Mike,
I don't want to explain the whole world, just one system that is open. Open means it interacts with other open systems. Interact means it can receive events, unexpected events, for example it learns something new (the market has crashed!), or something random happens (a radiocative atom has decayed). And these events initiate causal chains in my system. So I describe all that with a computer that represents the causal chains, but can also receive input which starts new causal chains. Or with a causal set, which can "learn" or "grow" meaning it can contain many chains and new chains can be added anytime (the market has crashed, hence I must buy gold). I make the chains on the go, as I learn. And one can go further. I can study situations that occurred in the past. What if the market crashes? Well, then ... and this is causal. That's called a Theory! I can make theories for causal chains that do not exist, then create the causal chain and store it. Then, if I learn that something actually happend, I know the consequences and can survivfe better. And yes, as I said many, many times, it is all infinite. There is an infinite variety. Every single sentence you make I have accounted for. Nothing is missing. Nothing is new for me in your letter. "Radically different predictions" I have explained before. It has to do with the butterfly effect in causal sets. Very small difference can lead to big changes. It's a chaotic world. > The real world consists of individual-ly formed, unbounded webs of inconsistent and imperfectly known objects. Of course. So also are my causal sets. >As Bertrand Russell insisted, logic doesn't apply to the real world. I know. He thought it did and tried very hard, and failed. Church and Turing proved him wrong. But that was boolean logic. Nobody is proposing boolean logic anymore. Sergio From: Mike Tintner [mailto:[email protected]] Sent: Friday, June 22, 2012 10:05 AM To: AGI Subject: [agi] Real World Reasoning/Inference Sergio: But algorithms are causal. Computers are causal, our brains are causal..! Sergio, What was the cause[s] of the economic crisis/French revolution/Vietnam war/ ipad's success? - Any historical event whatsoever? What was the cause of your writing your post? What will cause people to buy EI, or Opencog? What will cause any product on the market to be successful? How successful will shale oil be? What will cause any technology to be successful? In the real world, yes, events/effects have causes, and causal actions will produce effects. Unfortunately, events have *infinite* causes ultimately - a worldwide web of causes. And ditto causes have a worldwide web of consequences. We are also typically uncertain as to which causes produced a given effect, or what effects given causes will produce. Not just statistically uncertain, but fundamentally Knight-ianly uncertain. And we are plain ignorant about lots of relevant causes and consequences - and don't factor them in. A great deal of the time we have to do some research and experimentation in order to unearth new causes. That's why in the real world there are almost always radically different, historical explanations of events - there is no such thing as a right explanation, identifying a right set of causes. And there are almost always radically different predictions as to the effects of significant causal actions - and as to the possible sets of effects, - like the future course of the stockmarket . Again there is no such thing as a right prediction. Plus, every example of an event is individual and new and different from every other example - the causes of the last financial crisis were different from that of the '80's and the Depression. So algorithms (and sets) (and logic) don't apply to real world reasoning/inference - the real world isn't uniform like the artificial worlds of logic and algorithms, and can't be similarlyh inferred *every effect/event has an infinite set of causes - it is arbitrary as to which you choose to highlight - there are no algorithmic or logical rules *every effect/event is individual - somewhat different from others - algorithms and logic can only handle a uniform modular world, not the real individual, nodular world *real world events are INCONSISTENT - just because something or someone produced an effect yesterday, doesn't mean it or he will do so today - living creatures often turn round and do the opposite *we always have deeply imperfect knowledge of real world workings Put that more generally, in the real world, there are *NO UNIFORM CLASSES - every class consists of individuals which are exceptions to the rule, - so you can't GENERALISE (without making individual qualfications) *INFINITE WEBS of relationships between objects - you cannot define FINITE SETS of relationships such as causes, comparisons, classifications *INCONSISTENT OBJECTS (incl. both classes and individuals) - you cannot assume CONSISTENCY in objects * IMPERFECT KNOWLEDGE of objects - not PERFECT KNOWLEDGE Logic and algorithms can only work with, and infer from uniform, consistent sets of objects of which we have perfect knowledge (wh. includes statistics) - that's why they're strictly used in artificial worlds and environments only. The real world consists of individual-ly formed, unbounded webs of inconsistent and imperfectly known objects. As Bertrand Russell insisted, logic doesn't apply to the real world. AGI | <https://www.listbox.com/member/archive/303/=now> Archives <https://www.listbox.com/member/archive/rss/303/18883996-f0d58d57> | <https://www.listbox.com/member/?&; ad2> Modify Your Subscription <http://www.listbox.com> ------------------------------------------- AGI Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-c97d2393 Modify Your Subscription: https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-2484a968 Powered by Listbox: http://www.listbox.com
