This is a longitudinal study of development in coordinated mother–infant vocal interaction from 4 to 10 months (N = 41) focusing on the development of turn-taking patterns and time spent in coordinated vocal interaction. Data analyses were carried out using multistate analysis. Both mothers and infants were found to coordinate their own vocalizations with their partners’ vocalizations. The infants had a predominant pattern of covocalizing, whereas the mothers had a predominant turn-taking pattern at all ages (4, 7, and 10 months). However, a significant reduction in the duration of covocalization was found between 4 and 7 months, due to less covocalizing compared to turn-taking from the infants. In addition, time spent in coordinated vocal interaction increased significantly between 4 and 7 months and a development for the infants was found from repeated vocalizations toward single vocalizations between maternal turns. Taken together, these findings indicate a developmental process of fine-tuning and increasing the turn-taking format for vocal communication, characterizing adult communication, before the development of speech. In addition, our results indicate that this development starts earlier and is more prolonged than has been hypothesized from previous research.