Monday, August 03, 2009

Is there too much communication in science?

OpenScience (Dan Gezelter), "How Open Would You Want...", and CommunistSocialistSwine (?) have discussed the question whether science should be "open" and whether partial results should be shared. They have many right ideas but I find most of their attitudes one-sided and misidentifying the real problems and tradeoffs.

When we advocate "open science", our goals are:
  • Transparency in experimental methodology, observation, and collection of data.
  • Public availability and reusability of scientific data.
  • Public accessibility and transparency of scientific communication.
  • Using web-based tools to facilitate scientific collaboration.
These results are apparently desirable because it seems that they speed up the progress, by giving many people the access to the information, the data, the theories, and their known flaws. Such policies are also likely to make cheating harder. However, there are good arguments in favor of "closed science", too:
  • Credit is more likely to go to the actual authors which motivates people to work harder and find some new results.
  • Good new ideas are not being attacked just because they disagree with the average opinions (the "consensus") of the contemporary people: people don't get discouraged by premature criticism.
  • Complete, final results of a "hidden" scientific process are likely to be of a higher quality than "shared" partial results.
Because of similar reasons, communication is classified as a "bad thing" by some of the bloggers above. But is it the real culprit? Well, I don't really think so. At least a "proper" kind of communication is not a bad thing.


The list of reasons why science should be transparent looks pretty much obvious. Especially when it comes to some "routine" scientific work, the advantages look indisputable. For example, if the taxpayers pay some teams for their measurements of the temperatures, it seems crystal clear to me that the measurements - and all the calculations - should be as transparent as possible.

Temperatures can't be - and shouldn't be - classified. If the taxpayers have access to the measured data as well as the algorithms used to manipulate with them, many of them may verify them and exploit them for other purposes that are useful for them and mostly for the society, too. So I fully endorse the efforts of Steve McIntyre, Anthony Watts, and others to make the climate and weather data and algorithms collected and created by government-funded (or even internationally funded) institutions publicly available.

The availability of the information is a major criterion that describes the quality of their work. Because the taxpayers want this work to be good and useful, they should also try to raise this transparency. In this viewpoint, the people who measure the temperature are hired as workers who serve the society. Their work is pretty mechanical, comprehensible to pretty much everyone with scientific or technical education, at least after a finite amount of additional training, and they should't pretend that they're unreachable angels flying above everyone else.

They're surely not, especially because their field is certainly not among the most technically demanding ones.

Concerning the arguments for "closed science", we may say the following. If the scientists have to decide about their contributions, they must do so in a way that doesn't make their results classified. As far as the partial results go, well, the temperatures and theories describing them will always be partial - so we simply can't "ban" partial results. The accuracy of the measurements and the theories will gradually increase - at least if we are optimists: it hasn't increased much during the most recent USD 57 billion of climatology, as we will discuss later.

But the most important topic I want to discuss is the middle argument in favor of "closed science", namely the opinion that communication hurts great new ideas. But does it?

First, I want to mention one general observation about partial results and communication. If the amount of communication increases, there are two detailed things that increase as well: the amount of published partial results (because they're a part of the communication); and the amount of criticism directed against these partial results (because criticism is a part of communication, too).

The first group, i.e. published partial results, are likely to make the composition of published results "more incomplete" in the sense that a higher fraction if incomplete results is published. On the other hand, the second group, the criticism, is likely to reduce the amount of published partial results because their authors are afraid of - and discouraged by - this criticism. They must try to make their work more complete.

That's why the amount of communication doesn't have a simple, one-sided effect on the percentage of published work that is partial. Communication is neutral in this sense. So let me ask once again: does communication hurt new great ideas?

Well, some communication does. Communication may become a tool to promote groupthink. But is a "reduction of communication" the right tool to fix any of these problems? I don't think so because communication is also a tool to improve the legitimate, sensible, creative, individual research. You would be throwing out the baby with the bath water. In a vast majority of the negative cases, the real problem is not that the people communicate too much. The real problem is that the wrong people are determining the results of such a communication: meritocracy is broken.


Because communication between honest, sensible people does speed up the process of finding the right answers, at least statistically. There's no doubt about it. Every person who has written at least one exciting paper co-authored by a team must know why. A single researcher often gets stuck. He is often attracted to a potential well around a subtle error he is unable to avoid.

Feedback from decent peers (who don't severely intimidate or blackmail others!) can't really hurt him during the initial stages of a project. If he knows that they're wrong, he just ignores them. In many cases, the opposition even energizes him. The wrong old statements may also help the revolutionary to locate and articulate what is the core of the new idea and how it differs from the old framework. If he thinks that they're right, it means that he received the information that seems valid, and according to his best knowledge, he was probably on a wrong track.

Of course, the feedback could have been wrong. But it could have been correct, too. Most of the bloggers above assume that the author of the feedback is always correct (or always wrong), while the first person is always wrong (or always correct). But that's surely not the case. After all, they're two random people without any canonical "stickers". Each of them can be wrong and each of them may be right.

Moreover, the author of the wrong negative feedback doesn't have to agree with the "majority", as some bloggers also seem to incorrectly assume. Many great ideas that were very compatible with the "common" cutting-edge science have been slowed down by influential people in the vicinity of the revolutionary scientists who believed in wrong fringe theories and discouraged the brilliant discoverers of the new insights.

The very fact that only a few percent of the "new revolutionary ideas" turn out to be promising - statistics that is easy to obtain - actually implies that the author of the feedback is right in a vast majority of the cases. It's just completely wrong to assume that the authors of feedback are always "bad guys". Criticism is absolutely crucial for science. However, it's  sometimes correct and sometimes wrong.

What the feedback means is that the author of a new idea has received some information that can be either right or wrong and that the person can take into account or not. This simply can't hurt by itself. What can hurt is something different: that the new ideas and theories are often unfairly judged when they're released, even if these theories are "completed", or that the authors of new promising ideas "know" that this will probably happen, and this expectation discourages them from pursuing the idea.

Before I discuss the failures of meritocracy in detail, let me mention the reasons that are actually behind the well-known inefficiency of many teams in science:
  1. The credit and/or responsibility for the work is diluted which discourages the individuals from intense work - a classical problem with communist forms of ownership and sharing.
  2. The teams contain many people who are not competent enough and who are pushing the research in a wrong direction: the average talent and expertise of the group has a lower quality than the average of a few of its best members (or just the best one). The worse members become a liability for the best members.
Note that these two problems have nothing to do with the amount or the very existence of communication: they are about the ways how credit is shared, and about the competency of the people.

So if the authors of new promising ideas are actually discouraged and if these new ideas are prematurely "killed", this problem is completely different from a surplus of communication. In fact, most correct revolutionary ideas are being incorrectly thrown away by deluded people controlled by groupthink exactly after an unusually small amount of communication (and partially because of it!). It is very clear that the communication about some kind of questions (e.g. whether the Moon landing was shot in Arizona or New Mexico) is likely to be counterproductive while the communication about another kind of questions (what is the relative influence of oceans and solar activity on climate change) is likely to be productive.

The real difficulty for the society - or a group of scientists - is to find out which questions should fall into the first category and which questions should fall into the second category: and the answer is usually less clear than my examples suggest. These issues are manifestly decided by some people whose relative influence differs. The only invariant trouble is that this categorization may be up to the people who are incompetent, in many cases even according to the very sensible objective criteria that are already known.

Climate science

It shouldn't be surprising that climatology will be my first example. Despite 57 billion of U.S. dollars that were recently spent on this discipline, there has been no significant improvement in the accuracy of its predictions, retrodictions, and explanations. In fact, many people question the very fact that the climate was changing at all (by tenths of a degree per century) before the civilization began to influence it.

How did the funding, communication, groupthink, and meritocracy behave and interact in this field which became the ultimate contemporary symbol of a failed scientific discipline?

Well, there has surely been a lot of funding and a lot of groupthink. On the other hand, there has been very little sensible communication about the key questions - what are actually the main effects that explain the majority of one class of observations or another. In fact, not only the green activists among the laymen, but even many members of the climatological community argue that the "debate is over". The discipline is "settled". They can never say what the "settled" insights actually imply, and whenever they try to do it, they're proven wrong very quickly, but the field remains "settled" for them.

Why do we spend billions on a discipline that is "settled"? Research is only meaningful - and should be funded - when the questions are not quite settled yet.

The increased amount of funding for climatology hasn't helped the discipline, either. In fact, it has inevitably added a lot of people who are not as good in climate science as the original climate scientists (in average), who are not equally motivated (in average), whose main goal is not the science (but the gold or the political consequences), and the groupthink of these new weak people has actually hurt the quality of the research, despite the annual costs that have jumped by more than one order of magnitude.

Has there been too much communication in the field of "climate change"? Give me a break. The debate was over before it started. And did the small amount of communication protect the discipline from groupthink, as a German writer about physics generally hypothesized? It doesn't seem to be the case. The groupthink becomes most powerful exactly when the communication doesn't exist - especially when it is de facto outlawed (or "politically incorrect").

Communication - the inflow of ideas and objections from independently thinking, competent people - is actually the main weapon that fights against the groupthink. So the actual trouble occurs when communication is being de facto outlawed by people who are not independently thinking and/or who are not competent.

In other words, what's important for the sensible progress in a scientific discipline is that competent people are not being removed from the scientific communication by incompetent people who suffer from collective thinking. Now, an important question is who is actually competent or incompetent and who suffers of groupthink or who is independently thinking.

It's very clear that if one uses these terms that carry clear signs - "competent" is good while "groupthink" is bad - everyone is gonna scream that he is competent and unaffected by groupthink. But in both questions, at least one of the groups is clearly incorrect, despite the screaming.

In my opinion, scientists who are not independently thinking and independently evaluating the evidence shouldn't be called scientists at all. It's clear that there has to be a lot of people (a majority of the world's population!) whose job is composed of relatively mechanical tasks, or the promotion of other people's results, and so on (even within the scientific teams). But these people simply shouldn't be classified as scientists and they shouldn't be able to influence the selection of the insights of the scientific research. This distinction should be made clear in many contexts, including the staff of the LHC, indeed.

In climate science, the clearly negative word "groupthink" has been renamed to a neutrally sounding word "consensus", but it is clearly the very same thing. Whoever thinks that the calls that people should agree with the "consensus" are scientific arguments should be instantly fired from all research positions. All these people are freeloaders, a huge liability that only steals taxpayers' money and contaminates the scientific process by amplified noise.

There's clearly no consensus about these questions, and even if there were one, it would be completely inconsequential for a scientist's decision about the scientific truth (according to any proper definition of a scientist). Consensus, when it exists, may be a consequence of the scientific selection of correct (and elimination of incorrect) hypotheses, but it can never be its cause i.e. a criterion by which the truth is being found.

But needless to say, groupthink exists not only in climate science. In theoretical physics, we have people who love to paint themselves as the "warriors against groupthink" even though they are the clearest, hardcore pornographic champions of groupthink that you can ever imagine.

One of them, harbored by the Perimeter Institute, even gave a whole talk arguing that "science is like democracy" and science should democratically vote about the truth in science, so that the majority wins. Huh! The people who are able to believe this man that he fights against groupthink must be complete simpletons. I was stunned that this set included many people whom I used to consider somewhat sensible in the past.

People like Lee Smolin may claim to dislike groupthink but they're the ultimate representatives of groupthink and behind-the-scenes intimidation that tries to suppress any discussion and communication. The stories of his complaints against my review of his crackpot book were just shocking. I think that people doing things like that on this scale and in this context should be arrested, so when I found out that even the support of my position was somewhat "mixed", I had another reason to escape from the environment. Such essential issues simply cannot be ambiguous because nothing less than the very scientific integrity is at stake.

I just couldn't live in an environment where competent people are not allowed to openly disagree with the pseudointellectual trash energized by the most stupid journalists and by the anonymous scum on disgusting blogs who are more than allowed to attack the most prestigious scientific discipline and its researchers. They suffer from the most obvious groupthink you can imagine and they're very aggressive.

In fact, the situation especially on the Internet is so bad that this blog is literally the only blog on this blue planet that argues and explains why e.g. string theory is the only consistent theory of quantum gravity we know today, among hundreds of other questions. All other active blogs either focus on flowers or they're silent about the question, in order not to offend the group that dominates the Internet, or they're a part of the gigantic organized group movement of crackpots themselves. Despite this all-but-one dominance by idiots and hypocrites, the idiots are still willing to paint themselves as victims of groupthink of others!

It's just amazing.

So the right recipe to improve the efficiency of a scientific discipline is not to suppress the communication. Instead, it is to guarantee that the appropriate people - with a higher chance of being right - are being put into chairs that ultimately influence how much a particular scientific work affects the scientific knowledge (and policymaking, when we talk about applied sciences).

Such chairs have to exist, communication has to take place, but the results are ultimately determined by the quality of the people who matter.

And that's the memo.


  1. Is there too much communication in science? NEVER!

    I appreciate all of your efforts, Lubos, to expose the fraudulent "climate" science - and exposing all of the efforts that are made to keep perpetuating this fraud by keeping "science" a secret.

    Claims of global warming? Won't share the data? Won't share methods of arriving at the conclusion?

    We see this over and over with the data on ocean warming, polar warming, etc etc, it's just a methodology to cover up fraud.

    Anybody discovering something should be all too happy to explain what it is they found and be the best critic of the discovery.

    Technology is an exception - people keep technology secret, of course, to make money from it.

    But technology isn't "science" and never will be, and if it's kept secret and it isn't technology and it's called "science" - it isn't - it's just some more FRAUD

  2. i bet climate science community is the most open in terms of data and model availability to general public with a decent internet connectivity.. and if the theory looks shaky, that is if we are guessing things wrong about such omnipresent things as air and water!!!... then only "god" can help all the theories abour the very small and the very large :-P