The second book of Aristotle’s Organon, the Peri Hermeneias (Fī al-‘Ibāra), deals with theories of simple and compound utterances, of propositions formed from them, and of the mutual logical implications of such propositions. Generally speaking, though topics such as homonymy and intention are discussed in this book, language is treated not as a system of communication, but as divested of particular meanings and as symbolic for logical operations. The Arabo-Islamic tradition of logic in the pre-modern period retains this attitude to theories of language for much of its career; and this makes sense, given that the subject matter of logic was understood to be second order concepts, which need only have an accidental relation to ordinary language. In the seventeenth century, however, a fascinating transformation in the treatment of logical utterances takes place in the Indian logic text Sullam al-‘ulūm. The author of this text, Muḥibballāh al-Bihārī (d. 1707), and his commentators and glossators expend a great deal of energy presenting logical propositions in terms of theories of language developed in the field of poetics and uṣūl al-fiqh. In other words, they invest logical propositions with characteristics of ordinary communicative language—such as metaphor and allusion—that do not allow for the classical rules of logical implication; and this development, in turn, makes the entire structure of formal logic vulnerable. This paper examines the theory of language presented in this pre-modern text and the strategies that were devised to neutralize the effects of this theory on the larger project of formal logic.