During the last years there has been an increasing interest in examining the brain responses to word order variations. In one ERP study conducted in Spanish, Casado, Martin-Loeches, Munoz, and Fernandez-Frias (2005) had participants read Spanish transitive sentences with either an SVO (subject-verb-object) or an OVS order. The word order of a sentence was determined by semantic information (semantic condition) or by syntactic information (syntactic condition). We will focus on the syntactic condition, which suffers from a number of serious pitfalls. By using both linguistic evidence and data from off-line rating studies, we will first show that the authors are wrong in claiming that the OVS sentences in the syntactic condition are grammatical. Secondly, we will show that these sentences are not interpreted as OVS sentences but as SVO sentences. In light of these pitfalls, we conclude that their results in the syntactic condition do not inform issues concerning brain responses to word order variations, but of the detection and repair of an ungrammatical string. (Contains 1 footnote.)