This research aimed to describe the patterns of language development and frequency of codeswitching in Pakistani heritage children. Receptive and expressive language skills of 167 participants aged between 2;6 and 7;2 were sampled in their home language, Mirpuri, Punjabi or Urdu. In terms of MLU, Mirpuri children developed language in the same way as their monolingual English peers with comparable scores for children aged 3;0 to 4;5. Expressive language samples contained a high frequency of intrasentential codeswitching with a stable mean of between forty and fifty percent of multiword utterances. Data were examined using the Matrix Language Frame (MLF) and 4-M models (Myers-Scotton 2006, 2002, 1993) and found generally to conform to adult codeswitching patterns. Noun insertion was more frequent than verb insertion. A monolingual frame language was employed for over ninety eight percent of utterances, even when English verbs were incorporated into Mirpuri utterances. A Mirpuri compound verb structure consisting of noun plus an operator was a common site for codeswitching. First described by Romaine (1986) in adult Panjabi codeswitching, the Mirpuri noun was often replaced by an inserted English noun or verb. Forty five novel codeswitched compound verbs were found in the data, nineteen formed from an English noun plus a Mirpuri operator, and twenty six from an English stem verb plus operator. The large number of codeswitched compound verbs suggests that this structure is highly creative and novel verbs were not examples of borrowings. These findings challenge surface constraint codeswitching models and support the MLF and 4-M models. EAL acquisition was examined. Codeswitching from the home language into English was rare and EAL patterns contrasted with monolingual English acquisition patterns. Three case studies of children presenting with specific language impairment confirmed the hypothesis that codeswitching patterns may assist in identifying SLI in a child from this community.