This study, part of an on-going investigation, analyzes the syntactic aspects of code-switching. A series of empirical studies has confirmed that there are only two general linguistic constraints where code-switching may occur, the free morpheme constraint and the equivalence constraint. This study describes in formal terms how the two constraints determine the way two monolingual grammars may be combined in generating discourse containing code-switches. The data were selected from recordings of 20 individuals including both balanced bilinguals and speakers who are fluent in Spanish but not in English. The data, already analyzed in a previous study (Poplack, 1979), were reanalyzed within a formal grammatical framework based on the direct generation of surface phrase structures by a context-free grammar. The analysis sketches surface grammars for the Spanish and English spoken in the corpus, as well as for the code-switching mode. The analysis also studies the syntactic aspects of code-switching heard in the speech of Puerto Rican bilinguals. One point that emerged is that code-switching constraints are surface phenomena and that phrase structure grammars for each language can be combined to form a code-switching grammar which generates grammatical monolingual and code-switching sentences. (AMH)