Sunday, June 13, 2010

Truth and its Consequences

History is a beast. In our ancient past, what people believed to be true often turned out to be incorrect or only partially right. Over the last three thousand years or so, our view of our world has been radically altered. The shape of our world, the solar system, zero, negative numbers, chemistry, biology, genes, DNA, determinism, computers, etc.

We've undergone a massive change in our knowledge, one that has been gradually shaping how we see the world around us. Although one doubts that we've really 'evolved' in such a short time, we've definitely civilized ourselves. Truth, it seems, gradually has an effect on us, as it sinks in generation after generation. It takes a while -- knowledge is sticky, and people are resistant -- but as we know more, and more people come and go, our thoughts and actions gradually get more sophisticated.

Still, we've reached that point were we are surrounded by so much knowledge, and so much of it is of dubious quality, that we seem to be lost in what we know. Unable to assimilate or deal with it. Once we made it so easily available in large qualities, the side-effect appears to be a form of snow-blindness.

What the world really needs now is a truth-telling machine. One that when given a statement will determine whether or not the statement is true. And not, as one might expect, true in the present context, but instead 'universally true'. That is, true for all time, and for all places. This type of invention would allow us to once again size up what we know, show the falsehoods for what they are, and then collectively move on to a better place. It's a great idea. Crazy. But still great.

In a universal sense, while there can be an infinite number of incorrect answers there can only be one correct one. That perspective, however, is far too black and white for our universe, and how we live in it. It makes more sense to see some of the false answers, as being tied to a specific context. That is, "the sun always rises in the east" is a valid statement for the contextual period in which both the planet, and the sun still exist. In a universal sense, some day that statement won't be true, either the east won't exist, the sun won't exist or the relationship between the two planets will have been altered. The statement then, is not a universal truth, it is just contextually true today, as far as we know.

Most people however, don't want to deal with the world around them and their knowledge on a universal/geological scale. Most people, it seems would prefer that their own 'universal' truths be limited to their lifetimes, or even shorter periods. That, however is most likely the type of tunnel vision that has continued to plague our understandings. People dig in, and don't want to know the truth, or at least more of it, so the old 'truths' linger like bad odors, until the last of a generation die off. Given all of the generational overlaps, collectively what we know, and what we believe is a mis-mash of all of these overlapping incorrect versions of our world. To fix things in a way that allows us to move farther ahead, faster, we have to fix these stagnation problems first.

The problem with universal truth is that we can't possibly know what that it is. We start out in a context, and are trapped by it. Only in hindsight can we see that something isn't as we believed, and that's only if we willing to accept that it might be wrong.

Given all of the major discoveries and turmoil in history that have gone on, we can see that we generally grow our understandings in leaps and bounds. There are clearly times were generations stagnate or go backwards, and then there are times full of great discovery and change. Our societies ebb and flow, as the different organizational or cultural issues push or suppress us into examining our world.

And within our growth, everything we know is dependent on a huge underlying foundation. Each fact sits on a sea of other facts. We can only build new upper layers when we've managed to get discover enough of the lower ones. Knowledge, it seems, is the ultimate house of cards. That relationship, and the fact that there are always missing supporting cards, defines the frailness of what we know. There will always be assumptions, so there will rarely beuniversal truth.

Even if we can't have universal truth, one of our more recent discoveries, computers, do give us an unprecedented opportunity. Computers are stupid, but they beat elephants for having a long memory. That, and they are never bored, so you can just set them up to do something, seemingly for ever. Memory and patience open up the ability for us to do something fascinating. While we can't determine the universality of what we know, we can however use computers to chart out the degrees of freedom in what we say. That is, we can take all of our statements, one by one, and examine how and why they might very.

A really sophisticated program would be nice, but in truth something far cruder might still reach the same basic level of understanding. We'd just like to know for any very large collection of statements where the degrees of freedom lie. So, for my example "the sun always rises in the east", it would be nice to know that the sun is a planet. That we are assumed to live on another planet, and that east is a relative pointer based on the relationship between the two planets, and that planets have a finite lifespan. That is, for the very simple knowledge in that statement, what are all of the related things on which that statement depends?

We can't know if something is universally true, but we can figure out absolutely all of the places on which the statement is dependent. The truthfulness of a statement isn't a black and white variable, but rather a complex graph of underlying relationships on which the overall quality of a statement is formed. It's the structure of those relationships that ultimately sets the data quality. If most of the underlying dependencies are true, then the statement can been seen as mostly correct (and still not be the truth, for example "political spin").

In that sense, if we had a huge database of a massive number of relationships, we could add in a new statement and quickly see how likely that the quality of that statement might vary. We could test any new knowledge statement against the sum total of what's already there. We could easily see where the weakness and assumptions lie.

As miraculous as this sounds, domain experts do this all of the time. They can hear a statement and based on their internal understandings, come first to an intuitive belief that the statement is unlikely to be true. It may require further digging to confirm it, but human beings are quite capable of seeing how something doesn't quite fit an overall pattern. We can match what we hear up against what we know. We've effective pattern matchers. And it's not just the experts.

For whatever reason we've been inundated the last few decades with a huge amount of scientific research. The newspapers and TV are forever announcing a vast array of new discoveries. Most of these discoveries are probably fine, but the media has a tendency to want to announce the more exotic or dramatic sounding results. Because of this, it is inevitable that some less scrupulous researchers will bend their results to say things entirely for the point of achieving notoriety. We expect this, and we see it often.

Sometimes scientific discoveries are counter-intuitive, that does happen. However, we see far too many 'surprising' results, many of which get huge media coverage but then just disappear into the shadows. Mostly these examples of 'bad science' are obvious if you spend some time contemplating them. We're able to see how unlikely they are, even if we're not experts. They just don't fit. There may be a few false negatives, but mostly if its covered in the media, and it sounds off, it probably is.

Understanding our current knowledge is good, but our biggest problems come from where we are headed. Even if we know what we know, and we know that it is 'true enough', we still can't correctly analyse statements about what might possibly happen in the future. But it is these questions which cause us the most anxiety.

People are forever trying to change the world, and some of these changes are downright stupid and dangerous. Too see this in advance, we'd have to build and examine sophisticated models of our circumstances, then apply the changes. Intrinsically these models would have some likelihood of being incorrect when probed or extended. Nothing short of a full simulation of our existing universe would ever be complex enough to be entirely synchronised with our would.

Still given these limitations, the models only need to be accurate enough to given us an indication of the success or failure of the changes. And it is far easier to build such a reasonable model if we fully understand the variability underneath. The mechanics of the model match the underlying degrees of freedom. That is, we can only model what we understand, and we can only simulate it along the lines of the were we think it might change. But if we are working from a full structural view of the model's statement dependencies, then at least we know the fullest extent to which the model may vary. We know its domain.

Another thing we would get from these structures was a sense of the missing cards. There is always some intrinsic symmetry and pattern occurring because of the nature of our physical world. It may be ultra-complex like fractals, but we're still able to pick up on the patterns in nature where and when they occur. In examining the relationships, there will no doubt be places where the violation of a pattern or some lack of symmetry is noticeably obvious. Places that have been overlooked or not fully examined. Bringing these to the forefront will provide direction for researchers to investigate morethoroughly . We pretty much rely on inspiration for this now, but this would insure that what we know gets more fully well-rounded, that the small ignored gaps get noticed and filled. These holes may seem insignificant butgenerally they propagate upwards and distort the upper layers. Ultimately this slows down progression until we're lucky enough to have some genius come along and re-align the basis correctly. And that just takes too much time.

We have a fantastical amount of knowledge, but we really don't understand it well. It is openly available and has inundated our lives, but hasn't really improved them. Given that we really can't trust a lot of what we hear, or what we read, something, anything, that would give us a sense of the underlying truthfulness would be a huge boon to our society. It would allow us to contain the explosion of information, and turn it into something useful. Dependable. It would allow us to examine the frailness of our understandings. With so much information, and so much of it wrong, all we are going to do is make our population more resistant to change. We're foolishly forcing ourselves to have to ignore what is going on around us. To ignore what we know. We're not utilizing our own collective intelligence.