Sunday, October 11, 2015

Shut Up and Calculate

Over the last few days, I have completely submerged myself in an intellectual world with which I'm familiar, but am by no means any kind of expert in. Namely, Quantum Mechanics.

Physics has always been a particularly interesting subject for me, in that, as a scientist, I have naturally developed a causal-mechanical thinking-style, if you will, and take great pleasure in knowing that if I compute things a certain way, I will almost certainly get one, very specific answer. That A causes, or reacts with B to produce C, and that this can be predicted to occur with regularity, given a set of environmental circumstances, is nice and clean, and gives me a warm gooey sense of knowledge about the nature of reality.

Then came my first introduction to quantum mechanics.**

**I'd like to take this moment to restate that I am a quantum physics enthusiast out of sheer philosophical and scientific curiosity; I do not claim to be an authority on the subject.

If any of you are at all versed in rudimentary quantum mechanics -- and it doesn't matter much if you are or not, in order to follow along with the gist of this post, so don't worry -- I'm sure you'll understand why this threw me off. (Nothing about quantum mechanics is ever totally predictable, in the traditional, strict deterministic sense. Quantum mechanics relies on probability distributions, described by the Schrodinger equation. In other words, just like Schrodinger's famous cat: the universe is probabilistic, not deterministic.)

This tends not to jive with most physicalists/determinists in science and philosophy, who claim "we're all just physical machines," because their picture of reality tends to revolve around the concept that life at the "classical level" somehow reverts to traditional, Newtonian mechanics (the world of billiard balls bouncing off of one another in predictable patterns), and that the realm of quantum mechanics is virtually meaningless to us, at least in our being able to utilize it to determine anything meaningful about actions and reactions at the order of magnitude at which we live. And, to some extent, I can sympathize with this view, from the perspective of the mathematics of quantum mechanics. For instance, have you ever been shown how to solve the Schrodinger equation for the hydrogen atom? Can you fathom doing it for a pancreas? Me either. That said, however, recently conducted experiments clearly demonstrate that large 60-C molecules called "buckeyballs" display wave-like interference patterns, similar to photons and electrons. This phenomenon has been replicated over and over, in various ways, and, in each case, seems to hold true. To me, this begs a pivotal question: where do we draw the line? At what point does the nuisance that is quantum mechanics fade away and give rise to the nice, clean, mathematically-precise world Newton described so well? It should be clear, at this point, that we can no longer suggest that "quantum mechanics only describes ultra-microscopic particles." Do we simply keep moving the line up and up? Or, perhaps the line is a figment of our imagination, and there really is only one reality; and it's quantum mechanical. Just a thought.

Without delving too deeply into the specifics of any aspect of fundamental quantum mechanics, or the dynamics of subatomic particles (as much as I find myself wanting to), let me just say this:

Quantum mechanics has never failed experimental replication. Ever. Along with Einstein's theories of general and special relativity -- although both are still technically falsifiable; and, interestingly, neither should be able to exist in the same universe as the other, yet very clearly do; hence the search for 'quantum gravity' -- quantum mechanics is quite possibly the most successful scientific theory that has ever been proposed, to date.

Although it is true, I daresay, that there is no more a successful theory in science than quantum mechanics, there are several competing interpretations of its so-called "spooky" nature, which has been observed since the time of Bohr and Heisenberg, and card-carrying theoretical physicists to this day have not come to any consensus as to which of these purported interpretations makes the most sense. Although there is universal agreement among physicists that quantum mechanics is an amazingly accurate and successful theory that describes reality in a seemingly infinite number of ways**, and that the so-called "wave-particle duality" is a fundamental aspect of the theory, there is no such universal agreement on how best to interpret the meaning and the implications of this theory and its application to reality.

**e.x. If quantum mechanics did not work as detailed, cell phones would not exist. Neither would personal computers, and a host of other things. Life as you know it couldn't be. (I say this to make sure you understand that this isn't just idle theorizing on the part of a number of incredibly smart men and women, there are vitally important practical implications of quantum mechanics.)

They know that quantum mechanics describes reality, but not necessarily the intricate hows and whys of the theory and its auspices. (Which, it seems to me, tends to be true of much of our scientific understanding in biology, as well.)

So, anyway, I was amid a discussion with two friends of mine, this week, wherein each of us held a distinct and opposing position on how we view the strangest aspects of quantum mechanics

I am of the opinion that the Copenhagen interpretation of quantum mechanics seems to be the most accurate way of describing the nature of reality. Specifically, I think there's an implicit role for consciousness (or, rather, the conscious observer) in the collapse of the wave-function. Therefore, I fall into what's called the Von Neumman interpretation camp.

One of my friends falls into the camp that believes in the Many Worlds (or Many Minds) interpretation of quantum mechanics. He doesn't believe the wave-function collapse occurs, at all; but, instead, that all possible realities exist and are retained in infinite states of probabilistic superposition. (Basically, that any and all possible realities that can occur do occur, simultaneously.)

My other friend believes in something called the Pilot-wave theory or the de Broglie-Bohm interpretation of quantum mechanics. (Or, at least denounces that, if the wave-function collapse indeed takes place, it needn't require a conscious observer to collapse it; thus, introducing the question: "What constitutes an observer?")

Although both of my friends disagree with me, and I with them -- which is totally fine, and all in good fun -- I cannot escape the idea that the wave-function collapse occurs, effectively condensing all the potentialities that exist into a single, unified, so-called "objective" reality. (e.g. Schrodinger's cat is no longer both dead and alive, simultaneously, after observation [re: measurement] takes place. Once observed, the wave-function collapses, quantum superposition fades -- rather, the state vector is reduced -- and the cat turns out to be either alive or dead, but not both.) I'm not a fan of the concept of "eternal superposition," which seems, to me, to be the natural philosophical progression of the Many Worlds view. And, despite the recent popularity of the many worlds interpretation of quantum mechanics, proponents tend to evade questions I would want answered, like: "What about the quantum Zeno effect?

There are, of course, a variety of different interpretations of the quantum mechanical phenomena that have been observed over the past decades, and every one of them (including the Copenhagen interpretation, to which I have sided, here) has issues in need of resolution. As well, there are many pop-culture references to things like quantum entanglement, some of which are quite accurate, although perhaps not well explained, while others are flagrant misrepresentations of these concepts. The point is, of the scientifically accurate and mathematically sound theories surrounding quantum mechanics, none seems to be able to delineate precisely which of these interpretations is most correct. And, as I say, physicists today still cannot agree.** Hence, my friends and I are free to speculate, and debate within our group discussions, about which concept seems to have a more solid basis in nature and why.

**This graph on Sean Carroll's website clearly shows what I'm trying to tell you; those polled, here, are card-carrying physicists, per se.

The point is that there's no consensus. And consensus, in science, is one of the predominant things that takes a bold, counterintuitive conjecture, or educated hypothesis and solidifies it into an accepted theory. There may not be universal agreement in the lay public about whether or not evolution, as a theory that describes biological adaptation over time, is true, but if you were to poll professional biological scientists, there would be an overwhelming majority vote that yes, in fact, evolutionary theory is true.

Could Einstein's relativity theory have made history, if no one came to realize and accept he was correct, after which followed the almost universal consensus that his ideas could be doubted no longer? I would think not. Then again, some would make the argument -- and, I find myself in agreement, here -- that it might have inevitably become the truth, at some point, because, mathematically, it is the only theory that appears to make sense and accurately describe the cosmological reality.

This focus on mathematical proofs, the quantifiable measurements we make which point us in one direction or another, is what drove the evolution of classical, non-relativistic physics toward relativistic physics and, then again, toward quantum physics. We go where the math takes us. Which brings me to the bottom 12% of the votes in Carroll's graph.

The bottom 12% that voted "I have no preferred interpretation," are often called the "Shut up and calculate" group. They are frequently cited as being the kind of hard nosed physicists that care only about the measurements and how the data adds up, and harbor a certain kind of cynicism about philosophical introspection. More importantly, they're not interested in speculating past what we can know beyond a reasonable doubt, from the data before them, and if they don't yet have sufficient data to come up with a reliable answer, then by God, they have yet to generate enough information; ergo, "your question will have to wait, we simply don't know yet; we need more data."

I would argue that the late, great Dr. Richard Feynman was in this Shut Up and Calculate camp. He cared very little for the philosophical implications of his (or others') work. I happen to think there's something immensely important about considering all the implications of our data and our scientific theories, past and present. But, in a world where there seems to be a lot more up-in-the-air speculation than nose-to-the-grindstone computation happening, I can certainly sympathize with his view.

Here's the message I'm trying to convey: In a world where physics is so often viewed as the golden ticket of truth and knowledge, and yet the fundamental basis on which the entire structure of physics is built is riddled with confusion, and not even the greatest minds in the field to date can agree on what hypotheses are more or less correct, how can we expect to have any hope, whatsoever, in determining what diet is best for human beings, when the system in question is unimaginably more complicated than those the greatest minds in the world are stuck on?

In a community where everyone has a divergent opinion on what's healthy and what's not, what's obesogenic and what's not, which diet is best for which people, under which circumstances, etc., I can't help but wonder if the nutrition and metabolism community might be better suited to follow the path of the aforementioned 12% minority [of physicists], and just shut up and calculate. Forget about what we think we know, and just focus our attention, instead of on ideals, on generating better, more generalizable data, utilizing solid statistical methods, carefully avoiding common missteps (like overemphasizing p-values, running too many significance tests, and using relative risk ratios rather than absolute risk and NNT, among other things), making all the data available for objective re-computation, analyzing quantifiable changes in biochemical assays, and more.

"But Ian," you might retort, "If we don't stand up for our ideas in this way, the raw food vegans will win!"

Um, no. No they will not. Because their data (i.e. their math) will not prove their bias. It never does. Either, they want to think it does, but vital information is missing or intentionally excluded, or they purport something to be true, but their own calculations prove them wrong!**

**e.x. Willet, et al. have continuously tried to vilify red meat as etiologic in the pathogenesis of cardiovascular disease. Yet, each time they present an epidemiological data set that "seems to suggest" a "trend toward significance" that "red meat consumption is linked to [heart disease, diabetes, etc.]" they fail miserably and are forced to write things like: "After adjusting for [BMI, smoking, drug use, added sugar, etc.] the significance of this finding was attenuated." (i.e. We found shit, but we want you to come away thinking meat is bad, so... let's skew the conclusion. But, guess what, they can't skew the mathematics! They can only try to cover things up in sneaky ways, for whatever reason, by omitting relevant data points. But, assuming all the mathematical methods are made available to you, you can re-compute their findings and confirm or deny their association for yourself.)

Perhaps if we spent less time analyzing the introductions and conclusions of new papers, from the perspective of our biases, and, instead, spent our time diligently comparing, calculating and re-calculating the tables, the results and the statistical methods, we might actually be able to realize some new insights. Or at least postulate more accurate assertions into the "reality" of human metabolism, based on more than our conceptions of "common sense."

Perfect example: Hall, et al. (2015) purported to "prove the insulin hypothesis of obesity wrong." Everyone freaked out over the potential for methodological bias, blah, blah, and yes, I agree those things are important to consider, on some level -- even though I mostly disagree that they really played a part in that particular trial -- but, for those who really wanted to "win the argument" on this one, all they had to do was look at the figures. Stop seething over your hurt feelings and do the math: No statistically significant difference between intervention groups, with respect to serum insulin concentrations, by the end of the trial. But, there was a significant drop in both groups, from baseline. Both groups lost [the same amount of] weight. Insulin decreased [the same] in both groups. Oh look, there's still a correlation between insulin concentrations and fat mass.

Shut up and calculate.

There's a special kind of power in mathematics and scientific computation. Use it.

And if, in time, the math proves you (or I) wrong, we must all suck it up and roll with the punches, because 2+2 will always = 4. Science readily adapts as its hypotheses are falsified, and this is almost always due to better calculations! Quantifiable data sits at the heart of "what is Science."

No comments:

Post a Comment