Can syntax appear in a mirror (system)?

Converging evidence indicates that the processing of some aspects related to the phonetic and the semantic components of language is tightly associated with both the perceptual and the motor neural systems. It has been suggested that mirror neurons contribute to language understanding by virtue of a neurophysiological response matching perceptual linguistic information onto corresponding motor plans. This proposal has sometimes been extended to advocate that the language competence as a whole, including syntax, may be ascribed to this kind of perceptuo-motor mappings. This position paper examines what kinds of empirical and theoretical challenges such general mirror neuron language accounts need to face in order to proof their validity - challenges that we think have not been adequately addressed yet. We highlight that the most important limitation is constituted by the fact that some core defining properties of human language, at the phonetic, semantic, and especially at the syntactic level, are not transparent to the bodily senses and thus they cannot be the direct source of mirror neuron perceptuo-motor matching.