The standard treatment of object pronouns in Latin American Spanish assigns a direct-object function to "lo" and "la" and an indirect-object function to "le." This study challenges this descriptive attribution in light of the contradictory and refractory evidence in Spanish morphosyntax. It is suggested that more detailed research, especially sociolinguistic research relating to Latin American Spanish, might approach the analysis differently. The forms of "le" indeterminacy that seem not to conform to the standard treatment include: cases in which "le" appears as a sociolinguistically determined substitution for "lo," the form standard explanations would dictate; pattern conformity; and structures involving ambiguity or semantic overlapping in addition to these patterns. An alternative analysis viewing certain verbs as homonymic predicates, either communicative (of the "decir" class) or agentive (of the "hacer" class), is proposed and illustrated, and it is suggested that structural symmetry is a confounding pressure in these cases. It is also proposed that "le" functions as an unmarked third-person pronoun and grammatically acceptable compromise. (MSE)