Re: Social Science, Science, and Empirical Study

[Luke]  Your first sentence makes it sound as if prediction is an almost subjectiv[ist] enterprise; it sounds as if – per my own conversation with Francesco Ranci both on and off WSN – the predictor’s ability and methologic range is very much shaped by his/her own mental operations; and thus the basic criteria used to make these predictions and to choose categories/criteria out of which one’s experimentation/results is based is actually a function of the analyst’s own mental activity in designing the experiment.

 

[Mike]  That wasn't my intent.  I was merely saying that the prediction must carry with it some idea of its precision so that it can be tested. 

 

[Luke]  I have my own reservation about what you’re saying regarding prediction and exact “certainties.”  Aren’t all the predictions we make as scientists actually probabilities?  If Science has mapped a precise set of relationships for weather and so many other kinds of phenomena on earth and in the cosmos at large and scientists truly understand these laws, then why hasn’t Theoretical Physics (for example) finally been able to put together a Grand Unification Theory (GUT) from such precise knowledge? 

 

[Mike]  Yes they are all probabilitities, but extremely high probabilities  we often call certainties.  The fact that mathematical descriptions for heat and mass and energy transport (the processes underlying weather and much else) have been formulated doesn't mean that this has been done for all fields in physics.  A GUT hasn't been found yet because nobody has found it.

 

Complexity and Chaos theory are misundertood by many people.  Weather cannot be accurately predicted because of the non-linear nature of the expressions describing transport phenomenon. Unlike linear equaltions, non linear equaltions can "blow up".  That is, a small discrepancy at the initial condition rapidly leads to ever larger discrepancies as time goes by.  An example is prediction of the position of a billiard ball after n ricochets off the walls of the table.  Given precise  initial direction and velocity of the ball we can quite accurately predict the path of the ball after the first ricochet.  Indeed pool players routinely do this.  After several ricochets we no longer have an accurate projection of the path of the ball, even wiht extremely accurate initial conditions.  After 20 bounces the ball could be anywhere on the table, we have absolutely no idea where the ball will be.

 

Even if we had "perfect" knowledge we cannot know the exact initial conditions of the ball because of the Heisenberg uncertainty principle.  Hence it is theoretically impossible to have any deterministic idea where the ball will be after some number of bounces.  The best we can do is may a statistical prediction.  This despite the fact that we know the exact mathematical laws that govern the ball's movement.  The complete failure to use these exactly-known laws to make a useful prediction is a direct result of the nonlinear nature of the mathematical equations that describe billiard ball motion.  This phenomenon is what is known as chaos. 

 

[Luke] The point I was driving at in drawing my distinction was that use tests and/or better yet “natural experiments” (of the kind mentioned in Jared Diamond’s Guns, Germs, and Steel are really different from strict laboratory experimentation, the experimentation and empirical methodology par excellence. 

 

[Mike]  I don't see the difference (other than convenience).  Laboratory experimentation is fast.  This is why so much science has been done in the lab.  Consider, rather than waiting to see oif a prediciton comes true (which can take years for each experimental iteration) or searching for years to find the necessary data to conduct a "natural experiment" along the lines Diamond discusses, in the lab we can artificially produce the conditions that allow us to test our prediction and still get home in time for dinner!  Is it really so surprising that so much science has been conducted in a laboratory? 

 

Non-laboratory sciences progress more slowly (except for a few like astromony which has a whole universe from which to pull natural experiments).

 

[Luke] I’m wondering; what’s your opinion of this notion that in order for something to be truly empirical and scientific, then it has to have lab work as its central operating principle?

 

[Mike]  Not so.  In fact laboratory science is rather recent.  Astronomy is one of the oldest sciences and it is not laboratory-based.  Newton's laws were based on astronomical data and Galileo's observations of projectile motion.  The painstaking measurements of stellar positions undertaken by Tycho Brahe and others (before the invention of the telescope) surely are empirical data--yet involved no laboratory.

 

[Luke] That is, in terms of the Stock Market examples we’ve been tossing around, there’s the level at which we’re using economic analytical tools (like Elliott Waves and Japanese Candlesticks) and the patterning of the Market itself to make detailed probability-predictions [and looking for specific results that match our particularized projection criteria].  That is the first level of empirical meaning I mentioned. 

 

[Mike] Elliot waves and Japanese Candlesticks aren't economic analytical tools.  They belong to what is called "technical analysis" which is a kind of stock market astrology.  Some market analysts employ astrology.

 

[Luke]  The second meaning of “empirical” would involve our consideration of the systematic Market process itself (and all aspects of its specified detail) in its natural state regardless of our observations/predictions as analysts.

 

[Mike]  I don't understand this.  What is the "systematic Market process itself" and how is this any different (in a useful sense to people who actually particpate in markets) from our observations, predictions and actions (i.e. to buy or sell).  The market IS the observations, predictions (beliefs  about future movements) and the actions taken by the people involved in the market.