Monday, 10 February 2014

Why We Never Have, & Never Will, Predict Anything New

Hang on, I hear you protest, of course we predict things, don't we? We make predictions in everything from carefully studied specialised subjects like evolutionary biology, to simple sub-conscious predictions like 'tomorrow darkness will follow the day'. This is true, but it's a somewhat different thing. To understand why we don't predict anything new, you have to understand what is meant by 'new', and also that what we think are predictions are actually arrangements of ideas based on prior experiences. So in saying we never predict anything, what I mean is, we never predict anything that isn't already part of our repertoire of prior experience. 

This was a groundbreaking observation by empiricist philosopher David Hume, and it has never been refuted to this day. Hume drew the distinction between the kind of predictions made from prior experiences (like evolutionary biology and anything related to scientific enquiry) and our inability to know or predict anything that hasn't yet been experienced. This is what he called the difference between 'causality' and 'causation'. With the necessary prior experience we can distil knowledge of cause and effect (causality), but we have no direct intuitive ability to distil a connection between cause and effect (causation) - {don't worry, this will become clearer in a moment}.

Examples of causality would be something like: apples fall from the tree because of gravity, or if I bang my head it will hurt. Because of prior experience of gravity and banging our head, we can predict that when an apple begins to disconnect itself from the tree it will fall to the ground, and we can predict that if we bang our head on a low hanging tree branch it will hurt. But they are assumptions under the classical interpretation of gravity and solid objects. We could not predict from the classical system that there is a force of gravity according to the principles of quantum mechanics (quantum gravity), or that the tree is largely made up of empty space. Experiences of the quantum world have enabled us to understand gravity and solidity in ways that augment the Newtonian lens of reality, but without the agency to perceive the quantum world (empirical experience, scientific apparatus, and so forth), quantum physics couldn't be predicted.

When Hume says that we can have no direct intuitive ability to distil a connection between cause and effect in terms of causation, he is saying we can’t reason prior to experience anything new about nature*.So to recap, Hume’s contention (with which I concur) was that because we only discover things from past extrapolations, we experience only causality, and not causation - and we have to experience the causality before we can know of causation. 

So what is this 'causation' that we cannot predict or know without prior experience? I'll try to explain with a couple of simple thought experiments. Suppose that as of midnight tonight a fundamental fact in the universe is going to change. While you are sleeping tonight what you thought was a fundamental law of nature gets changed; from 6am something is altered in one of nature's multi-dimensions and we no longer can have a magnetic field through the application of electricity. When you wake up in the morning, you wouldn't have any way of knowing this before you've climbed out of bed, because as far as you are concerned nothing has changed. It would be impossible for you to wake up in the morning and reason your way to the conclusion that it will be impossible to have a magnetic field through the application of electricity.

But you are about to find out; you get up at 7am with no knowledge of this fundamental change in nature - but as you go about your daily business you start to sense something is wrong, because you notice subtle empirical indications. What empirical indications do you notice?  You turn on the radio and the electromagnets that used to amplify the sound coming out of the speakers no longer do what they did yesterday. The postman rings the doorbell but nothing happens except silence.  When he pushed the button, where a tiny electromagnet would pull a metal clapper against a bell, now there is no sound at all.  You decide to pull it apart - you examine the length of conductive copper wire wrapped around the metal, you replace the battery, yet still no magnetic field around the coiled wire.  Perhaps the circuit has been interrupted so that no electricity will flow. 

You decide that that must be it - the circuit has been interrupted - so you put it out of your mind for a while. Now a bit later in the day you have been informed that you are not alone - everybody's doorbells are the same.  The electronic circuitry no longer closes an electrical loop, meaning the circuit is no longer completed - a fundamental law in nature really has changed - there is no electricity flowing, and there is no magnetic field created to enable the clapper to become magnetised. 

Now here's the rub. We could arrive at the knowledge that a fundamental law of nature had changed when we learned about new causalities, but not before we'd experienced them.  This is what the whole edifice of science is like – everything new is experienced, not predicted.  If it can be predicted, it isn’t new, it is merely extrapolated from experience already held – and this is the brilliant legacy that Hume left us.

You may think you know of some predictions – you may have thought that it was through Newton's Law of gravity that we predicted the existence of the planet Neptune; that it was through discoveries of relativistic quantum mechanics that Dirac's equations gave rise to anti-matter; and that Maxwell's theories led to the prediction of radio waves.  But once again, these aren’t instances of predicting something new – they are only instances of extrapolating from prior experiencing and arranging ideas into prescient statements about how those ideas may play out in the future.  If we could reason our way to predict things without experience, then scientists wouldn’t be in labs, they would be in quiet rooms pondering in silence or talking in groups.

For those who are still not convinced
If there are still some of you who are still not convinced that without experience we would know nothing, here’s an extreme illustration I once thought up that should help you see where you're going wrong. Imagine we grabbed a baby boy from birth, stuck him in a room, kept him alive with food and liquid for 18 years, but at the same time we disconnected his ability to see, touch, taste, feel, and listen. Consider what we'd find after 18 years. The 18 year old brain would have none of the sensory perceptions or experiential abilities that ordinary minds have - he would be mentally moribund, devoid of any of the concepts we have acquired through our experiences of seeing, hearing, touching, smelling, tasting, and all the learning that accompanies that sense data.

Now imagine that at 18 the young man is suddenly given all his senses. He gets to sense the outside world for the first time. He has no experience of anything yet, but in front of him is a billiard table and two billiard balls. He is about to be tested on causality by observing an experimenter rolling one billiard ball to hit the other (the billiard ball illustration is the one Hume uses, so for the sake of faithfulness I am using that too).  Because Hume maintains that we know nothing without experience, and hence, we cannot predict anything without prior experience, he would argue that there is no way our 18 year old could have any expectation of what would happen when ball 1 hits ball 2. 

An 18 year old with absolutely no prior experience is not going to have the necessary experiential capabilities to infer the billiard ball causality from ball 1 to ball 2. That is to say, it only seems obvious to us that ball 2 will move after being hit by ball 1 because we have experience of physical reality – this young man has none, so he wouldn’t know if ball 2 will roll, disappear, shatter, liquefy, melt, change colour or stay motionless.  If ball 2 stayed motionless or liquefied he would have no reason to think it strange, just as a baby would have no reason to think it strange if he saw a man flying in the sky.

Hopefully these illustrations have conveyed not just a picture of Hume’s empiricist legacy, but also a clear picture of the distinction Hume made between causation (which we don’t have) and causality (which we do have), and the still relevant knowledge he gave us that there are no new predictions without prior experience.

* In his treatise Hume speaks of the kind of cause (causality) we can know as being “An object precedent and contiguous to another, and so united with it, that the idea of the one determines the mind to form the idea of the other, and the impression of the one to form a more lively idea of the other" Or to put it another way, Hume’s notion of causality is to do with ideas built upon a succession of activities related to how we perceive the material world, which is an object of study that belongs within the sphere of physical science.

 ** Photo courtesy of