Explanation in linguistics

Over centuries, the concept of explanation was unclear. Several approaches, such as the type of teleological explanation (i.e. the attempt to explain something by its purpose) failed in the end because they could be shown to be logically flawed.

The only known logically perfect kind of explanation is the deductive-nomological one: An explanation is the subsumption of a fact under a set of laws and boundary conditions (gr. nomos - "law"). Thus, the specific fact can be logically deduced from the laws and boundary conditions.

This can be illustrated by the "Hempel-Oppenheim Scheme" (Carl Hempel 1965a):

G1, G2, ..., Gn

R1, R2, ..., Rm,

_____________

E

The part above the line represents a set of universal laws and a set of boundary conditions and is called the "explanans", the line represents the logical deduction, and "E" represents the fact, the "explanandum", which is being explained.

The explanation is valid only if the fact E is covered by at least one of the laws and if the relevant boundary conditions are met. Thus, the fact that closes are, on the average, shorter in sentences with more clauses than in sentences with less clauses is explained by the Menzerath-Altmann Law and the boundary condition that clauses are immediate constituents of a sentence.

A special kind of the deductive-nomological explanation is the functional explanation (not to confuse with teleological explanation), which is the appropriate form of explanation in linguistics. The simple variant, represented by the Hempel-Oppenheim Scheme, or attempts to explain linguistic facts and phenomena using a cause-effect relation must fail. A functional explanation is much more complicated (cf. Hempel 1965b and Köhler 1968). Its logic is, to some extent, similar to biological evolution.