Science is the Art of Approximation

Occasionally (all too rarely!), someone will offer a guest post for LOTRW publication. Below is a thoughtful and welcome contribution from John Plodinec, who’s posted here before.

Scientists may seek Truth, but Science is the Art of approximation. Just as artists use their media to represent the truths they see about the world, we scientists use our equations, models and case studies to represent the truth as it has been revealed to us. But in those dark weary hours in the dead of night, we acknowledge to ourselves that we have provided only a representation of reality, an approximation of a deeper truth. Our models and conclusions are like shells in which we hope to hear an echo of the great ocean beyond.

Just as artists are judged by how well they have captured and conveyed the beauty or the ugliness, the emotion and the feeling of the scenes they portray, so we scientists should be – and are – judged by how well we represent the reality we see. George Box, the Renaissance Man of statistics, famously said – “Essentially, all models are wrong, but some are useful.” And this statement, I believe, reveals a prime criterion for judging how well a scientific model represents reality: its usefulness. As an example, look at how long it took for the elliptical model of planetary orbits to displace the older Ptolemaic system, which most now recognize as an inaccurate rendering of the Music of the Spheres. It was only when it was proven (applying Newton’s physics of gravitation to Tycho Brahe’s measurements and observations) that the elliptical model provided greater accuracy and predictive power (i.e., was more useful) that Ptolemaic epicycles were abandoned and eventually forgotten. Thus, the usefulness of a model depends on how accurately it predicts things that we are interested in, recognizing the uncertainties in what we see and measure.

What, then, should we make of claims that “The Science is settled,” or that “The consensus is clear,” in any domain? And, more importantly, what should we do if our models predict that bad things are going to happen?

To the first question I resoundingly reply that Science is never settled: we may – or may not – be satisfied with our models’ usefulness, but we must expect that more precise or different kinds of observations will eventually change our views on the usefulness of our models. Just as the experiments to confirm the Theory of Relativity demonstrated the limits of Newtonian physics, we can be certain that any model we find useful today will be at least modified and perhaps superseded by future scientific advances.

And what of Consensus? This is a much more difficult question because it is more multi-faceted. If we look at the world of Art, 150 years ago there was a clear consensus among the cognoscenti that the works of Manet, Monet, Renoir and other Impressionists were scandalous and did not conform to the standards required of true Art. It took decades for some of their works to be accepted as offering a new and useful way of looking at the world – one with greater light, more vibrant colors, more realistic scenes that echoed what each of us encounters in our daily lives. In other words, the consensus slowly changed and eventually reflected the belief that Impressionism provided a useful representation of our world.

If we apply this analogy to scientific models, we have to say that while consensus is important when it speaks to the utility of our models, it is also evanescent: today’s unthinkably radical representation of reality is tomorrow’s accepted truth; eventually to be displaced by something even more radical – and more useful. And a consensus about an assumption on which a model is based (e.g., that humans are affecting climate) says little about the usefulness of the model itself. Models must stand or fall based on their accuracy and predictive power.

The last question – what should we do if a model predicts something bad will happen – is the most difficult to answer, because it transcends the world of reality into the realm of values. And while I am relatively fearless in talking about science, I am not – quite – fool enough to venture anything but carefully into that minefield.

The preceding is not intended to drive opinions toward either of the antipodes of the climate debate, but rather to tone down the rhetoric. The lack of civility and the degree of intolerance demonstrate an unseemly – and unscientific – lack of appreciation of the Art of Approximation. We know much about our climate, but that knowledge seems dwarfed by the uncertainties inherent in that knowledge and the certainty that there is so much yet to learn. This, to me, inspires humility rather than the arrogant certainty that seems to abound.

John, I loved this piece. Many thanks for taking the thought to craft it, and for the spirit behind the thinking. It deserves reader comments, which I hope will be forthcoming. And I hope as well that others will accept my standing invitation to submit posts. A range of views are welcome.

This entry was posted in Uncategorized. Bookmark the permalink.

One Response to Science is the Art of Approximation

  1. John Bates says:

    John, wonderful and thoughtful framing. Thank you for sharing this. I’ll be passing it along. Regards, John Bates

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>