, 2.2) used as 1. purpose clause 2. complement clause of ? desiderative verbs (e.g. has-'want') ? manipulative verbs (V of assistance only; e.g. kaa'll-'help') ? modal verbs (e.g. dand-'can, be able, may') ? phasal verbs (e.g. ke'-'start') ? achievement verbs (e.g. akeek-'try') ? commentative verbs (e.g. makk-'be suitable
, 3) used as 1. purpose clause 2. complement clause of ? utterance verbs (xa'mm-'ask') ? propositional attitude verbs (e.g. amma'nn-'believe') ? knowledge and perception verbs (dag-'know', xuud-'see') ? manipulative verbs (e.g. ass
, A grammar of Sidaama, 2000.
, A grammar of Supyire: Das K'abeena. Deskriptive Grammatik einer hochlandostkuschitischen Sprache, 1994.
DOI : 10.1515/9783110883053
, A linguistic geography of Africa, pp.228-249, 2008.
, When 'say' is not say. The functional versatility of the Bantu quotative marker ti with special reference to Shona Reported discourse. A meeting ground for different linguistic domains, pp.253-287, 2002.
, Grammar and meaning: A semantic approach to English grammar, 1995.
, A grammar of Sidaama (Sidamo), a Cushitic language of Ethiopia, 2007.
, Kambaatissa xifati seerrata [Rules of writing the Kambaata language], Liiranso Wotango & Tiitoos Heegana, Duuraame: Educational Department of the Kambaata-Alaaba-Xambaaro Zone, 1992.
, All the same? The emergence of complementizers in Bislama, 2002.
, Language typology and syntactic description Complex constructions, pp.52-150, 2007.
, Exploring the periphery of the Central Ethiopian Linguistic Area: Data from Yemsa and Benchnon, Language contact and language change in Ethiopia. Cologne: Rüdiger Köppe, pp.59-81, 2009.
, A typology of purpose clauses, 2009.
DOI : 10.1075/tsl.88
, A grammar of Alaaba, a Highland East Cushitic language of Ethiopia, 2007.
, Predicate conjoining in Hadiyya. A Head-Driven PS grammar, 1989.
, Adverbial clauses Language typology and syntactic description Complex constructions, pp.237-300, 2007.
, The Dhaasanac language, Cologne: Rüdiger Köppe, 2001.
, A grammar of Kambaata Part 1: Phonology, morphology, and non-verbal predication Cologne: Rüdiger Köppe. ?, 2008b: Relativization in Kambaata from a typological point of view Interaction of morphology and syntax: Case studies in Afroasiatic, pp.161-206, 2008.
, Shared multifunctionality in Ethiopian languages: A case study of 'like'. Paper presented at the Journée scientifique Afro-Asiatic languages and cultures. Cambridge: Cambridge Scholars Publishing. ?, forthcoming b. Switch-reference in Ethiopian languages, Highland East Cushitic. In: Ghil'ad Zuckermann Journal of Language Contact, 2011.
, ???Dire??? et finalit?? en bedja : Un cas de grammaticalisation, Journal of African Languages and Linguistics, vol.25, issue.2, pp.133-53, 2004.
DOI : 10.1515/jall.2004.25.2.149
, The verb in Cushitic, 1975.
, The second consonant of a glottal stop-sonorant cluster is generally written as double, although the cluster only consists of two phonemes, e.g. <'mm> = /?m/; this convention helps to distinguish these clusters from glottalised sonorants, which are written <'r> and <'l>. Word-final unstressed /i/ does not occur orthographically, irrespective of its phonological status. 6 The only counterexample is in fact an erroneous example published in Treis 189, ex. 567) Instead of the SS purposive fanqalóta 'so that I return' (SS), the example should have included the DS purposive fanqálunta 'so that I return' (DS) (information by Deginet Wotango 7 If the subject of a converb clause and an adjacent clause are not entirely coreferent (part-whole or inclusion relationship), the DS converb form is used (Treis forthc. b). 8 Switch reference-sensitive converbs (see Table 1) are even more important referent-tracking devices in Kambaata because they have a higher frequency than purposives. 9 In the translations, focussed constituents are printed in small caps. 10 The morphemes are usually not broken up in the glosses of the examples. 11 With the exception of affirmative imperatives and verbal nouns. 12 For details on object marking on verbs see Treis 13 The morpheme =tannée(ha) originates from the feminine dative demonstrative pronoun tannée(ha) 'for this one (f)' (Treis 2008a: ch. 9.3) 14 For other reason clause constructions see Treis (2008b: 195ff). 15 The subject agreement morphemes are closest to the verbal stem: 1s/3m ? , 2s/3f -t, 3hon -een, 1p -n, and 2p -teen (see the initial morphemes in all columns of Table 3) They are often subject to morphophonological processes (assimilation, metathesis and epenthesis); see e.g. (17) where -t-un 3fJUS is realised as -s-un after the verb stem le'-is-'make grow'. 16 The morpheme is reduced to -u in the 1s jussive. 17 The demonstrative ta underwent multiple grammaticalisations in Kambaata (Treis 2008a: 128ff, 410) 18 The affirmative imperatives have no person morphemes and their endings are not related to those of the jussive. The singular form ends in an unaccented, devoiced i; see ít[ i ] 'Eat!' The 2pIMP ends in -é (triggering palatalisation and/or gemination on single stem-final vowels) or in -yyé (if the stem ends in a geminate consonant or a cluster); see e.g. ichch-é 'Eat (p)!' 19 Sim (1989) treats jussive and purposive verb forms under the label " subjunctive " . The DS purposive ( " subjunctive 2 " ) is identical to the jussive. 20 Cf. Sim (1989: 153) referring to comparative work on the Cushitic verb by 21 The list of verb classes taking purposive complements may not be exhaustive. The labels for the different types of complement-taking verbs (or predicates) are taken from Noonan (2007). 22 I cannot exclude that dative-marked VNs may be negated periphrastically with hoog-'not do'. However, no periphrastically negated VNs in purpose clauses could so far be found in the corpus. 23 Irrespective of whether it is used in deontic or epistemic function, the verb dand-'can, be able; may' permits dative-marked VN and purposive complement clauses. 24 Note that phasal verbs like ins-'start' and jammar-'start' (< Amharic) usually govern accusative-marked VNs; see e.g. áayyee yú (mACC) insitóo'u 'they started to mourn' (lit. 'they start to say aayyee'). 25 there is a noun gara 'way, manner' in Sidaama, a HEC language closely related to Kambaata. 26 This simplifies the matter because there are exceptions to this rule: some verb forms (progressive verbs and some forms of the e-perfective paradigm) do not undergo stress shift; in these cases, the relative forms are indistinguishable from the main verb forms) I had written that relative-based purpose clauses can be marked for different verbal aspects, ) quotes one example from the Gur language Supyire The Kambaata data is written in the official orthography (to which I added accents to indicate the position of phonemic stress The following graphemes are not in accordance with the IPA conventions Treis) its modifiers (incl. negative relative verbs) end in o/-ua (masculine oblique). 30 This tendency still requires exact quantification. 31 The use of the 3fPOSS 'her' rather than the 1sPOSS 'my' in this direct speech complement is probably an error (in other quotative constructions of the corpus the pronouns are not adapted to the viewpoint of the speaker of the entire sentence but the viewpoint of the, pp.76-338, 0200.