After hearing Jiahuang's derivation of Boltzmann distribution from first principles of thermodynamics(btw, Jiahuang is in J1) recently, it struck me that utilitarianism might not be as useful as I had thought after all.
There are two interpretations of utilitarianism. One is that "We
should always try to maximise total utility in a society." The other is that "Society
can be understood in terms of maximisation of personal utility." I have felt that the first interpretation is pretty problematic, but I saw potential in the second interpretation.
One impetus for the second interpretation is that it readily allows us to apply the powerful tools used in physics and mathematics to understand society in a statistical way. After all, in a reaction chamber, whether an individual molecule would react or not is a very tricky business, but at a large scale and long time frame (reactions take place to moles of particles in femtoseconds), regular patterns are observed at the macro level (such as temperature and concentration). We do not need to understand quantum chemistry to study reaction kinetics at a macro level; so perhaps, you also do not need to fully consider the complexity of individuals to understand macro-level measurements (perhaps things like as inflation, population levels and Human Development Index?). In other words, are there social phenomena and measurements that are
emergent from considering human interactions at a statistical level? (I believe this is similar to Durkheim's "social facts".)
Let us consider thermodynamics again. As Ricardo once said, "If you ask an atom what the temperature is, he wouldn't know. But he can tell you how much kinetic energy he has." Note that in a box of gas, temperature in Kelvin is equal to twice the average kinetic energy of molecules divided by the degrees of freedom and Boltzmann's constant. It is therefore not hard to imagine that some measurement that is fundamental to our understanding of society would appear pretty elusive and unintuitive to an individual. Also, you can't just take any statistical measurement of society and expect it to be fundamentally significant. If you worked with the average magnitude of the velocities in a box of gas, or their harmonic mean weighted, or their median, or their geometric mean, good luck trying to derive something meaningful - because you still haven't figured out that mass and degrees of freedom are important. While trial-and error might eventually work, you might want to note that the above amazing relationship between temperature and KE of gas particles was not obtained by trial-and-error; it was derived from Newton's Laws. And we are not even talking about humans, we are talking about
gas particles - incompressible, completely elastic, infinitesimally small, with negligible attractions or repulsions, not there weren't that many factors to tweak anyway. But maybe! Just maybe, that elusive fundamentally meaningful property that arises from statistical measurements of society is looking right at us now. Can
utility save the day?
Maybe. But problem is, you can't measure utility. That doesn't mean we should give up on it though - the guys in the white lab coats couldn't measure entropy either, but they didn't give up. If you understand entropy, you'd probably think that the discovery of entropy is incredibly awesome. Unfortunately, I don't understand entropy, so I can't share your joy. But entropy is very fundamental to thermodynamics. So fundamental that the thermodynamic viability of anything (yes,
anything, even the collection of dust in your keyboard) is understood in terms of entropy. From what I've learnt in the first week of H1 econs, utility cannot be measured directly, but can be detected when exchanges occur. (It's kinda like temperature and heat, don'cha think?) Good, so we've got the zeroth law of Economics down. When are the next three laws coming? Ok, never mind, let's just assume that utility works like the negative of potential energy. That would neatly parallel the first law of thermodynamics.
A huge problem arises here. Potential energy only makes sense in conservative fields [i.e., there is no closed path which a particle can take in the field that would cause it keep gaining energy, or simply put, where energy is always conserved], and it only makes sense if there is also a social equivalent of a force, and a social equivalent of momentum, a social equivalent of kinetic energy, and social equivalent of displacement etc. In other words, you need the social equivalent of Newton's laws. Unfortunately it either doesn't exist or hasn't been discovered.
Also, even if we assume that the social equivalent of Newton's laws exist, utility is clearly not a conservative field - it is supposed to increase with every exchange. It is also not strictly increasing either, because while we are guaranteed that total utility increases at the point of exchange (I guess that's why people like shopping so much?), utility can increase or decrease by a great deal in between exchanges. Since we spend more time working or consuming goods than making transactions, it is reasonable for us to suspect that the utility changes between transactions would be significant enough to off-set the utility gain from transaction.
Then there is the oft-stated argument of human irrationality, which I would not repeat here. There is this counter-argument from my econs teacher: "Some people would buy more than it is rational while some people will buy less than it is rational, so overall their effects would cancel out, and this assumption still works at a statistical level." But Dan Ariely says something like, "No, even statistically speaking, humans are irrational in predictable ways." "In fact, we have some idea of how to make them behave in a certain way without them realising it," Thaler and Sunstein might (probably) add. Given commercial interest in making people irrationally profligate at a statistical scale (by advertising and market research), I have serious doubts that this assumption is in any way justified.
Lastly, is there a sufficiently large number of events at a sufficiently long time scale for us to suppose that perturbations from anomalous probability distributions, unstable equilibriums and random events have negligible effects on society? Even physical systems with well defined characteristics can behave in a chaotic manner - what more humans who are hosts to memes? - memes that are struggling to survive and propagate with our minds and technology as their mediums, which evolve at human time-scales as well. And that is a huge destabilising factor not present anywhere else. This means that there is no convenient analogy to human society that we can just adapt and use.
Oh no. I guess utility isn't that fundamentally meaningful property that arises from statistical measurements of society that we are looking for then. How now brown cow.
