Wise People Think Alike

Within your brain there is a neuron that makes bad calls. However luckily there are a lot of other neurons that produce the correct1 response. The way your brain "processes" these mixed signals is by basically taking all the right and wrong signals and combining them with various weightings2 into a final result that is passed along. The benefit of such a system is that it is fault tolerant and flexible. The odd bad neuron or two does not stop the system working and more importantly allows it to try other configurations and other signal processing strategies. This is crucial to allow improvement of the processing ie to allow learning. This is an excellent metaphor for science.

Within science individual scientists or scientific groups form their own hypotheses. As I explained previously these hypotheses can be mutually exclusive even if they are supported by entirely valid research and peer-reviewed papers. Some of these hypotheses will in the fullness of time turn out to be wrong and others will be proven correct. However this time delay could be decades or more so science still has to continue in the meantime. Science would not be very useful if every time some scientist proposed a new hypothesis all work in the affected field stopped until the idea was vindicated or rebuked.

In working practice if the scientist/group proposing a new hypothesis is unknown and they do not have particularly convincing evidence it is likely they will simply be ignored. If they have a proven track record and loads of evidence some might start to look into the offered hypothesis/theory3. If more and more evidence appears that backs them up more and more scientists will get convinced. Eventually the majority of scientists would accept the new idea and it would become accepted as mainstream theory. In other words eventually the scientific consensus would shift to be that the majority of evidence says that the proposed theory is correct.

(An important tangent: Often when showcasing science people tend to ignore this gradual realization, model refinement and slow acceptance. Instead generally the "discoverer" revises history and claims that they, generally alone in some poetic setting, realized something was wrong, solved it in a flash of inspiration and then announced their obvious solution to rapturous applause at the next conference. This lone genius with the Eureka moment tale is one of the most persistent and unhelpful myths in science. Yes it benefits individual scientists by making them appear as singular geniuses however it also totally misrepresents the scientific process and the work of hundreds of other minor but crucial people in the discovery. I would go so far as to say nothing interesting in modern science can be realized in a flash or by an individual in isolation. An idea might be planted but it is the supplementary background reading, writing up, convincing of peers and public explanation which is the actual "science bit". If science progressed through flashes of inspiration and nothing else scientific institutes would merely be drug dens.)

So in the interim period before consensus/mainstream-acceptance on the new theory who is right? Group A and their model backed by their evidence or group B and their opposing model backed by their differing set of evidence? The answer is that both views are entirely scientifically valid as neither has been conclusively falsified yet. Either group leader could go to a conference and make their case using their own evidence. A scientific view is merely an evidence based opinion and there is no necessity for them to be representative nor pre-demonstrated as correct. The scientific consensus is entirely different and crucially the view of no individual scientist or group4 matters when it comes to forming the consensus.

Take the example above of the leaders of groups A and B going to a scientific conference. Both lecture on issue X and lay out their group's theory on it. The consensus of that conference would be whomever the majority choose to believe. Of course whom they believe is not just a random coin toss, scientists are generally smart and they know how to evaluate sources. The average scientist is generally better than the average member of the public at seeing through biased presentations at how good the evidence for the presented idea really is. Of course better than the public should not be read as perfect. Scientists are still influenced by industrial propaganda, strong personalities, authority, vested financial interest, well delivered lies and other nasty tricks5. Many will be fooled but like the bad neurons the hope is that the majority will work around them to produce the correct outcome.

Incidentally this is not merely down to integrity as scientists have a very vested career interest in not being fooled. Every time a scientist reiterates an idea they place themselves in one camp. If that camp turns out to be wrong people remember this and reduce their opinion of the scientist. The more vocally a scientist supported an idea that turns out to be wrong the more credit they lose. The upside of this being that the average scientist is sceptical and cautious when announcing their scientific opinions.

As such the majority views of a scientific conference is a pretty good indicator of the presented evidence. It's basically a quality weighted evidence assessment. Of course this is merely a conference consensus. Conferences themselves can have massive sampling problems. For example a climate conference at the Cato institute will have a certain slant to say the least. To get the scientific consensus on a certain issue one has to take the consensus opinion of all scientists with interest in this issue. Unfortunately it is normally impractical to ask every scientist. Some fields especially those with contentious or very financially sensitive issues have conferences where they attempt to state the consensus. This can be useful but tends to induce conspiracies and accusations of selections bias. A better way is to look at some open access metric that reflects the scientific views. For example surveying working practices is popular in some medical topics as is counting the published papers6 that support or reject a particular idea.

The scientific consensus is the only scientific view that can be stated as THE mainstream scientific view or the accepted scientific viewpoint. Everything else is simply too open to abuse, too unrepresentative or simply too open to interpretation to be useful. Is there any downside to using the scientific consensus? Well yes there is actually, for one it is very slow to change. The averaging nature of it means that new undisputable evidence will not immediately change it. As such it can be a bit behind the times. However this is merely a consequence of its stability. You simply cannot be both reliable and on the cutting edge (as any software developer can tell you). The robustness against manipulation7, inherent reliability, transparency and democratic nature of the scientific consensus means that in my opinion it is the only scientific position that I think should be promoted outside of science. For public outreach, basing political decisions and teaching I personally consider it scientifically reckless to promote anything else8.  


Footnotes

  1. "Bad" and "correct" are a bit vague given the individual signals are combined to give the correct output hence the bad signal might be just as important to the stability of output as the correct.  
  2. Being a "neural network" The weightings and combination are famously adapted according the assessed success of a configuration. 
  3. "Hypothesis" and "theory" are not generally interchangeable words in science but this particular paragraph is about ideas which one scientific group regard as a theory whilst others would regard as a hypothesis.
  4. Unless it's a hypothetical incredibly niche field in which only one scientist has any interest and his proposed model doesn't affect more popular fields.
  5. How prevalent these tricks are depends both on your field, it's transparency, the public attention it gets and the money made relating to it. 
  6. Papers are effectively snapshots of the authoring scientists' viewpoints. 
  7. I am totally ignoring conspiracy theorists who believe that the world's scientist might conclude to avoid change. If you believe these you are an idiot as scientist rarely agree and many truly hate each other. Worldwide scientific collusion is simply impossible.   
  8. Of course saying what you think or are doing is fine but giving your opinion is somewhat different to promoting it.