Developmental theories of intonation are typically based on the idea that linguistic pitch patterns reflect biological or emotional “natural tendencies”. Accordingly, these theories predict that intonation will develop along with the infant's earliest speech-like vocalizations. This article examines the empirical and theoretical basis of the claim that children master intonation before the onset of speech. As a background for studying prosodic development, Section 2 describes three components of intonation: (1) intonation-groups, (2) forms of intonation, and (3) tonal meanings. Section 3 presents criteria for evaluating children's development in each of these areas. Section 4 focuses on studies of children's development of intonation from 3 to 25 months of age, with special attention to the grammatical, pragmatic, and affective/attitudinal functions of pitch patterns. Most studies were not designed to meet the criteria specified in Section 3. Although this makes any conclusions very tentative at present, our review suggests that children control some core features of intonation before they produce two-word combinations— but not before they produce their first words. Register seems to be the first component of intonation to stabilize in the single-word period. Except for pre-intentional associations between pitch level and affect, natural tendencies are not sufficient to account for children's development of intonation.