We investigate asymmetries in the behavior of NP subjects and objects in Sipakapense (Maya) across generations of Sipakapense-Spanish bilingual speakers. Often, the two languages investigated are typologically classified into two groups, with SVO word order assumed as the traditional sequence in Spanish, and VSO in Sipakapense. In Sipakapense, SVO is acceptable if the subject is topicalized, and OVS is acceptable with object focus. With SVO order, the verb is marked with an inverse suffix (the focus antipassive) that detransitivizes the verb. A conscious purification of the language is spreading across Mayan communities. Prescriptivist Maya argue that the inverse suffix is obligatory, and that failure to use it with SVO word order is due to contact with Spanish (Pakal B'alam 1994). Barrett (2003) observes that bilingual speakers born in the 1960s-1970s consistently use SVO with Sipakapense, whereas older Maya speakers use VSO. A still younger generation of Sipakapense-Spanish bilinguals (born 1980s-1990s) uses the inverse suffix at the same rate as their grandparents, yet the youngest speakers have a highly variable word order in Sipakapense and frequently use SOV (subject topicalization with object focus). Grandparents use SOV very rarely in quite restricted contexts, while parents use this order in 6% of their utterances. We argue that the key factor in the generational shift to SVO is not directly related to a "dominant" influence of Spanish, but instead is based on independent bilingual strategies. We illustrate that the typological partition is artificial: in Spanish, as in Mayan, VSO can be the "basic" declarative word order, where the subject maintains its internal-VP base position with a neutral interpretation. Like Sipakapense, Spanish SVO occurs from subject topicalization. Declarative order allows a focused subject final, VOS. Since VSO word order is available as a basic sequence in both Spanish and Sipakapense, the attested SVO order occurs as bilinguals "economize" their cognitive load, applying the same operations to similar VSO structures in both languages. A principled explanation exists along the same lines for the youngest generation's adoption of SOV order. Barrett (2003) shows that grandparents elicit higher rates of code-switching (due to army or finca experiences where pidgins developed for communication across Mayan languages), even while parents make conscious efforts to keep Spanish distinct and to speak "puro Maya." From birth, the youngest generation is reprimanded for mixing Spanish and Sipakapense. As we demonstrate, the young bilingual speakers' SOV preference cannot be due to Spanish interference, since SOV order with full NP objects is less common in Spanish. Rather, we claim that another type of bilingual "economy" of resources has taken place. In the face of ambiguous and/or confusing input from the grandparents' and parents' Mayan speech, young bilingual speakers have created a salient contrast between the languages by extending the one form that is possible in Sipakapense, but not widely used in Spanish. In other words, SOV order unambiguously signals Sipakapense, and allows the youngest and most proficient bilinguals to economically differentiate their two languages by introducing a general distinction not utilized by previous generations.