In this paper, we examine the emphatic focus doubling construction in American Sign Language (ASL) and Brazilian Sign Language (Libras), in which one element of the sentence appears in its base-generated position within the sentence and one copy appears in sentence-final position. We review the existing focus doubling data in the literature, as well as a previous syntactic analysis of the construction that we think is the best available option on the market (Nunes and Quadros 2005). Diverging minimally from this analysis however, we propose that movement of the focused element proceed not to the head of an emphatic focus projection, but rather through the specifier of that projection; this modification nicely precludes the need for excorporation and c-command out of a dominating non-terminal node. We then examine an asymmetry between focus doubling in Libras vs. ASL, namely that doubling is permitted in indirect questions in the former but not the latter, an asymmetry not addressed by Nunes and Quadros. We suggest that there is a ban on multiple instances of focus-driven movement in ASL, and briefly discuss how a striking parallel with restrictions on multiple foci in Modern Greek may ultimately hold the answer to resolving the asymmetry, at the same time raising interesting questions about the way that information structure maps onto phonology and syntax in different languages.