Published: January 2, 2026
Some years ago, for a column on the philosophy of science that I was writing at the time, I thought about the frightening shadow presence in modern scientific theory that is friction. Now that may seem like lofty words but friction is an interesting and underappreciated part of modern physics. Obviously, is is an ubiquitous phenomenon. Friction is everywhere but it is in fact also fundamental to any macroscopic system as we know from thermodynamics. Given the 2nd law of thermodynamics, we know that for any spontaneous process the entropy of the universe must have increased. That increase in entropy typically results from the transfer of heat out of the system into the environemnt. The heat in turn is what we typically attribute to friction.
So there is something fundamental about friction in the sense that all real processes are not ideal in the sense of perfect reversibility and hence any real process will involve an increase in entropy that in the majority of cases we attribute to friction. Think of the stone rolling down the hill, in an idealized (high-school physics) version of the world that process is theoretically reversible and includes no friction and also no increase in the entropy of the universe. Evidently, in the real world this process is not strictly reversible as the stone will rub against the ground, creating friction and thus dissipating the air around it increasing the globabl entropy. That energy accounting exceeds the idealized version of potential energy being converted into kinetic energy without losses.
Now my claim in that column was that in a different - perhaps stranger - world, we would have based the fundamental laws of our physics theories on friction rather than idealized, perfect equations of frictionless motion. I do believe this is possible but is also pretty obvious that it would have made for a poorer physics. While friction is fundamental, strictly necessary, and ubiquitous on our macroscopic scale of things, it is true we don't really see friction in quite the same way on a quantum scale. That links to a larger question of how macroscopic theory (including general relativity as well as time irreversibility in thermodynamics) can co-exist with contradictory statements (equally powerful in explaining empirical data) from quantum theory. A probable analogue of friction in the quantum world is uncertainty as that in turn relates to information which is related to entropy via Landauer's theorem and alike.
What I was interested at that time however was not the prominence of friction in modern physics (in practice we do know well enough how to deal with it as any aerospace engineer can attest to, after all those big metal birds don't fall out of the sky as soon as their theoretical predictions meet actual frictionful air) but rather the epistemological prior our loveless relation to friction reveals. With friction I mean essentially anything that makes a theory deviate from its idealized mathematical form in practice. Think also turbulence, some types of boundary effects, numerical imprecision. You can recognize friction easily as that part of the mathematical model that nobody discusses and which typically includes one large, collect-all term that probably will be dropped out at the point you reach your homework assignments. This is how we typically deal with friction: we neglect it from the start and only introduce it as a mathemtical afterthrought when we have to.
But friction on the scale of physics humans live it is not an afterthought. It may sound trivial but a wheel wouldn't turn on a road were it not for friction. You could not drive a car without a seat-belt. Without all forms of turbulence an airplane wouldn't fly. Yet unlike essentially all other forms of physics, friction never merits its own study. We cannot really mathematically model most types of frictions outside of using them as a perturbation to some other law. We do not analyze friction in classical mechanics outside of its perturbation to Newtonian (or equivalent) equations of motions. Why however did Newton, while writing down his eponymous laws of motion, pay so little love to our friend friction? And, more interestingly still, why does seemingly noone ever object?
When one sets out to construct a theory of science, one must begin with a large collection of a priori unrelated empirical facts where in the process of theorizing some of those will become subject of shared, unifying laws. These laws must have predictive power and thus at least implicitly continue this set of finite empirical observations into an infinite set of possible observations. But some facts will lay by the roadside. Newton for example could not explain the precession of Mercury's orbit. There remained a discrepancy between the predictions of the fundamental laws and the empirical observation that also could not be resolved by more precise measurements. In this case, the discrepancy could only be ruled out by new theory, general relativity. But most of these deviations we (absolutely correctly) attribute to friction.
What I mean by an aesthetic choice is what subset of empirical facts we choose to focus on. Where does one begin to put the larger and ever expanding set of empirical observations around us into a rigid architecture of predictive theory? Newton chose to begin with the pure and idealized laws of motion, i.e. that a mass in motion will remain on an infinite straight trajectory unless acted upon by an external force. It takes a long time until one learns that one of these external forces may be friction. And then it takes even longer to learn that (again thanks to the 2nd law of thermodynamics) that force will always act on the mass in one way or another. Why did noone ever start a theory from the friction first? And, again, why did Newton's choice seem so intuitive that I have never heard anyone argue this choice with their teaching assistant in a classical mechanics class?
Newton's choice to construct laws of motions from idealized, frictionless motion was on some level an aesthetic choice. And it was the right one as I admitted to above. I do believe it would be possible to construct a theory of classical mechanics starting from friction first. It would certainly be an interesting gedankenexperiment. It is more difficult (aesthetically more difficult perhaps?), at least as I see it, because friction does not have a clear, idealized, archetypical form. Motion is easy to idealize: an infinite straight line. What is archetypical friction? Friction so strong an object cannot move and we treat empirical data as special cases of lower friction? Not quite. Probably a friction-based theory of mechanics would actually begin somehwere in thermodynamics and derive itself from versions of the 2nd law. A much more daunting challenge but more alluring promise if actually achieved.
In any case, it seems a universal aesthetic preference by people to treat friction as a perturbance to something more clear, universal form of the world. We seem to have an in-built preference for the ideal, the platonic archetype of objects. I was thinking about this problem again more recently in a very different context however. Another pretty widely accepted empirical fact (more sociological than physical this time) is that people, young people too, spend much less time in social contexts. Less time partying, less time socializing, fewer people are in sports clubs, associations, political parties. We are an increasingly private and fractured society.
There has been a lot written about the reasons and possible interventions given this increasing isolation. One conceptualization I quite like is by the sociologists Ramon Oldenburg and Dennis Brissett from the early 1980s of The third Place. A place that is neither work nor family but a third place for socialization where more distant but close (i.e. non-professional) social connections can be made. I've increasingly begun to think about social friction when going out. I do think our increasingly online society, particularly when online interactions are increasingly how we first learn the ropes of society as children and young adults, enforces our preference for idealized, frictionless social interactions. We want the perfect wedding, a perfect friend that we can depend on and that (if we're honest) doesn't really need us in that same way, we want a bofriend 6' 5", blue-eyes, finance. The true idealized social interaction perhaps is a relationship with a chatbot. A social interaction that entails all the hallmarks of a human interaction - minus the friction.
It becomes a truism when written out: real social interaction does not work like that. True social interaction is very frictionful. I for one increasingly want to think in a friction-first theory of social interactions. That the friction you feel - the disappointment, the annoyance, the ick of it all - is not a perturbation from some better, some idealized state of things, but a necessary property of it that only enables it to exist at all. The embrace of the beautiful, frightening frivolousness of social friction.
← Back to main.