PSEUDO_EMPIRICISM Another way to look at the problem of BS trying to pass for “empiricism”, or “evidence-based science”.
– When one discusses events related to a casino roulette wheel, one can easily make a distinction between realizations and the property of the roulette wheel. Anyone can see that it is highly irrational to make comments and build theories on past history, except *in their relation to* the properties of the roulette table –otherwise, one is fooled by randomness, or committing the gambler’s fallacy.
– When discussing a given history of gambles and making claims, we know *exactly* whether we are referring to the *outcomes* from the roulette wheel or to the roulette wheel.
– Now in real life, when we talk about *crime*, or incidence of *ebola*, or such things, alas, we use the same word “crime” or “victims” to refer to both the realizations (i.e., the history of the process) and the *generator* or what here is equivalent to the roulette wheel. Verbalistic effects cause us to conflate the two.
– But when we write down things mathematically, we clearly see that one is called “random variable”, the other is called “realization” so mathematically the two are separate. When we discuss series of random variables we talk about “stochastic process”. Analytical claims are always made about the random variable or the process, not the realization.
– What I have just illustrated seems simple. But, alas, many people make fallacies and we have to fight journalistic imbeciles and social scientists conflating past history of, say Ebola with its properties. (More technically, the history of Ebola is NOT an empirical claim about its properties, except that they correspond if and only if the process has thin tails.)
PS- Which brings me to the value of mathematics: its value lies less in the computation, the numbers, than in the use of clear-cut definitions that avoid sloppiness of language. I see mathematics, particularly in probability, close to legal theory, where everything is as explicitly defined as possible. When things are well defined, their relations (or lack of) become immediable visible.
via: Facebook
