The objective of the article is to identify different formulations of indeterminacy thesis and various trials of dealing with it. Indeterminacy thesis was first identified by Quine as a consequence of his theory of meaning. According to the thesis, although we can determine the kind of facts or events of the world, which a particular observation sentence type refers to, we are not able to determine the referents of words the sentence consists of. And that is because of the fact that on the basis of data we normally posses in the process of interpretation we can construct indefinitely many equally good transactions manuals (due to any one of them a reference of a particular word type can be something different than due to any other). In the article four variants of indeterminacy thesis will be identified and the relations between them will be showed. The resulting thesis will be that if we do not want to accept ad hoc and improbable semantic assumptions that at the start undermine the indeterminacy thesis, we should reduce three first variants of the thesis into the fourth one. The last one, however, is such that essentially seems to resist virtually any argument against indeterminacy thesis (granting the general assumptions of semantic theories of which it is a consequence).