This thesis presents an in-depth analysis of infants’ acquisition of onomatopoeia – an area of phonological development that until now has been largely overlooked. Infants produce many onomatopoeia in their earliest words, which are often disregarded in phonological analyses owing to their marginal status in adult languages. It is often suggested that onomatopoeia may be easier for infants to learn because of the iconicity that is present in these forms; this corresponds to Imai and Kita’s (2014) ‘sound symbolism bootstrapping hypothesis’, as well as Werner and Kaplan’s theoretical work Symbol Formation (1963). However, neither of these accounts considers the role of phonological development in infants’ acquisition of onomatopoeia. This thesis presents a series of six studies with a range of perspectives on our central research question: is there a role for onomatopoeia in phonological development? Two analyses of longitudinal diary data address the nature of onomatopoeia in early production, while two eye-tracking studies consider the nature of iconicity in onomatopoeia and whether or not this has a perceptual advantage in early development. The role of the caregiver is then considered, with a prosodic analysis of onomatopoeia in infant-directed speech and a longitudinal perspective of the role of onomatopoeia in infant-caregiver interactions. The contributions from thesis are threefold. First, we offer empirical evidence towards an understanding of how onomatopoeia fit within an infant’s wider phonological development, by showing how onomatopoeia facilitate early perception, production and interactions. Second, our results illustrate how these forms are an important aspect of phonological development and should not be overlooked in infant language research, as has often been the case in the development literature. Finally, these findings expand the iconicity research by showing that onomatopoeia do not present an iconic advantage in language learning, as has so often been assumed by theorists in the field.