This chapter will look at how affective and grammatical information is encoded temporally and spatially on the human face, and argue that the temporal dynamics (i.e., how facial expressions unfold over time) and spatial frequencies (i.e., the resolution of visual features such as fine or coarse details) are crucial for disambiguating affective and grammatical facial expressions, particularly in contexts where emotions and grammatical markers overlap in sign languages. Furthermore, in addition to facial expressions, body language provides critical emotional information, often complementing or enhancing facial cues. In sign language, especially, the body plays a crucial role in conveying emotion, and it provides an additional channel of information that helps disambiguate emotional content, especially when facial cues may be subtle or ambiguous. This computational perspective on how the brain must solve the multimodal functions of facial expression in sign language interactions will be based upon the analytic approach proposed by Marr and Poggio (1976). This approach breaks a cognitive problem down into three distinct levels of description. First, is the computational description: what is it that the brain is required to compute, and why? Second is the algorithmic description: how does the brain perform those computations and what kinds of representations must it create and manipulate to do so? Then there is the implementational description: how is this algorithmic computation implemented in the human brain?