The following article, in the form of the author’s accepted manuscript with page numbers inserted in double brackets, has been published as (2018): Factual vs. evidential? - The past tense forms of spoken Khalkha Mongolian. In: Ad Foolen, Helen de Hoop & Gijs Mulder (eds.) Evidence for evidentiality. Amsterdam: Benjamins: 45-75. https://doi.org/10.1075/hcp.61 The article is under the copyright of John Benjamins Publishing Company, so for any permission to re-use or reprint the material in any form the publisher should be contacted. [[45]] Factual vs. evidential? - The past tense forms of spoken Khalkha Mongolian * Benjamin Brosig Academia Sinica Past tense forms of spoken Khalkha Mongolian distinguish between established (-sang) and non-established knowledge, which is then either based on direct (-laa) or indirect (-jee) evidence. Time of acquisition thus determines whether information source is marked, though vivid recollection (-laa) and deferred realization (-jee) overrule it. Conversely, attempted recollection in questions (-l=uu) doesn’t presuppose sensory perception. A fourth suffix (-v) is used if well-established events still surprise the speaker. These suffixes may also be used in a discontinuous fashion to refer to the future and then modally qualify predictions as inevitable (-sang), apprehended [but preventable] (-v), based on sensory evidence (-laa) or inferred (-jee). The distinction between unsourced -sang and sourced -laa/-jee is thus not about factual stance, but codes the extent to which information is consolidated in memory. Keywords: direct evidence, discontinuous tense, indirect evidence, mirative, preventive, time of acquisition * I’m grateful to Bettina Zeisler, Dolgor Guntsetseg, Östen Dahl, Henrik Bergqvist, Lamont Antieau and Maria Koptjevskaja-Tamm for comments on the draft of this chapter, to all my informants for their enormous contribution, and to all who supported me during this stay and later on. My research was financed by Stockholm University as my employer and the Sven och Dagmar Saléns stiftelse who funded my 7-month stay as a visiting researcher at the National University of Mongolia in Ulaanbaatar with 53270 SEK.