B.00 APPENDIX B - Anecdotes             (Where, When, How and Why, did Evolution Need to Change?) B.01 The Change Over     Everything was going swimmingly. Literally. Microscopic organisms were swimming about in tepid bath water ... or something like that. I exaggerate, but not by a lot. Then something happened. Evolution became different. Or it seemed to. Instead of just more of the same ... there had been a slow steady increase in diversity of creatures, but it was a very slow increase indeed for all of three billion years. And then the rate of increase of that diversity, changed. That was during a period called "The Cambrian Explosion". The question was - why? Why wait so long and why then?     And now ... two anecdotes ... of personal experiences which persuaded me that my idea was plausible despite the apparent difficulties. I invite you to read my anecdotes and to reflect on similar experiences you may have had. B.02 Anecdote (1) Sunset: Recently, I read a report of a car crash - involving one of those new driverless cars. The account I read claimed that the system which provided this car with visual perception had, in effect, been dazzled by very bright sunlight.     And then I remembered being similarly dazzled by the setting Sun when I was driving my car Westwards across the North Yorkshire moors. The sun was directly in line with the road ahead. I could tell, easily enough, that my visual perception mechanism, was not operating very well.     My response was to drive my car very slowly (while hoping not to be struck from behind by another vehicle) ... until I found a parking place where I took a nap until the Sun was well below the horizon. Note that. To respond in that way I needed to be able to predict that the Sun would indeed sink below the horizon and roughly how long it would take to do that. It seems to me that to be able to tell that you are NOT seeing properly, you need to be able to see (in your imagination) what it is you should be seeing.     Turn your mind back to the previous section, where I showed you a picture of various shapes with which a brain might compare a silhouette of an arbitrary shape, which it needed to identify. That was in Appendix A (algorithms - A.09). If that ability was already available, it could also be the evolutionary precursor of the ability to imagine what it should be able to see. And therefore be the evolutionary precursor also of an ability to tell that it wasn't seeing what it should be able to see. B.03 Anecdote (2) Colour Change: The previous anecdote was about a change from a normal environment to one that was too brightly lit. This one is about a change to an environment where the perception of colour was all wrong.     Once, a long time ago, I gave bad advice to a stranger who was lost in the district where I lived. He asked me how to reach some place and I said he had to wait at a bus-stop (that one right there) for a red bus. This was at dusk, but there was still ample daylight.     Later, after dark. when the place was lit by yellow sodium lamp standards, I was passing, and I again saw the same man still standing at that bus stop. I asked what was wrong. He said no red buses at all had come that way. The only buses he had seen were dark blue. At that moment a red bus did arrive, but to my surprise I saw that under those yellow sodium lamps it appeared to be very dark blue - almost black. And the change of colour was total.     I pondered on that experience. How could we correct for that false perception? Perhaps we should all carry around a colour chart, with patches of colour and the names of the colour against each patch which it would appear to have in normal daylight. Quite so. That would work. But it might be considered somewhat too onerous a cure, for a mistake which did not happen very often ... unless ... unless the result of that mistake was usually fatal.     Is there a general principle here that could help us design driverless cars that will not be dazzled by bright sunlight or can see red buses when they are illuminated by yellow lamps - or in whatever unusual circumstances it may find itself?     Perhaps what we need is a team of very very small expert advisors. One for sunsets, and another for coloured street lamps. That's two. Perhaps what we need is a whole football crowd of very very very small expert advisors. Why do they need to be small? Because we do not want having to carry around a head that is very heavy.     And that suggests a new system which might work like this .... When you need them, you ask each expert to test the environment to see if their expertise is appropriate in the prevailing circumstances. They do this test themselves. You do not need to do any testing. To save time each expert does his/her own self-test in his/her own time (and all at the same time). I call that a auto-adequacy-test (or AAT). Then, when each advisor gives you advice on how to respond, they also each tell you how reliable is that advice according to the AAT. If it is not very reliable - forget it. Choose the one that comes at the start of that list of suggestions. So the choice is automatic. You do not need to waste time thinking about it. B.04 Thoughts     In the older and simpler progression of evolution, one pattern of behaviour is replaced by a newer version, which is found (in certain circumstances) to be more effective. But that improvement may be only temporary and be due to current (but somewhat unusual) environmental circumstances. So what is then found out (too late) is that the mechanism which works best most of the time has been thrown away.     A better strategy might be to keep both forms of behaviour available for deployment - (or a even large collection of alternative forms of behaviour). These alternatives could then be offered - each in association with its own standard and relevant measure of merit. So - if these alternatives were placed in a sequence from which the first alternative would normally be chosen for performance, the subsequent choice could still be very rapid indeed.     That suggests that what is needed is a whole lot of semi-independent brain-units which are more or less autonomous and which can be called upon to do their own special thing. To test their own abilities to see if they can operate properly, and then to offer their expert opinion on what action should be taken.