Search This Blog

Projection or Prediction

Most formal scientific cosmological investigation is preferential to the determination of the earliest events in our universe's history. As we tease away the noise, we illuminate conditions and dynamics ever closer to the moment of origin. The notion, reasonable enough, is that the more robust our knowledge of the earliest conditions, the greater our ability to predict the scope (limits) of all future events.

Prediction, remember, is the purpose of all intelligence, all science, all knowledge, all structure, and all evolution. So, yes, if one is interested in advancing our power and accuracy of prediction, it is methodologically reasonable to understand the earliest conditions of our nascent universe.

But somewhere along the scientific way, investigators revealed an interesting attribute of systems, of all systems, the attribute now called "The Second Law" of thermodynamics. What makes the second law interesting and unique is that it predicts the same future for all posible systems. That future is maximal dissipation. The second law says that all systems at all times are moving en mass towards disorder… are falling apart.

The 2nd Law was discovered by people interested in the flow of heat. Specifically, their interest was the maximal efficiency of steam powered equipment, factories, and transportation. They wanted to get the most production bang for their coal fired steam buck. And what they found was rather frustrating to a factory or train owner. What they found was that all systems no matter how well designed, leak a lot of heat, a lot of potential power, power that would ideally be used by the factory to make carpets, or by the locomotive to pull freight from point A. to point B. To make matters worse, the nature of this leaking of energy problem was such that it was irreversible. Once energy leaked out to the surroundings, any effort to recover it would cost more energy than was lost, lots more.

Now you might expect that there is something special about factories and locomotives that makes all this leaking energy so big a problem. You might be inclined to hope that leaking thermodynamic energy is specific to man made or artificial systems. You'd be wrong. Second law demanded energy dissipation is as true of natural systems as it is of man made systems. But the real kicker is that 2nd Law dissipation has nothing specific to do with heat or steam or coal, manmade or otherwise, but to all systems and all forms of energy applied in any way and under any conditions. This universality of dissipation became crystal clear in the 1940's when Claude Shannon of the Bell Labs in New Jersey, USA independently discovered the same dissipation dynamics in information and communication while trying to do for the telephone industry what the original thermodynamics investigators had attempted for the steam power industry a century and a half before. Shannon found that trying to shove signals down a wire or through the air resulted inevitably in noise that degraded the original signal and that insuring accuracy or distance in communication was a costly affair where more and more energy must be pumped into the system with less and less of that energy resulting in actually moving that signal from point sender to receiver. The final kicker came a few decades later when it became clear that computation suffered the same dissipative pitfalls as had been earlier discovered in communication and the conversion of power to work.

OK, so what does any of this 2nd Law stuff have to do with the question I posed at the top of this post, essentially: what is a more effective path towards predictive understanding of a universe, knowing when and how it started or knowing how it will all end?

Until the 2nd Law, all scientific effort resulted in understandings that started with initial conditions and worked forwards in time. Newton's laws of acceleration are a great example. If you know where and object is and what forces are brought to bear on that object, you can use newton's math to accurately predict that object's position at any time in the future. Einstein's work simply reinforced Newton's laws and provided a more robust contextual understanding of why they worked and when they could be expected not to work. But the 2nd Law is a strange bird indeed. The 2nd law simply doesn't care how a system starts, or what it is made of, or what forces pertain. The 2nd Law focuses our attention on the way systems move into the future, and mostly, on what systems become in the end. That end, on the grandest universal scale is something called "heat death". Heat death isn't really an end, the time doesn't stop, its just that things fall down and fall down, the dissipate and dissipate until less and less becomes posible. The slide into maximal dissipation is what we call an asymptote, it is an end never actually met. An end that in effect keeps ending. The universe is scheduled to become yet more dissipated forever. But the lion's share of that forever will look almost the same from eon to endless eon. The 2nd Law end is an end that never quite ends.

From a scientific perspective, at least from the perspective of most of the short history of science, the 2nd Law predicted end is absolute and perfectly knowable and absolutely independent on the initial state of our universe or for that matter, of any posible universe. Previous scientific knowledge had settled in on the idea that the future is only predictable to the extent that the past is known, that the laws of the dynamics of the universe are known, and even then, the predicted future becomes fuzzier and fuzzier the deeper one looks.  Yes the 2nd Law is strange indeed, flipping prediction end to end, it says that the end state of any universe is absolutely known, and the intractable part is instead the path towards that end. Got it? No, its not an obvious idea to grasp.

So now lets revisit the original question I asked. Is there any point in the full arch of a universe's life, when what is known of its past is less important than knowledge of its end?

To answer that question one might want to look not to the beginning or to the end, but to vast middle. In both predictive models, the classical causal model that predicts the future by knowing the past, and the entropic model which says that the past is always just a ramp towards a perfectly knowable end state, it is the middle ground that is the the most intractable. A thermodynamically determined universe is one in which falling down is the determining factor. A thermodynamic universe is one in which the end is absolutely known but the path getting there is not known. In a thermodynamically determined universe, each new moment presents a new set of conditions that must be computed upon in order to make the best posible prediction of the shortest path from that indeterminate here and now to the perfectly determined eventual then and there.

A universe locked into the 2nd Law dance would seem to be a universe in which the dynamics of change is a dynamics that becomes better and better and understanding its own dynamics. We have come to table this cumulative understanding "evolution". Evolution it would seem, is the process by which a universe becomes better and better at playing the only game a universe can play, and that is the game of getting to the end state as soon as possible. So we must reframe our original question and ask which knowledge is most evolutionarily potent, knowledge of the past or knowledge of the future? Or, in the spirt of my original question, is the situation more complex, more dyanmic, does the answer to the question vary depending on the particular epoch one asks it? Is the past more determinate in an early universe, and the future more determinate in an older universe? The reverse hardly seems reasonable.

If not, if predictions are more dependent on knowledge of the end than they are of the beginning, and if this is true no matter when in the arch of the lifespan of a universe one asks, what can be said of the value of traditional cosmological formalisms? If evolution is a process by which a universe finds the shortest path from any beginning to its entropic end, how important is knowledge of a causal classical dynamics in the solving of any of the moment to moment shortest path computations that must be eternally computed?


… to be continued …