This contribution examines how early Arabic linguistic thought moves from the collection of language data to their deployment as structured argument. Rather than treating Classical Arabic linguistic scholarship as descriptive repositories of linguistic material, it explores how scholars transform diverse sources into epistemic evidence for linguistic argument. Central to this process are the main corpora that recur throughout the tradition: the Qurʾān, kalām al-ʿarab (“the speech of the Arabs”), pre-Islamic poetry, and, more selectively, the ḥadīṯ tradition.

The analysis focuses on how these materials are transmitted, but also actively filtered, categorized, and mobilized within argumentation. Particular attention is given to internal variation within kalām al-ʿarab, including varietal differences and speaker-based distinctions, as well as to the interpretive practices through which individual attestations are assessed and rendered usable for linguistic theory. The paper ultimately discusses how variation is not removed but managed, and how linguistic knowledge emerges through the interaction between transmitted material and analytical frameworks.