Argumentation or illustration of theoretical ideas by referring to “wrong” (*X) or “doubtful”
(?X, ??X) constructions has long been an ordinary element of linguistic research. “Incorrectness”
or “doubtfulness” of the constructions is usually not justified, being considered obvious
to any competent native speaker. This practice is based on the postulate of equivalence
(or at least of absence of significant differences) of competent speakers’ linguistic intuitions.
Another issue that this research addresses is the problem of correct sampling for a population.
This is one of the most important problems in experimental sciences, but it is often underestimated
by linguists, who by default rely on the premise that all native speakers have the same
linguistic intuitions and, therefore, are interchangeable in linguistic experiments. This study
provides strong evidence against this assumption, basing on assessments of correctness of
some semantic constructions made by different groups of native speakers. The authors show
that gender, level of education, and occupation are factors that have to be taken into account
in sampling linguistic experiments based on a comparison of native speakers of different languages
or first and second language speakers. Finally, the article outlines possible directions
for a fully-fledged experimental study of linguistic competence.