Michael Arbib's book proposes a scenario of language evolution that begins with pantomime, progresses to proto-sign, and then develops together with proto-speech in an “expanding spiral” to create a language-ready brain. The richness of detail in Arbib's hypothesis makes serious appraisal of each of its aspects possible. Here I describe findings about established and emerging sign languages that bear specifically upon the interaction between sign and speech proposed in the Mirror System Hypothesis. While supporting the central role that Arbib attributes to gestural/visual communication in understanding language and its evolution, I point out some kinks in the spiral that potentially disrupt its smooth expansion. One is the fact that each modality relies on an entirely different motor system. Another is the type of relation that holds between the articulators and grammatical structure, which is radically different in each system as well. A third kink disrupts the proposed continuity between holistic pantomime (gestural holophrases) and signs. Given such differences, instead of a scenario in which speech grew out of sign, it seems more likely that the two modalities complemented each other symbiotically throughout evolution as they do today. If so, then the modern ability to spontaneously create sign languages reveals the extraordinary richness and plasticity of human cognition, and not an evolutionary stepping stone to speech.