We investigated the extent to which parents' prosocial talk and negations relate to the quantity and diversity of parents' spatial language production. We also examined similar associations among children. Participants included 51 children of ages 4-7 years and their parents recruited from South Florida. Most of the dyads included mothers and were Hispanic and bilingual. Dyads constructed a Lego house for 10 min. Sessions were transcribed and coded for instances of parent prosocial talk (praises, reflective statements, and behavior descriptions), child general positive statements (all positive contributions to the interaction), and parent and child negations (criticisms, corrections, and disapprovals) using the Dyadic Parent-Child Interaction Coding System. Transcripts were also coded for quantity and diversity of spatial language including shape terms (e.g., "square"), dimensional adjectives (e.g., "little"), orientations (e.g., "turn"), locations (e.g., "middle"), and spatial features/properties (e.g., "edge"). Parents' prosocial language, but not negations, were significantly associated with the quantity and diversity of parents' spatial language. Children's general positive statements were significantly associated with children's spatial language quantity. Exploratory data analyses also revealed significant associations between parent-child talk about shapes, dimensions, and spatial features and properties. Findings suggest that variability in parent-child prosocial and spatial talk during collaborative spatial play relates to aspects of their own--and each other's--spatial language production.