Marcelo Gleiser, Appleton Professor of Natural Philosophy at Dartmouth College, is a theoretical physicist who has worked on a diverse set of topics: cosmology, particle physics, phase transitions, condensed matter physics and biophysics. He is also a well-known author and public science communicator. A couple of months ago Marcelo suggested a guest piece for Cosmic Variance, and I’m delighted to be able to post it below. I hope you enjoy it, and I’ll encourage Marcelo to look in on the comments section and contribute there if he’d like.
Here are some thoughts on something that has been bothering me for a while. How do we know the world is the way it is? Easy, a pragmatic person would say, just look and measure. We see a tree, a chair, a table; we hear the wind, music, people talking. We feel heat and cold against our skin. Once our brains integrate this sensorial information, we have a conception of what is real that allows us to function in the world. We know where to go, what to eat, what not to touch; we enjoy a good meal, a nice hug. But what happens when we go beyond our senses, using tools to extend our conception of reality? We don’t see galaxies with the naked eye (well, maybe Andromeda on a moonless, dry night) and much less a carbon atom. How do we know they are there, that they exist?
When Galileo showed his telescope to the Venetian senators in 1609, some refused to accept that what they saw was real. More recently, late in the 19th century, physicist and philosopher Ernst Mach refused to accept the existence of atoms, claiming they would never be seen and hence couldn’t be proven to exist. Mach and the Venetian senators were wrong. What we see through telescopes is, of course, perfectly real; we capture photons—particles of light—that a celestial body emits (or reflects, for planets and moons). If the source doesn’t emit in the visible and is so dim that we can’t capture photons between red and violet, we capture photons from radio or infrared radiation, no less real even though our eyes can’t see them. When atomic electrons jump from orbit to orbit, they also emit (or absorb) photons that can be detected by instruments or, in the case of certain transitions, by our eyes. The instruments we use in the study of natural phenomena are an extension of our senses. This amplification of reality is one of the most spectacular feats of science, allowing us to see beyond the visible. So far, so good.
The situation gets complicated when the complexity of the phenomenon forces us to filter the data, and we select to study only part of what is happening. Our brains, of course, do this all the time, what we call “focus”; otherwise, we would be flooded with such an absurd amount of sounds and images that we wouldn’t be able to do anything. When we look at a star with the naked eye or with an optical telescope, we only see part of it, what it emits in the visible. A complete view of the star would incorporate all of its emissions, in the infrared, ultraviolet, x rays, etc. This fact has a simple but, to my mind, profound consequence: our construction of reality, being necessarily filtered, is incomplete. We only know what we can measure.
In the case of elementary particle physics the situation is even more alarming. The Large Hadron Collider, for example, should start working this coming summer or early fall. In its full capacity, it should produce around 600 million collisions per second. This translates to about 700 megabytes per second of data, more than 10 petabytes (1015) per year. That’s more than a million hard drives, each with a gigabyte. To make sense of this flood of information, physicists have to filter the data, selecting events deemed “interesting.” This selection, in turn, is based on our current theories that speculate on what’s beyond the standard model of particle physics, that is, theories that speculate on stuff we don’t know is there. Although these theories are mostly pretty solid (the Higgs particle as universal giver of mass; extensions of the standard model using more than one Higgs, supersymmetry or/and more than three spatial dimensions…) they can only be confirmed through the very same experiments whose outcome they are trying to predict. Given this mechanism, there is a risk that unexpected phenomena, not predicted by any current theory and hence not included in the subset of collisions deemed interesting, will be eliminated by the data filtering process. In this case, and in a paradoxical way, the theories that we construct to amplify our view of physical reality will actually limit what we can know about nature.