ABSTRACT Background Language treatment for bilinguals with aphasia has been shown to result in gains in both the treated language and the untreated language (i.e., cross-language generalization). However, cross-language generalization is not consistently found. This inconsistency may be due to several factors, such as the age of acquisition of, and proficiency in, each language. One often-overlooked factor that may influence whether cross-language generalization occurs is the manner in which bilinguals learned their second language (L2): in a formal educational context (explicitly) or naturalistically through exposure to the language (implicitly). Prior research suggests that implicit L2 learning results in greater overlap in the representation of the first language (L1) and L2 in the brain, particularly for grammar, compared to explicit learning. In contrast, lexical processing in L1 and L2 is proposed to rely more on shared brain regions regardless of the manner of L2 acquisition. Greater overlap should provide a greater likelihood of cross-language generalization effects from treatment. Aims The goal of this study was to determine how the manner of acquisition of L2 may affect cross-language generalization following treatment in L2 separately targeting the lexicon (object naming) and grammar (sentence construction). Methods & Procedures Two Spanish-English bilinguals with aphasia each completed two treatment phases in English of 2-4 weeks each, in succession, with one week between them: semantic feature analysis (SFA) targeting object naming, and Verb Network Strengthening Treatment (VNeST) targeting verb retrieval and sentence construction. Pre- and post-treatment assessments and weekly probes were completed for each phase. Participant P1 learned English explicitly in an educational setting, while participant P2 learned English implicitly. Outcomes & Results As predicted, P2 showed cross-language generalization after verb/grammar treatment (VNeST) whereas P1 did not. However, contrary to the prediction that both participants should show cross-language generalization after noun treatment (SFA), only P1 showed cross-language generalization of object naming. Conclusions Cross-language generalization was observed for both participants but for different aspects of language. The findings suggest that naturalistic second language learning may lead to stronger links between languages in the grammatical system, whereas formal second language learning may lead to stronger links between languages in the lexical-semantic system. Future research should further explore the effects of manner of acquisition as a predictor of language co-activation and cross-language treatment generalization in bilinguals with aphasia.