The findings for adverbs and adverbial phrases in a naturalistic corpus of Miami Haitian Creole–English code-switching show that one language, Haitian Creole, asymmetrically supplies the grammatical frame while the other language, English, asymmetrically supplies mixed lexical categories like adverbs. Traces of code-switching with an English frame and Haitian Creole lexical categories suggest that code-switching is abstractly BIDIRECTIONAL. A quantitative methodology that codes the language-indexation of the token in addition to the surrounding lexical items was used for all mixed (e.g. xYx/yXy, xYy/yXx, yYx/xXy) and unmixed (xXx/yYy) adverbs. Discourse position, especially the left-periphery, is found to be a significant factor in adverb code-switching. Sociolinguistic and psycholinguistic analyses which acknowledge the ‘low’ status of one language and the ‘high’ status of the other explain better the frequency of mixed English adverbs in a Haitian Creole frame and the rarity of mixed Haitian Creole adverbs in an English frame than a minimalist approach, such as MacSwan's (1999 and subsequent work), which uses phi-feature valuation and entails asymmetry without bidirectionality. While I provide confirmation for Myers-Scotton's (1993) Matrix Language Frame approach, I emphasize that trace bidirectional data need to be accounted for by a theory that is grounded in the sociolinguistic and psycholinguistic realities.