Drawing on a broad range of primary sources in Ottoman Turkish, Arabic, and English, such as governmental correspondence, memoirs, and journal articles this paper explores Ottoman practices of knowledge production and transmission in the Province of Yemen between the establishment of this province in 1872 and the end of Ottoman rule in early 1919. Historians of the late Ottoman period from the beginning of the Tanzimat in 1839 to the end of empire have equated the production of official knowledge about the political, social, cultural, and economic realities in the different regions of the empire (including Yemen) with acts of writing – be it that population figures were recorded in census documents or the “manners and customs” of local people detailed in the form of petitions, memoranda or book-length accounts. These documents, so the argument, were archived and thus became repositories of knowledge on which the Ottoman central government would draw for its efforts to build a modern, more intrusive state. While scholars have shown that orality retained a prominent role in the daily practices of Ottoman governance, these studies concentrate on the 1840s and 1850s and do not look at knowledge production. Building on this body of scholarship, I argue that during the period under study the spoken word was of crucial importance in the context of Ottoman knowledge production and transmission on Yemen. For example, governmental correspondence and the memoirs of governors-general such as Mahmud Nedim or Mehmed Tevfik suggest that it was common practice for the central government to interview individuals, deemed knowledgeable about Yemen, on issues, such as the influence of the Zaydi imams, without recording in writing the information they provided. Similarly, officials about to deploy to Yemen for the first time relied primarily on conversations with colleagues, who had served there before, to acquire local knowledge on a wide range of topics. This began to change during the early 1900s, when the near collapse of Ottoman rule in Yemen in 1905, the end of the Hamidian regime in 1908-09, and new notions of professionalism prompted officials to view existing practices of knowledge production and transmission as partly responsible for the government’s failure to control Yemen. Now, they increasingly associated knowledge with published accounts, while considering unrecorded knowledge unstable and, hence, unreliable. In introducing greater complexity into our understanding of Ottoman practices of knowledge production my paper adds to a growing literature on late Ottoman governance.