Wednesday, December 12, 2018

Scientific Consensus and Uncertainty

There is a line of thinking, you may have encountered it, that says something to the effect of "there is no such thing as consensus in science."  The argument goes that in science we never prove, only disprove and we are always discovering, learning, tweaking and we should always be holding what we believe to be true, especially the theories and laws most dear to us, with skepticism.  Therefore, we should not or cannot ever have consensus.  There should always be room for debate in science and consensus is a tool of the powers of authority to stop those that would challenge this authority. 

The real difficulty with this line of argument is that it takes something that is true and good about science (that we only disprove, that we constantly amend our theories when new observations or information is collected, that we should remain skeptical) and takes its logical end creating such epistemological relativism that would dictate we can never know anything or ever do anything with that knowledge.  Which is demonstrably not the case.  We do know things and we do things with this knowledge. 

Let's think about how this applies to engineering.  We've studied steel.  Steel is great; it's fairly plentiful (rather it's components are), it's very strong, it's relatively easy if somewhat energy intensive to make stuff with.  So we have an idea: let's build stuff with it!  Let's use it to build some buildings!  Modest at first but eventually, maybe we can build some really tall buildings!  Now if your position is really that we cannot build scientific consensus you'll say: "No, no!  We need to continue to learn about steel!  Perhaps there is some hidden property that we're unaware of that would make it unsafe to build with!  What if there is an as-yet-undiscovered reaction that occurs when it reaches a certain elevation that makes it impossible to build tall buildings with?"  But we do build with steel.  We've reached a level of confidence that allows us to say "we know enough about this material, we understand its properties, let's move on and do some engineering with it." 

While it is true that nothing in science is ever proven in the strictest sense, confidence is a technical term for how statistically likely it is for future experiments or observations to hold with past experiments and observations.  So while we might not have proven that steel is a good building material, the statistics are very much in our favor. 

We also do have what we call scientific facts.  A fact is a piece of knowledge that, again, has been tested over and over and over to such an extent that statistically we have very strong confidence that it is for practical purposes "true."  One of the go to examples of this would be if you hold an object in your hand (that has greater density than air) and let go of it, it will fall.  Steel's high strength is a fact.  Atoms consist of protons, neutrons and electrons: fact.  Now could we discover new aspects of the natural world that counter these observations?  We could.  It's just very, very unlikely. 

"Ah!" Says the anti-consensus line of thinking.  "But in science there is no "truth."  We've been wrong before and we could be wrong now!"  Yes, many many scientific facts, laws and theories that were consensus for very long periods of time have not held up to the scrutiny of the scientific method and new evidence.  It's quite true.  We once believed in something called "the aether," a substance through which light waves could propagate.  It doesn't exist; we now know that light and other electromagnetic waves are self-propagating.  We once believed there was a substance called "phlogiston" within all combustible materials.  We had lots of evidence based on observation and experiment that such a thing existed.  It turns out we were wrong. 

Scientists and consensus can get locked into what is "accepted" and fail to pick up on evidence that would lead us to new theories, facts and laws that better explain our observations.  This is a very real aspect of the scientific method.  But saying that we cannot ever have enough confidence in our observations to do anything with them is throwing the baby out with the bath water.  And, again, demonstrably false.  Engineering is nothing but taking scientific facts, measurements, theories, etc and applying them.  We reach consensus whenever we build new infrastructure and technology.  Whenever we develop medicine (and yes, sometimes we make mistakes) we are reaching consensus.  When I know that jumping from a great height is a bad idea because gravity will accelerate me quickly to the ground and severely injure or kill me, I'm working with scientific consensus.   

You may see this argument applied to climate science and climate change.  That because science never proves anything, saying that there is a consensus is antithetical to the scientific method and merely serves to silence dissenters.  It prevents us from making new discoveries.  But here's what we know: it is a scientific fact (in the technical sense) that carbon dioxide and other gases trap heat in our atmosphere.  It is a historical and scientific fact that humans have added a bunch of rampant carbon dioxide to Earth's atmosphere.  Therefore, we can infer (and have measured) a rise in average atmospheric and sea surface temperatures.  We have observed ice that has existed since humans have melting.  We have a very, very high level of confidence in these things.  And we have a very high level of confidence that Earth's atmosphere will continue to warm and that this will have other effects on our climate. 

Do we know exactly what will happen with the climate?  No, of course not.  We have some decent models but there are a host of variables we can't predict very well (mainly the social variables, i.e. what are humans and governments going to do as we move into the coming decades).  Humans are notoriously difficult to predict.  But just because we don't have a perfect climate model with extremely high confidence doesn't mean we don't still have some basic facts that allow us to act in a logical, reasonable way. 

Good scientists are always skeptical, always open-minded.  Good scientists do their best to truly think critically and examine if their observations are in conflict with what they believe.  But good science also allows us to create, to heal, to feed, to protect.  Good science allows us to act and not only study. 

Sources:

Kuhn, Thomas.  The Structure of Scientific Revolutions.  Chicago and London: The University of Chicago Press, 1962.

https://www.visionlearning.com/en/library/Process-of-Science/49/Uncertainty-Error-and-Confidence/157

https://eic.rsc.org/feature/the-logic-of-phlogiston/2000126.article

https://ncse.com/library-resource/definitions-fact-theory-law-scientific-work

No comments:

Post a Comment