This dissertation is a study of the strategies employed in the indigenous languages of North America for distinguishing grammatical subjects from grammatical objects. The degree to which intransitive subjects, transitive subjects, and direct objects are the same or different in their form is a property known as alignment. An interesting feature of alignment is that the sameness or difference of subjects and objects can vary according to certain grammatical conditions. This phenomenon, known as split-alignment, has been the focus of linguistic investigation from a range of theoretical perspectives since the 1970s. It is frequently remarked that some grammatical conditions are more likely than others to induce split-alignment. In this study, I survey fifteen indigenous North American languages in order to (a) determine all the grammatical conditions responsible for splitting alignment systems in these languages; and (b) identify the factors contributing to the preponderance of certain conditions over others. The fifteen languages are selected at random in a manner which controls for genetic and areal affiliation. I determine that at least thirteen split-inducing conditions are present in these languages. The most common are status of a subject or object as a speech act participant, verbal semantics, lexical specification, and grammatical number. I interpret the skewing toward these conditions by appealing to four generalizations which highlight those factors involved in the common, and several less common, conditions. Split-inducing conditions which do not immediately fall out from these generalizations are tentatively treated in diachronic or other language-specific terms.