Saturday, 26 May 2012

29. What is Reality?

It is believed by many that there is something called 'absolute reality', and that it is a different matter whether or not we humans can know all of it. Certainly there are limits to what we can know. But even within the knowable domain, reality today may not be the same as reality in the future. I shall come to this time aspect later in this post. But even at a particular instant of time, there is no such thing as absolute reality. As argued convincingly by Hawking & Mlodinow(2010), all that we have is MODEL-DEPENDENT REALITY; any wider or deeper notion of reality is a baseless myth.

Does something or somebody exist when we are not viewing it? There are two opposite models for answering this question. Which model of reality is correct? Naturally the one that is simpler, self-consistent, and most successful in terms of its predicted consequences. This is where materialism wins hands down. The materialistic model is that the entity exists even when nobody is observing it. This model is far more successful in explaining reality than the opposite model. And we can do no better than build models of whatever there is to observe, understand and explain.

Suppose 100 persons are asked to describe an object, including its colour, and all of them say that it is a chair. Further, suppose 98 of them say that it is a red chair, but the other two disagree about the colour seen by the majority. If further investigation shows that these two persons have a colour-blindness problem, the model of reality we humans build is that the object is a red chair.

But suppose it turns out that these two persons are not colour blind, and no matter what tests we carry out, we are unable to explain why they do not see or describe the chair as red. We then go (tentatively) by the majority view, or consensus. Of course, any model of reality can change in the light of new data and insights. This is the approach we adopt in science for building up our knowledge. We build models and theories of reality, and we accept those which are most successful in explaining what we humans observe collectively.

A model is a good model if: It is elegant and self-consistent; it contains no or only a few arbitrary or adjustable parameters; it explains most or all of the existing observations; and it makes detailed and falsifiable (i.e. testable) predictions.

That brings me to the M-theory (cf. Part 14) and the cosmic-inflation model (Part 17) in cosmology. Are they good models of reality? There are eminent scientists who vehemently attack both of them, and have even proposed alternative models. Nothing unusual about that. At the cutting edge of science, the edge is blunt or nebulous, rather than sharp: Experts disagree on many issues, and fight it out. But out of this informed debate consensus emerges gradually, usually when additional ('issue clincher') data become available, or when some genius formulates a great new model. M-theory and the multiverse idea are the most accepted we humans have at present, even though there are many arbitrary-looking parameters, and loose ends. In due course the models would get either established or rejected, but they are the best (i.e., most accepted, even beautiful) models of reality at present.

The cosmic-inflation model of reality ties up so many loose ends in cosmology, and explains so many observations, that some form of it is highly likely to survive in any scientific version of cosmology.

Reality is nothing deeper than the best available scientific model for it. Often a phenomenon or entity is so complex that no sensible model has been formulated yet. In such a case, we have to wait till science makes more progress.

I now come to the time dependence of reality. Let us consider our current understanding of the origin of our universe. Scientists agree, by and large, that our universe began with a Big Bang 13.72 billion years ago. This model stands on three pillars: (i) the observed 'Hubble expansion' of the universe; (ii) the observed cosmic-wave background; and (iii) the fantastic agreement between the predicted and observed relative abundances of the light elements hydrogen, helium, and lithium. Lawrence Krauss has discussed them in detail in his recent (2012) book A Universe from Nothing. I shall consider only the first of them here.

Observations show that the rate of expansion of our universe is increasing (cf. Part 15). What is more, the observable universe is expanding at present at a rate that is not much lower than the speed of light. Galaxies which we can see today will one day recede from us at a speed greater than the speed with which even the fastest possible signals (electromagnetic radiation) from them can reach us. They will then become invisible to us, permanently. This will happen ~2 trillion years from now.

We and our solar system will die out in ~5 billion years. But other civilizations and advanced science can emerge elsewhere in the distant future. Suppose you are one of those astronomers in the very distant future. When you look around the cosmos with your highly advanced telescopes, you would see none of the ~400 billion galaxies we humans see today. What will be your model of reality then? Certainly not the same as ours at present. There is no such thing as absolute, unique, invariant reality.

We are 'lucky' to be living in a period when evidence for the Big Bang is available. Astronomers far enough into the future will have no evidence and no reason to believe that the Big Bang occurred at all, and that there were galaxies other than their own.
We live at a very special time . . . the only time when we can observationally verify that we live at a very special time! (Lawrence Krauss et al.)

You can see a timeline of the far future in the Wikipedia.

Saturday, 19 May 2012

28. Free Will is an Illusion

You are not in control of your mind  -  because you, as a conscious agent, are only part of your mind, living at the mercy of other parts. You can do what you decide to do  -  but you cannot decide what you will decide to do. [Sam Harris, Free Will (2012)]

What is 'self awareness'? In the human context this term means that a person is aware of herself/himself and has a self-image. You 'know' where your body ends and something else begins. You also have the self-image of a live and thinking person.

Cut to evolutionary robotics. This field involves a good deal of simulation and generalization work, rather like the creation of ‘invariant representations’ in the human brain. Following Hans Moravec, let us consider an autonomous robot that is able to continuously update its CPU about its own configuration and that of the environment, using simulation algorithms and the continual incoming stream of information. Suppose further that the robot can carry out these computations a little bit faster than the real rate of change in the physical world. Such a 'smart' robot can then compute the consequences of its intended action before taking the action. If the simulated consequences are not desirable, the robot would change its ‘mind’ about what would be a more appropriate course of action under the circumstances. Such a robot can be viewed as having an inner life or 'consciousness', right?

The free-will idea says that every person is the conscious source of his/her thoughts and actions. For all practical purposes we subscribe to this effective theory. Otherwise, how to fix responsibility if, for example, a murder is committed? Our legal system has little use for the assertion that we are just automatons or machines, governed by the laws of physics.

But the free-will notion clashes with determinism, although attempts have been made to reconcile the two. How can there be many alternative ('freely chosen') effects of the same set of causes? Free will must be an illusion only.

If determinism holds, our actions are uniquely determined by previous events, and we are not free. Even if indeterminism holds, our actions are random, and once again we are not free to 'decide' to act one way or another. Either way, the free-will notion ends up as a logical impossibility.

In philosophy, two opposite positions have been in vogue: compatibilism vs. incompatibilism. The former accepts the possibility of both determinism and free will. The philosopher Daniel Dennett subscribes to it.

And incompatibilism says that determinism and free will are incompatible. This leads to three possibilities: (i) Choose determinism ('hard determinism'). (ii) Choose free will ('metaphysical libertarianism'). (iii) Reject both determinism and free will ('hard indeterminism').

According to a recent paper in Nature, 'Haynes, a neuroscientist at the Bernstein Centre for Computational Neuroscience in Berlin, put people into a brain scanner in which a display screen flashed a succession of random letters. He told them to press a button with either their right or left index fingers whenever they felt the urge, and to remember the letter that was showing on the screen when they made the decision. The experiment used functional magnetic resonance imaging (fMRI) to reveal brain activity in real time as the volunteers chose to use their right or left hands. The results were quite a surprise. .. The conscious decision to push the button was made about a second before the actual act, but the team discovered that a pattern of brain activity seemed to predict that decision by as many as seven seconds. Long before the sub­jects were even aware of making a choice, it seems, their brains had already decided.'

The authors argue that consciousness of a decision may be a mere biochemical afterthought, with no influence whatsoever on a person’s actions, so free will is an illusion.

The neuroscientist V. S. Ramachandran is rather ambivalent on the question of free will being an illusion. In his 2010 book The Tell-Tale Brain he talks about a patient suffering from 'ideomotor apraxia', i.e. an inability to perform suggested skilled actions. This is what he writes on page 131: 'What he lacks is the ability to conjure up a mental picture of the required action  -  in this case combing  -  which must precede and orchestrate the actual execution of the action. These are functions one would normally associate with mirror neurons'. Perhaps it is the conjuring-up part (a semi-conscious action?) which is showing up in the recorded neural activity before the actual action in the above-mentioned experiment.

Sam Harris has recently argued at length about why free will is just an illusion. Here are some points made in his book which are of practical importance:
  • The realization that free will is an illusion makes us more humane when it comes to reacting to crime, hatred, etc. How can you hate a person when you know that he/she is not fully in control of the bad or undesirable behaviour?
  • Nevertheless, when somebody commits a crime, society must react effectively, not in the spirit of punishment or retribution or hatred, but for providing deterrence against future crime.
  • Being aware of the contribution of the unconscious to our apparently 'freely chosen' actions can make us more responsible and sensible about how we react to situations and take decisions.
Within a religious framework, a belief in free will supports the notion of sin  -  which seems to justify not only harsh punishment in this life,  but eternal punishment in the next. And yet, ironically, one of the fears attending our progress in science is that a more complete understanding of ourselves will dehumanize us. [Sam Harris, ibid.]
For more on free will and consciousness, please see my article Evolution of Intelligence and Consciousness.

And watch this video by Sam Harris for the latest update:

Saturday, 12 May 2012

27. Actions, Reactions, Interactions, and Causality

The question of free will, which I mentioned briefly in Part 26, is linked to that of causality. The causality principle says that every phenomenon or event is an 'effect' for which there must be a 'cause' that precedes it. I shall argue here that this principle is, at best, only an effective theory, often useful in day-to-day situations, particularly when it comes to understanding simple (or simplifiable) macroscopic systems. But an effective theory may not always have a sound logical basis, and is often just a label or nomenclature for something we do not understand very well.

Habitually we tend to subscribe to the idea that any action is followed by a reaction. In fact, we have Newton's 3rd law of motion, which says that to every action there is an equal and opposite reaction. But can we really talk in terms of actions and reactions, or causes and effects, always? No.

Consider two protons, moving towards each other. The force of repulsion between them is not much when they are far apart, but increases as they approach each other, resulting in a bending of their trajectories. They cannot get too close to each other because of the repulsion, and go their separate ways after coming as close as they can.

Both proton trajectories have been affected. Can you tell which is the cause and which the effect? No. Instead of cause and effect, or action and reaction, it makes better sense here to talk only of an interaction. Bring in a third proton and it would become even easier for you to agree with me.

Cut to our solar system. Does the Sun go around the Earth, or does the Earth go around the Sun?

There are complications because of the Moon and other planets etc. (and other celestial bodies), so imagine a simpler situation in which we have only the Sun and the Earth. Most people will still say that it is the Earth that goes around the Sun. The psychology of such an attitude is that the Sun is much heavier than the Earth. But the reality is that the two go around each other. There is a 'centre of mass' for the Sun-plus-Earth system taken as whole, and they both go around that point.

The human tendency is that the larger object is taken as causing the effect on the smaller object(s). The real thing is that there are only interactions, rather than actions and reactions. Philosophers have been tying themselves into knots by carrying the action-reaction or causality idea too far. The absurdity of it all becomes palpable when they even talk of 'downward causality'. Why not just talk of interactions, rather than actions (causes) and reactions (effects)?

As I said, the cause-and-effect approach is at best an effective theory, albeit convenient to use in a large number of practical situations, with the proviso that the effect never precedes the cause. And the last part of this statement is subject to what the special theory of relativity theory demands (cf. Part 10), namely that a signal from the cause cannot travel faster than the speed of light. What this implies is that the meaning of the word 'simultaneous' is observer-dependent. The cause precedes the effect for all inertial observers. The cause and the effect are separated by a 'timelike' interval, and the effect is in the future of the cause.

And according to the general theory of relativity, the effect must belong to the future 'light cone' of its cause, even when the spacetime is curved.

When we come to quantum field theory, the causality idea gets linked to the much-debated 'principle of locality'. But let us not get into those things here.

Consider a beehive. It is a complex system. It has 'swarm intelligence'. No one is in command, not even the queen bee. Each bee follows some very simple 'local rules', and interacts with other bees in the hive. The effect here is the 'emergent' property of swarm intelligence. What is the 'ultimate' cause of this intelligence? Not the action of any one bee. The beehive is the archetypal example of a system in which it is meaningless to talk about causes and effects, or actions and reactions. It is interactions, through and through. This is not an isolated example. Complex systems are generally like that.

The causality idea is well-entrenched in the human psyche, in spite of the above-mentioned limitations. There is no need to abandon it, of course. In fact, much of our conventional science is based on it (conventional or 'traditional' science follows reductionism and constructionism). Logical reasoning in conventional science is one big chain of cause-effect-cause-effect-cause- .... interpretations: An effect becomes the cause for the next event or process, and so on. But conventional science is often quite inadequate for tackling complexity-related, highly highly nonlinear, problems. Radically new thinking is needed when any simplifying assumption can destroy the very essence of the complex system being investigated, or when it is impossible to model a system in terms of an adequate number of differential equations, or difference equations. But some of the unconventional approaches formulated have met with strong resistance from many practitioners of conventional analytical science. Mindsets do not change easily.

Be prepared to think in terms of interactions and correlations when necessary, rather than actions and reactions all the time. Such an approach will help you better understand the properties of complex systems, and keep you away from absurdities like 'downward causality'.

Saturday, 5 May 2012

26. Scientific Determinism, Effective Theories, Free Will

Quantum physics might seem to undermine the idea that nature is governed by laws, but this is not the case. Instead it leads us to accept a new form of determinism: given the state of a system at some time, the laws of nature determine the probabilities of various futures and pasts rather than determining the future and the past with certainty. (Hawking & Mlodinow (2010))
Laplace was perhaps the first to clearly enunciate the basic tenets of scientific determinism (cf. Part 25), according to which, given the state of the universe at one instant of time, a complete set of natural laws fully determines both the future and the past. This was classical scientific determinism. The fantastic success of quantum theory has necessitated a change: We now speak the language of probabilities, rather than certainties, as also of several futures and pasts (cf. Part 4), rather than the future and the past.

Real-life situations are usually so complex that it is not enough to have knowledge of the 'complete set of fundamental natural laws' for explaining all phenomena. It is often found necessary to formulate additional (empirical) laws as 'effective theories'. An example is the gravitational force experienced by a macroscopic object on the surface of the Earth. The gravitational interaction is present between any two atoms, but we cannot formulate and solve the equations governing the gravitational interaction between every atom in the macroscopic object and every atom in the Earth. Instead, an effective theory is formulated in terms of the mass of the object and a few other numbers like the value of the gravity constant (g) at the surface of the Earth.

Similarly, in chemistry we cannot hope to formulate and solve the totality of equations describing the interactions among all the positive and negative charges in a system. Instead, an effective theory involving concepts like valence deals with how chemical reactions occur.

This approach continues as we go up the ladder of increasing complexity. Details at one hierarchical level of complexity are 'summarized' or 'integrated over' to generate some effective parameters which are used for describing the details of the next higher level: From particle physics to macroscopic physics and chemistry; from chemistry to biology; and so on.

An effective theory is essentially a framework we create for modelling certain observed phenomena, without describing in detail all the underlying processes.

The computational limit (cf. Part 25) to understanding the reality of our complex universe manifests itself in a dramatic way when we ponder over the notion, or rather the illusion, of free will. Michael Shermer (2006) described it like this: 
As with the God question, scholars of considerable intellectual power for many millennia have failed to resolve the paradox of feeling free in a determined universe. One provisional solution is to think of the universe as so complex that the number of causes and the complexity of their interactions make the predetermination of human action pragmatically impossible. We can even assign a value to the causal net of the universe to see just how absurd it is to think we can get our minds around it fully. It has been calculated that in order for a computer in the far future of the universe to resurrect in a virtual reality every person who ever lived or could have lived (that is, every possible genetic combination to create a human), with all the causal interactions between them and their environment, it would need . . . .  1023 bits of memory. Suffice it to say that no computer in the conceivable future will achieve this level of power; likewise, no human brain even comes close.

The enormity of this complexity leads us to feel as though we were acting freely as uncaused causers, even though we are actually causally determined. Since no set of causes we select as the determiners of human action can be complete, the feeling of freedom arises out of this ignorance of causes. To that extent, we may act as though we were free. There is much to gain, little to lose, and personal responsibility follows.
The often-handy notion of free will can be regarded as another effective theory, like the many mentioned above. Even though free will is only an illusion, one can make progress in cataloguing and understanding psychological phenomena by pretending that people have free will (Hawking and Mlodinow 2010).

One goes a step further when modelling economics. The effective theory we use in economics is that people have a free will, and that their behaviour may or may not be rational, and that their decisions may be sometimes based on a defective analysis of the limited data at their disposal.

Although the free-will notion serves as a convenient effective theory for certain situations, we should not lose sight of the fact that biological processes, including those occurring in the brain, are determined entirely by the laws  of physics at the fundamental level (call it 'quantum determinism' or 'probabilistic determinism', if you wish). Our actions are therefore determined entirely by the laws of physics. A large number of experiments done on humans under controlled conditions have demonstrated this. For example, electrical stimulation of appropriate regions of the brain can make a person 'want' to move the hand, arm, or foot. Such findings are in line with the scientific idea of physical causes leading to all effects, and show free will its place: It is only an illusion, or, at best, a useful (though logically unsound) effective theory for certain practical purposes. 

More on free will in a later post.