This project analyzes the dialectal variation of insubordination, the phenomenon whereby a formally subordinate or dependent clause is conventionally used as a main or independent clause (Evans 2007). The empirical focus is on the functional load of que-insubordinate clauses, which are the most frequent insubordinate constructions in Spanish. Que-initial sentences can display several functions in interaction: third-person imperatives, wishes, negative evaluation, echo, disagreement, topic change, etc. Considering their high polyfunctionality, this project addresses two interrelated questions. The first research question (RQ1) concerns the historical dimension of insubordination: how is the diachronic development of insubordinate constructions to be modelled precisely? To answer this question this project builds on the assumption that synchronic variation reflects diachronic evolution (Hopper & Traugott 2003: 1). The second research question (RQ2) deals with the functional load of insubordination: if insubordinate constructions express more than one meaning, on what basis then should it be decided whether we are dealing with one (polysemous) construction or with more than one construction, each expressing a different meaning (homonymy)? In order to properly respond to these questions, the project is based is based on a quantitative and qualitative corpus analysis adopting an interactional constructional approach to language. The analysis will be performed on different corpora that include spoken spontaneous conversations of the main dialectal areas of European and American Spanish. The objective of this project is twofold. From a descriptive point of view, the project attempts to map the situated meanings of independent complement constructions and describe their distribution across selected varieties of Spanish. At a theoretical level and along the lines of the overall Construction-Grammatical approach adopted here, the project aims to model the situated meanings extracted from the corpora in the form of constructional networks, which acknowledge both generalizations and low-level patterns, as well as potential dialectal specificities. This should allow us to extend the framework of Construction Grammar, so that it can accommodate variational data. Moreover, since the networks resulting from the analysis represent synchronic variation of complement insubordination in Spanish, they can shed light on the current debate regarding the diachronic evolution of insubordination. Finally, the project investigates what types of contextual information are required to fully explain the use of these constructions in interaction (conditions on precedent and subsequent discourse, discourse-structural information, information structure, amongst other) and how this information should be formally represented.