Lately it seems that the number of people arguing against a scientific theory based on false assumptions of how science is supposed to work has increased. In particular, there seems to be an idea that anything that isn’t known 100% isn’t known at all. The people who posit these claims seem to repeat a single idea over and over – that a concept can only be science if you can test it out in a lab and get perfect results every time. Lets discuss, then, the concept of Hard Science vs. Soft Science. Please add your coments and disagreements, and please point to this thread when other threads get derailed by confusion between Truth and Theory. ________________________________________________________________________________________________ Chapter1: Hard sciences are the ones which strictly adhere to the scientific method - basic physics, anatomy, chemistry. In these disciplines, you can come up with a hypothesis, design a repeatable test, walk into your lab, and find out pretty quickly if your idea holds any merit. A classic example (though it took place outside): -Person: Galileo: -Idea: Heavier objects fall faster than lighter ones. -Test theory: Go to a tall place, and drop things of different weights. See if they hit the ground at the same time. -Test reality: take two spheres of drastically different weights, and drop them off the Leaning Tower of Pisa while an assistant stands at the bottom. -Result: Assistant witnessed both spheres hitting at the same time. Reproducing the experiment returned identical results. -Conclusion: Things with different weights (but similar air resistance) fall at the same rate, contrary to common sense. The reason that this sort of “test an example and make a statement about the world at large” is accepted is because of the ongoing effectiveness of “generalization”: the idea that you can take a well understood example and extrapolate that example to a larger set of objects in real life. We are ok with the idea that Galileo’s experiment applies to more than just his own two spheres, and more than just the Leaning Tower of Pisa, for two reasons: 1) Anyone can repeat the experimental theory, and will get the same results. 2) On a larger scale, many other ideas can be described and tested in a similar way. If a man eats nightshade in Spain and dies, and a woman eats nightshade in China and dies, it can be surmised that nightshade is the deadly similarity, and should be avoided. Not deductive proof, but a usefull conclusion! This idea of generalization is key to the concept of inductive reasoning; that the more of something that is witnessed, the higher the probability of that thing bend a trend that is likely to be repeated given the same starting point. It is not as solid as deductive reasoning, since it doesn’t provide 100% certainty in its conclusions, but certainly good enough for daily use. For instance, as a baby, you let go of a crayon. It falls to the floor. You pick up the crayon, and then let it go. It falls again. You can’t deductively determine why the crayon seems to move toward the ground every time (maybe the crayon like the ground more than the sky, or maybe there is an invisible string and an invisible puller); you can *inductively* determine, however, that “crayon + lack of support = fall”. Thus, through generalization, you can gather that “general object + lack of support = fall”. You can then test this hypothesis (a concept that describes your knowledge thus far *and more*) by picking up something other than a crayon and seeing if it drops when you let it go. The reason that this hypothesis/test methodology exists, however, is the very core limitation of the hypothesis stage: you can’t know everything. If we had the ability to go and test every possible situation, then we would simply KNOW things, and science wouldn’t be needed. Since we are not gods, however, we are limited in how much we can test. If we want to verify that “general object + lack of support = fall”, we can determine a fairly high confidence by testing it *a lot*, in as many different situations as we have access to. The more it turns out to be correct, the more confidence we can have in claiming it to be an accurate description of how the world works. After dropping things for a few decades, we are pretty confident that “that which goes up must come down”, and because we aren’t the only ones who have noticed this, we may agree to use the name society have given the phenomenon – gravity. But then we get a job a NASA. One day while shooting things into space, we notice that stuff isn’t coming back every time. Our rule, understood to be universal since our first memories, isn’t being followed! Stupid satellites!! This is where we encounter the choice that seems to be confounding a number of people. Our hypothesis, tested to the point where science would call it a Scientific Theory due to its accuracy, has now been shown to be inaccurate. What should we do? We have three main options (though going out for beer is also an option, it’s not really applicable to the conversation) 1) Abandon our hypothesis and the decades of successful tests we’ve already conducted because the idea isn’t 100% correct, and no longer believe that things fall towards the earth at all. 2) Accept the claim of another person (for instance “objects really like the earth”), and continue to believe our original hypothesis, but now based on different reasons (“that satellite must like the earth, so it is crashed”, instead of just “everything falls”) 3) Take our original idea, and adjust it given what we have now seen. Science would fall into that last group, in case you hadn’t already gathered that. Instead of “general object + lack of support = fall”, the simple addition of location makes everything again agree with what we have seen: “(general object + lack of support) near earth = fall”. Further generalization can then lead us to apply this rule to any celestial object, and to a gradual degree of “fall” based on how “near” the thing in question is to a celestial body. Chapter 2 ________________________________________________________________________________________________ We can never know that some new bit of data isn’t going to come along and make our existing idea no longer 100% correct. But the more data we have that agrees with the idea, the more confidence we have. By taking our confident ideas and adjusting them to fit contradictory evidence, we don’t foolishly throw away years of understanding when we don’t need to. So then, what should we do when presented with a question that cannot be tested in a lab? For instance, how can the effect of 0 gravity be tested 100 years ago, prior to the ability to have a 0G laboratory? Or how do we know how live worked 3 million years ago? Or how do we know what happened to a culture 300 years ago, when there are no written records, and the oral tradition is fragmented at best? Well, at that point, we *can’t* carry out a direct test, but just like we also can’t test every possible gravity experiment, we can use the tests that are possible to generalize hypotheses and increase the level of confidence we have in our ideas. This is commonly known as Soft science. It has provided us with tons of useful information which has later turned out to be accurate, at a rate significantly better than random guessing. It does not follow the scientific method directly, but it is not pseudoscience, as it still tests itself against as many varied sources as possible to avoid bias, and it allows for modification based on new evidence. The levels of confidence involved are much lower than with Hard Science, because we can’t easily test the generalized ideas, but they are orders of magnitude higher than with a simple guess. With a guess, all you have a simple idea thrown at the wall in hopes that it sticks. With Soft Science, you have ideas that have been formulated with the structure in mind, and with a testable set of non-options mapped out. By looking at (and testing) the results of what you can’t test directly, you can infer information about the originating causes based on (again) inductive reasoning. Until 100 years ago, we couldn’t send things into space, but we could study the orbits of the planets. We could also determine that when something on earth is thrown into the air, the point between “up” and “down” is comparable to “not falling”. We could (and we did), with confidence levels higher than pure guesses, study these items, and generalize the concepts out to 0G as a whole. We then also make predictions with the hope that someday, a direct test will be possible. We hope that at some point, Mir will exist, and a test can be performed to verify the hypotheses we have created. Once those tests are run, the scientists alive to run them can take our idea and change it as needed, now that Hard Science can be applied to the question. Similarly, we can’t go back in time and see what dinosaurs looked like, but we can look at the results of their existence. We can have a high-level of confidence that dinosaur bones are the results of living dinosaurs because of induction as well – fossilization of known living things occurs (petrified trees), and bones are left by known living things when they die. The similarities in structure between dinosaur fossils and modern animal bones allows us to conclude with a very high certainty, that they are the result of the death of a thing that was very much like and animal, but not identical to any that currently lives today. We can then look at the context of where those bones were found, and not finding any modern human remains, or any modern deer bones, infer that dinosaurs died in a way that didn’t involve people or deer. Since geological strata where people are found, near the top, don’t include dinosaur bones, and dinosaurs aren’t alive today, we can infer that the way dinosaurs died (that didn’t involve people or deer) was that they lived during a time when people and deer weren’t running around – farther down in the strata. This idea gains traction when we look at partially petrified trees; we can gauge their age using non-geological methods, and determine that down tends to equal old. Also similarly, we can look at human stories, and determine, independent of their content, the confidence level of the story’s accuracy. A book labeled as “fiction”, and found in the children’s section of a library has a much lower chance of accurately reflecting the true nature of the physical world. We have this confidence due to its context, even before we read the book - again, through induction and generalization. If a story is found in a temple, and it talks about how the God of the temple is the most powerful of all Gods, we can have a fairly low expectation of its accuracy, because of its context. The people who would be writing that story clearly have a vested interest in that version of the story, and so our confidence level in the hypothesis of it being propaganda instead of true fact is much too high; solely because of the context. Finding the same story in the temple of an opposing god would be fascinating, however – it’s context would suggest a vested interest in NOT writing such a story down. Our confidence in a forgery would go against the evidence at hand, and new theories would have to be determined. Chapter 3: Conclusion: ________________________________________________________________________________________________ Science, and inductive reasoning, can never be 100% sure in it’s claims, because we as humans are not Gods – we can’t know everything about a subject before we form an idea about it. However, we equally can’t withhold any decision-making until after we become gods; daily survival requires us to make choices at any given moment. We can either simply listen to what another person tells us (our teachers, our parents, our preachers, our gang leaders, our cult messiahs, our Richard Dawkins), or we can figure things out for ourselves. Science provides us with the ability to determine things about the world that we would normally not be able to; due to human limitations. But just as Soft Science deals with items that are untestable in a way that improves it’s accuracy over simple guesses, we don’t have the time to repeat every test made by every scientist over the last few hundred years. Thankfully, the soft science of understanding scientific writing can be used to form a level of confidence in those tests, and by repeating certain key experiments, we can increase our level of confidence in the overall field of scientific theory while we add to it. The past few thousand years, since logic was formalized, approaching life this way has changed our world, in both good and bad ways. But no one can claim that it didn’t change how quickly we as a species have progressed in the world. Logic has moved us from small-time farmers to space-travelers. From people who believe in stories to people who understand the world around them. From people who believe in something with an completely unknown and unstudied level of confidence, to people who know what they don’t know, and have an idea about what they do. We may find ourselves with more unanswered questions than those who answer every unknown with “God did it”; but at least we can tell you if an answer could be found, and predict with some accuracy where it would be found. At least our answers allow self-reliance. At least our answers drive further study, instead of complacency. At least our answers are willing to learn along with us.