We have all been soothsaid; some of us believe the people who can apparently look into the future with as much ease as we look out of our windows. Well, my article is not really about them. I am setting out to prove that we can never look into the future. Let me first define what I mean by looking into the future. Given the current state of being of a system, if we can predict the state of system in a future time accurately, we have effectively looked into the future.

The only scientific way of looking into the future would be to simulate the system in question. This approach hits an insurmountable boulder right at the beginning. The problem at hand is how to get the initial state of the system. This rather silly question has a well-established physical principle working against it, the uncertainty principle. The uncertainty principle simply states that you cannot accurately measure the position and momentum of a particle simultaneously. This means that we cannot have the initial condition of the system at hand to start the simulation. Even if we had the initial state, it is only an inaccurate estimate of the real state of the system. Heck! We cannot even determine what IS, forget what WILL BE.

Let us hope that Heisenberg is shown wrong and it is proved that we can get all the information we want. Accurate simulation will remain a mirage even then. Let me try to explain why I think this is so.

I will take the help of a concept called light cone here. I read about this concept in the book ‘A brief history of time’ by Stephan Hawking. If you drop a stone in a pond it creates ripples. A ripple starts as a point; it then expands as a circle of increasing radius. If you put draw a graph with time on z-axis and the shape of ripple corresponding to a particular time on x-y plane, you will get a cone. Now it is easy to see that a ripple can only affect events that occurred inside this cone. If you had a paper boat on the water, the ripple would move the paper boat only at the time when ripple is below the boat. At that time the boat will be within our cone, all the time before, it was outside our cone. Similarly light originating from a point will expand as a sphere of increasing radius (The radius being c*t, where c is the speed of light and t is the time elapsed). We can thus imagine a 4-Dimensionaly cone of light. Only the events that fall within the cone can be affected the by the event that started the cone. All events that are outside the cone are oblivious of the event.

Coming back to our discussion, In order to start the simulation we will need the initial state. Let the initial state reach the point where the system is being simulated in time T. Now if the simulation runs for time less than T, then we end state of simulation will be in past of the real system, the end state would have occurred in real system when the simulator was waiting for the input to travel and reach it! So the simulation must run for at least greater than time T, which in turn means that the system is in the light cone of the simulator. Which in turn means that the simulator can affect the system. Now the implication of this is profound, the simulator will have to simulate the effect of itself on the simulation! Now the act of simulating the effect of simulator on the system will have effects on the system, which in turn will need to be simulated. We can easily see that this is a classic case of infinite simulation! All of this discussion implies that a real simulation will never be possible even if we had an accurate initial state.

We have only argued that an accurate simulation is not possible. But we have left out an important kind of simulation, which is realistic. It is an inaccurate, probabilistic simulation. I leave the reader to ponder how feasible that is. I just ask you to consider the fact that our simulation should run ‘faster than time’ to actually peek into a most probable future.

Created on 14^th February 2005

• Home
• About
• Links
• Credits
• GuestBook
• E-Mail Me

• Technology
• Book Reviews
• C In DOS
• Reason

• Projects
• Cool Ones
• Headers
• TSRs
• Others
• Syllabus Programs
• User Files
• VTU Projects

• Standard
• Boxy
• Autumn
• Original