Selected data on the acquisition of German by a 3-year-old Spanish-speaking child (Andy) are presented and compared to other data on the acquisition of Spanish by this same child. Because Andy's productions show a dominance of Spanish over German, it is suggested to view the German data as second language phonological acquisition. In a traditional framework, some phenomena related to consonants in Andy's German pronunciation could be interpreted as transfer, whereas other phenomena affecting vowels could be accounted for by means of developmental processes. These results contradict Hecht and Mulford's (1982) hypothesis on L2 phonological acquisition, according to which transfer should have more weight in the acquisition of vowels and developmental processes in the acquisition of consonants. A solution is provided in terms of a principles-and-parameters view of phonological acquisition, assuming a nonlinear view of phonological structure. The parameters relating to the syllable and to the principle spreading phonological features are specially considered. By the time Andy begins acquiring German, the syllabic parameters have already been set according to L1, as shown by data on epenthesis, whereas the principle spreading vocalic features has not been parametrized yet and exposure to German input delays its setting, as shown by frequent cases of vowel harmony in German as well as in Spanish. An explanation for the nonsimultaneous fixation of these two parameters is attempted.