In the case of child bilingualism, much debate has been generated as to whether speech reflects one single syntactic system or two. Code-switching data from child bilinguals has served both sides of the issue. It has been suggested that code-switching, or language mixing, reflects either a) one syntactic system, with harmonious interaction between each language\'s lexical and morphosyntactic features, or b) evidence of limited ability in two separate systems (cf. De Houwer, 1990; Deuchar and Quay, 1998, 2000; Genesee, 1989; Köppe, 1997; Lanza, 1992, 1997; Lipski, 1998; Paradis and Genesee, 1996; Redlinger and Park, 1980). The controversy is fundamentally one of the have\'s versus the have-not\'s: bilingual children have either one large system at their disposal, or two small systems. Can the accuracy of either theory be determined by examining code-switching data from a child bilingual learning a second language? In this poster presentation, code-switching data is presented from an English-Swedish bilingual learning German in a second-language environment. Code-switching utterances involving the child\'s \"third language\", German, were collected over a period of 6 months (from age 2,8 to 3,4), and analyzed in terms of syntax and lexicon. Should the data have been limited to English-German and/or Swedish-German code-switching, the case for two systems would be supported. However, evidence of three-language code-switching suggests the existence of a single system, where switches to the \"third language\" reveal not limited resources in the other two languages, but rather an awareness of -as well as adventurous exploitation of- a further linguistic possibility.