How do we know what we know?

    Everything we know about the world comes from one source: observations.  They may be as systematic as a laboratory science, or they may be totally anecdotal and unorganized.  They may be personal observations, or they may come from the observations of others who may—or, admittedly, may not—have more expertise than we do.  Nonetheless, everything ultimately comes from some form of direct observation.

  We do, of course, also learn things by making extrapolations and interpolations from data.  Mathematics and logic do play roles in learning about nature, and we regularly make inferences—not all of them justified—from what we have observed.  We also make completely non-rational conclusions about the world, and many of these are correct; the activity of the right brain is inherently non-linear and non-rational, like throwing a crossword puzzle into the air thousands of times a minute until it lands in the correct order.  For any of these non-direct conclusions to be reliable, however, they have to be testable by observed data.  Without such testing, they are useless, dogmatic proclamations.

  We can work through an example of how that process.  I see you standing in a field with a stone in your hand; you let go of the stone and it falls to the ground.  I learned from this that, at one point in time you held this object, let it go, and it fell.  Technically, that is all I have learned, that it happened once.  Now, if I see you do it several times, and if, every time, the stone falls to the ground when you let it go, then I can conclude that when you let go of the stone, it will fall to the ground.  I am curious, so I get other people to do the same thing and, wonder of wonders, every single time someone holds a stone and lets it go, it falls to the ground!  Thinking I might be stumbling upon something profound, I get people to hold different objects, not just stones, and observe further that when any object is released, it falls to the ground.  I may not know the cause, but I do see what happens: objects that are released from the hand, or any other way they are supported, fall to the ground when that support is no longer there.  From this I deduce a general rule that when objects are dropped they fall to the ground.

   Here is an example of how I can learn from a direct observation and, as well, from an extrapolation from that data.  If we look closely, however, that “knowledge” I obtained from the extrapolation is not as reliable as the information I obtained from the first observation.  First, I have not, and cannot, observe every single incidence of something being dropped; it is at least possible that my universal conclusion is incorrect.  Furthermore, not only are my data incomplete, but I could interpret the data incorrectly.  Suppose in my “experiment” I asked you to hold and drop a rock, and other person to hold and drop a leaf at the same time.  Sure, I observe that they both fall to the ground as they had in the past, but I also note that the rock hits the ground before the leaf does.  Furthermore, whenever a leaf and a rock are dropped together, the rock seems to fall faster.

   What do I conclude from this?  I might conclude correctly that it is the shape of the leaf that is causing it fall more slowly—air resistance—or I might conclude, incorrectly, that it is weight that cause the difference, that heavier objects fall faster than light ones.  This is, in fact, the error that people made, and this error was not corrected until Galileo performed his famous experiments at the tower of Pisa.  If I do not have you drop two stones of the same shape but different weights at the same time, I am likely to draw the erroneous conclusion.  So, we have to admit these two caveats about empirical observation:  a direct observation can be wrong if I did not see what I thought I saw, but conclusions drawn from data can be wrong for several other reasons as well.  Direct observation will only be incorrect if I observed incorrectly, but extrapolations can be incorrect if any of the following occur: if I observe incorrectly the data will be wrong, if I extrapolate too without sufficient data, or if I make a failure in logic and reason improperly.

   Suppose I did guess the right reason. I am still likely to draw an improper conclusion, because I might be tempted make the conclusion more universal than it really is.  Unless I performed the experiment in a zero gravity environment, or in a vacuum with no air resistance, or with objects that would be subject to the same resistance, like two identically shaped objects,  I am unlikely to make the proper connections, and come up with a wrong explanation about the data.  A key part of explaining the data is to understand the parameters of the experiment, and limiting the conclusion to the variables of the experiment.  Yes, the conclusion will be limited, but the more limited, the more reliable.

   As I mentioned earlier, we cannot ignore the role that non-rational insight provides in advancing knowledge.  Collected data may well tell us that a conclusion at which we arrived earlier was incomplete, or even incorrect, but the data will not provide the new conclusion.  That comes from insight, the “aha” factor.  We examine the data, and in a moment of insight, guess—and at this point it is still a guess—a new explanation.  In religion we may assume that it was the gods, or the spirits, that provided this new explanation, but empirical research does not work that way.   Primitive people did not think that one would question or test the gods, but modern man does.  We don’t take this insight on faith; we test it.  We call this insight an hypothesis, and subject it to repeat testing. 

   Most hypotheses will fail the experiments devised to test them, and, even when they pass, the results are published so that others may test them.  Alternative explanations for the data are welcomed, not condemned, so that all possibilities can be investigated.  There is, unfortunately, always the possibilities of honest error, tunnel vision and, sadly enough, outright fraud.  The antidote, of course, is the kind of transparency that science considers the ideal.  No question is ever considered to be finally and completely answered.  On all matters, our conclusions are always tentative, always incomplete, and always open to correction.  There is no final, absolute truth as far as we are concerned, only better and more accurate conclusions.  Common sense will tell us, then, that if we want to avoid looking like fools, we should avoid making absolute statements outside of the realm of mathematics and formal logic.

   This all means, of course, that we have to regard the world as purely a natural phenomenon.  If there is no room for faith, or any form of divine revelation, then we can make no statements about things that cannot be observed.  Since we cannot empirically affirm there is any type of deity, we leave the question alone.  If we want to know the nature of reality, we turn, not to religion, and not to philosophy, but to physics.  Physics—and the other sciences—tell us what the world and life is, and the role of philosophy is to interpret those findings for their relevance to our lives.  In other words, we should understand nature, and human life, in the terms described by the best science, and then we should turn to philosophy to explain what those facts mean for our lives, for how we live. At this point all science can tell us is that we are an animal, more intelligent than others, but not inherently different, living on a planet that exists because of purely natural events.  Science gives us the facts; philosophy gives us the significance of those facts.

---