seems to be that you need to move more and eat less, especially eat less fat, to lose weight. As far as I can judge this is based on rather weak evidence. Finding hard evidence on nutrition is difficult, human bodies are highly complex, finding physical mechanisms is thus nearly impossible. The bodies of ice bears (eating lots of fat), lions (eating lots of protein) and gazelles (eating lots of carbs) are very similar. They all have arteries and the ice bears arteries do not get clogged by fat; they all have kidneys and the lions kidneys can process the protein and their bones do not melt away; they all have insulin, but the gazelles do not get diabetes or obesity from all those carbs. Traditional humans ate a similar range of diets without the chronic deceases we have seen the last generations. Also experiments with humans are difficult, especially when it comes to chronic decease where experiments would have to run over generations. Most findings on diet are thus based on observational studies, which can generate interesting hypotheses, but little hard evidence. It would be great if the nutritional sciences also wrote an IPCC-like report.

For a normal person, I find it completely acceptable to say, I hold this view because most of the worlds scientists agree. I did so for a long time on diet, while I now found that the standard approach does not work for me, I feel it was rational to listen to the experts as long as I did not study the topic myself. It is impossible to be an expert for every topic. In such cases the scientific consensus is a good guiding light and communicating it is valuable, especially if a large part of the population claims not to be aware of it.

4. Contrarians

The concept "consensus" is in itself uncomfortable to many scientists. Most of us are natural contrarians and our job is to make the next consensus, not to defend the old one. Even if our studies end up validating a theory, the hope and aim of a validation study is to find an interesting deviation, that may be he beginning of a new understanding.

Given this mindset and these aims, many scientists may not notice the value of consensus theories and methods. They are what we learn during our studies. When we read scientific articles we notice on which topics there is consensus and on which there is not. When you do something new, you cannot change everything at once. Ideally a new work can be woven into the network of the other consensus ideas to become the new consensus. If this is not possible yet, there will likely be a period without consensus on that topic. If there is no consensus on a certain topic, that is a clear indication that there is work to do (if the topic is important).

5. Scientific literature

A final aspect that could be troubling is that the consensus studies were published in the scientific literature. It is a good principle to keep the political climate "debate" out of science and thus out of the scientific literature as well as possible. It is hard enough to do so. Climate dissenters regularly game the system and try to get their stuff published in the scientific literature. Peer review is not perfect and some bad manuscripts can unfortunately slip through.

One could see the publication of a consensus study as a similar attempt to exploit the scientific literature. Given that all climate scientists are already aware of the consensus, such a study does not seem to be a scientific urgency. Furthermore, Dana Nuccitelli acknowledged that one of the many aims was to make "the public more aware of the consensus".

However, many social scientists do not seem to be aware of the consensus and feel justified to see blogs such as WUWT as a contribution to a scientific debate, rather than as the political blog it is, that only pretends to be about science. One of the first consensus studies was even published in the prestigious broadly read journal Science. Replications of such a study, especially if done in another or better way seem worth publishing. The large difference in the perception of the consensus on climate change between the public and climate scientists is worth studying and these consensus studies provide an important data point to estimate this difference.

Just because the result sounds like a no-brainer is no reason not to study this and confirm the idea. Not too long ago a German newspaper reported on a study whether eating breakfast was good for weight loss. A large fraction of the comments were furious that such an obvious result had been studied with public money. I must admit, that I no longer know whether the obvious result was that if you do not eat breakfast (like Italians) you eat less and thus lose weight or whether people that eat breakfast (like Germans) are less hungry and thus compensate this by eating less during the rest of the day. I think, they did find an effect, thus the obvious result was not that it naturally does not matter when you eat.

As a natural scientist, it is hard for me to judge how much these studies contribute to the social sciences. That should be the criterion. Whether an additional aim is to educate the public seems irrelevant to me. The papers were published in journals with a broad range of topics. If there were no interest from the social science, I would prefer to write up these studies in a normal report, just like an Gallop poll. However, my estimate as outsider would be that these paper are scientifically interesting for the social sciences.

Outside of science

An important political strategy to delay action on climate is to claim that the science is not settled, that there is no consensus yet. The infamous Luntz memo from 2002 to the US Republican president stated:

Voters believe that there is no consensus about global warming within the scientific community. Should the public come to believe that the scientific issues are settled, their views about global warming will change accordingly. Therefore, you need to continue to make the lack of scientific certainty a primary issue in the debate

This is important because the population places much trust in science. Thus holding that trust and the view that there is no climate change must produce considerable cognitive dissonance.

There is a consensus within the Tea Party Conservatives that human caused climate change does not exist. It is naturally inconvenient for them that this is wrong. However, I did not make up this escapist ideology. Thus for me as a scientist this is not reason to lie about the existence of a clear consensus about and strong evidence for the basics of climate change. Even if that were a bad communication strategy, which I do not believe, my role as a scientist is to speak the truth.

What do you think? Did I miss any reason why a scientist might not like the consensus concept? Or an argument why these reasons are weak if you think about it a bit longer? I will not post comments with flimsy evidence against The Consensus Project. You can do that elsewhere where people are more tolerant and already know the counter arguments by heart.

[Update, 23 Sept 2014. This post is now linked on Spiegel Online, where the local climate "skeptic" Axel Bojanowski needs no act as if I agree with him. I admit that the title suggests this, I was hoping to get a few "sceptics" to read it, but I was hoping that people reading the post itself would see that every single "reason" is countered. Thus Bojanowski was cherry picking, I hope it was not on purpose, but just by not reading carefully.

Axel Bojanowski calls the topic of the Cook et al. study a "banality". Because even the most hardened skeptics of the climate research do not doubt the physical basis that greenhouse gasses from cars, factories and power plants heat the atmosphere. (Selbst hartgesottene Kritiker der Klimaforschung zweifeln nicht an dem physikalischen Grundsatz, dass Treibhausgase aus Autos, Fabriken und Kraftwerken die Luft wärmen.) It would unfortunately be a great jump forward if Bojanowski was right.

The blog Global Warming Solved lists 16 people/blogs that agree with them that climate change is not man-made. In this list are well known people/blogs from the "skeptic" community: Roger “Tallbloke” Tattersall, The Hockey Schtick (often cited at WUWT), the German blog No Tricks Zone (Pierre Gosselin, who is followed by Bojanowski on twitter), Tom Nelson, Climate Depot (Mark Morano; CFACT), Steven Goddard, James Delingpole, Luboš Motl, and Tim Ball (regularly posts on WUWT).

Roy Spencer recently wrote a post with the "Skeptical Arguments that Don’t Hold Water". Most of which were somehow acknowledging that CO2 is a greenhouse gas, that was the biggest concession he was willing to make. He realized that even this was controversial in his community and wrote in the intro:

My obvious goal here is not to change minds that are already made up, which is impossible (by definition), but to reach 1,000+ (mostly nasty) comments in response to this post. So, help me out here!

He got "only" 700 comments, but the tendency was as expected.

At the main Australian climate skeptic blog, I once pointed out that even the host, Jo Nova, accepts that CO2 is a greenhouse gas. That produced a lively push back and no one came forward to say that naturally CO2 is a greenhouse gas.

I can only conclude that some "high profile" "sceptic" bloggers pay lip service to accepting that global warming is man-made (while many of their posts do not make sense if they would). And that at least a large part of their audiences is against accepting any scientific fact that is accepted by liberals. ]


Tweet

Related reading

In case you do not like people judging abstracts, there are also surveys of the opinion of climate scientists. For example this survey by the people behind the Klimazwiebel.

Andy Skuce responds to critique of consensus study in his post: Consensus, Criticism, Communication and gives a nice overview of the various possible critiques and why they do not hold water.

On consensus and dissent in science - consensus signals credibility


Photo: „Paris 2010 - Le Penseur“ by Daniel Stockman - Flickr: Paris 2010 Day 3 - 9. Licensed with CC BY-SA 2.0 via Wikimedia Commons.