Lars P. Syll — Econometrics — rhetorics and reality

The principal objective in teaching Logic 101 is for student to come away with the knowledge of the conditions for a sound argument.

A sound argument is one in which the logical form is valid and the premises are (known to be) true. 

Logical form is matter of rules that are independent of reality. Valid logical form is a necessary but not a sufficient condition for sound argument. 

The truth of the premises is also required that is not and cannot be given by looking at the premiere alone. Premises are models that make existential claims about what is modeled that so the proposition as model can be checked against the assertion or negation that the statement makes about how things stand in the world. If and only if the state of affairs that the proposition models corresponds with fact is the propoistion true. Understanding a descriptive proposition requires knowing what would have to be the case if the proposition is actually true. This enables testing propositions as models of possible states of affairs against what is modeled, that is, putative facts.

Logic and math are about possibility. Empirical testing is about actuality.

The sciences prefer to employ mathematical models instead of that conceptual models wherever possible since they are more rigorous and precise. Rigor comes from following the rules and precision is dependent on accuracy of measurement. A model can be made precise to whatever degree the model builder desires, but the actual degree of precision is limited by the ability to measure accurate. That is an empirical matter.

In a mathematical model, the model must valid, which can be known simply by looking at its construction in terms of definitions, formation rules and transformation rule. In math this is called consistency. It results in syntactical truth. 

The syntactical truth of consistency implies absolutely nothing about the semantic truth of the model as a faithful representation of reality. The output of the model is necessarily true semantically if and only if all the input is true semantically. Then the model necessarily corresponds with the modeled. The necessity involved is logical necessity whereas semantic truth is matter of empirics. Both are required for a scientific hypothesis to be significant as a test of the model. An untested model remains a description of a possible world that is not known to actually exist. 

This is the difference between theory and experiment. For example, physics is divided into theoretical physics, in which theoretical physicists develop models of possible worlds, and experimental physicists design experiments to test these models agains the real world though observation. The creative physicists that apply the resulting knowledge are called engineers and their output is called technology. It is through technology that the public learns pragmatically about advancements in the natural sciences.

Science is about constructing a general description of the world. To be testable, the general must be reducible to the the particular so that hypotheses can be generated that can be compared with the how things actually stand in the world. There are rigorous procedures in scientific method for doing this. 

Failure to conform to this widely accepted understanding of scientific method result in conclusions that are not based on sound argumentation.

While scientific modeling may be extremely complicated and highly technical, anyone that has taken Logic 101 would easily be able to grasp the basic justification behind the process. Of course, only highly trained professionals are capable of rigorously critiquing the application of method, including the empirics, and this is what the peer review process is designed to do. Then, the results are offered for debate and counter-argument in the profession.

How does this compare with the current practice and outcomes of conventional economics?

Lars P. Syll’s Blog
Econometrics — rhetorics and reality
Lars P. Syll | Professor, Malmo University