Kramer / Tröbs / Kastenholz (Hg.) Afrikanische Sprachen im Fokus, Köln 2011, S. 257-270 S-AUX-O-V-X in Manding and Mande 1 Henning Schreiber Introduction Since the beginning of Mande linguistics (MacBrair 1837: 31) the syntax of Mande languages has attracted a great deal of attention. A wide range of poly- functional morphemes and the poor inflectional morphology coincide with a spe- cific word-order type. This sub-type of SOV languages that challenge Greenberg‟s (1963) word-order generalizations is given the label „S-AUX-O-V-X‟ or „type B‟ in Heine 1976. Since the latter was published this word-order type required expla- nation and attracted much attention: Heine & Reh (1984), Claudi (1994), Gensler (1997) and later Bearth (1995, 2009), Kastenholz (2003, 2006), Creissels (1997), Nikitina (2009) and Tröbs (2009) and others. It is generally assumed that S-AUX- O-V-X type is a result of a word-order change in Mande and that the distinct syn- tactic slot-patterns and morphology of perfective and im perfective go back to asymmetric grammaticalisation. While for the imperfective the so-called predicate markers in AUX position are analysed as grammaticalise locative copulas, it is assumed that the predicate markers for the perfective go back to postpositions (Bird & Kendall 1986; Creissels 1997). However, in this paper, it is argued that, although both grammaticalisation and historical syntax are directly related to each other, they should be treated sepa- rately: while grammaticalisation theory delivers plausible arguments for change both at the empirical level and in the cognitive realm, historical syntax detects changes that may have occurred in the history of the language. In order to recon- struct previous stages of syntactic structure, actual change(s) need to be detected. Historical syntax makes wide use of grammaticalisation theory but also seeks to distinguish between plausible change and change that can actually be recon- structed. In other words, while the typological disposition of a proto-language 1 I would like to thank all the people who have commented earlier drafts of this paper, and in par- ticular Theda Schumann and Joe McIntyre for their help.