Malinowski's classic accounts of Trobriand sociality have left anthropology with many lasting conundrums. This two-part article examines two such puzzles revolving around contradictory reports over the agencies involved in magical chants (megwa). On the one hand, consistent with his pragmatic and functionalist theories of language and culture, Malinowski claimed that, although ancestral baloma and other spirits are typically invoked in most spells, those incantations' efficaciousness derived instead from the power of the enunciated words. On the other, as part of his evidence in support of Islanders' "ignorance of physiological paternity," he conceded that spells intended to produce pregnancy in village women were instead expressly aimed at eliciting appropriate ritual actions from baloma spirits as agents of conception and birth. On the basis of ethnographic data recently gathered at Omarakana village interpreted through specific adaptations of the "New Melanesian Ethnography" and Tambiah's earlier "participation" theory of ritual practice, I argue that for Trobrianders the magical power of words is the power of spirits, and vice versa. This insight has important implications for classic and contemporary debates over the nature of "magic," controversies over paternity and so-called "virgin birth," theories of personhood and agency, and the character of dala "matrilineage" relations.