in some languages. 50 Here is yet an-,)ration and Antipassive, further jus-section that the distribution of Anti-0un Incorporation over a wide rang~ unted for by analyzing Antipassive 10, thereby making it subject to the )Ies. Superficial differences between Illow from the fact that the former is a compounding root, together with a correlates of this distinction. This 2 rule of Antipassive in the grammar n languages with Antipassivization or absence of such a rule; rather it is : exists a lexical item with particular _~ that is both an argumental N and an .issives follow from the general prin-Verb Incorporation In the last chapter we studied in detail constructions in which a single mor-phologically complex word does the work of two words in English: noun-verb combinations which count as both the verb and the (head of the) direct object of their clauses. I argued that these were the result of XO movement, which adjoins the head noun of a noun phrase to the verb between D-structure and S-structure. This process is simultaneously morphological and syntactic: syntactic in that its distribution and its consequences for the structure are determined by syntactic principles involving government, X-bar theory, and case theory; morphological in that the resulting [N + V] structure is morphologically and phonologically indistinguishable from normal compounds or derived verbs in the language. In this chapter, we turn to another construction in which a single, mor-phologically complex word corresponds to two words in the English coun-terparts: namely, morphological causatives. In these constructions, a single verb corresponds not to a verb and a noun, but rather to two verbs. This possibility, together with Noun Incorporation, is the second major element of polysynthesis. Here again, we will find strong evidence that the forms are actually syntactically derived from two independent verbs by move-ment. Thus, causatives are VERB INCORPORATION (VI), directly parallel to Noun Incorporation and subject to exactly the same principles. One con-clusion of this will be that explicit rules are unnecessary to account for the properties of this class of GF changing processes as well. 4.1 CAUSATIVE CONSTRUCTIONS AS VERB INCORPORATION Consider the following causative paradigms from English and Chichewa (Bantu): (1) a. Bill made his sister leave before the movie started. b. The goat made me break my mother's favorite vase. 147148 Verb Incorporation (2) a. Mtsikana ana-chit-its-a kuti mtsuko u-gw-e. girl AGR-do-make-AsP that waterpot AGR-fall-AsP 'The girl made the water pot fall. ' b. Aphunzitsi athu ana-chit-its-a kuti mbuzi zi-dy-e udzu. teachers our AGR-do-make-AsP that goats AGR-eat-AsP grass 'Our teachers made the goats eat the grass.' (3) a. Mtsikana anau-gw-ets-a mtsuko. girl AGR-fall-made-AsP waterpot 'The girl made the waterpot fall.' b. Catherine ana-kolol-ets-a mwana wake Catherine AGR-harvest-made-AsP child her chimanga. corn 'Catherine made her child harvest corn.' (Trithart (1977» The English sentences in (I) are biclausal in all respects. In particular, they are biclausal in meaning, with an embedded clause appearing as a semantic argument of the causative predicate in the main clause. For each of the two clauses, there is a distinct morphological verb, as one would expect. The Chichewa sentences in (2) are similar; they correspond to their English glosses lexical item for lexical item and phrase for phrase. However, Chichewa has another way of expressing these notions, illustrated in (3). These sentences contain only one verb each, which happens to be morpho-logically complex. Nevertheless, sentences like those in (3) can be the-matic paraphrases of those in (2). Thus, the same thematic roles relate the same verb roots to the same Noun Phrases in (2a) and (3a). Furthermore, the sentences in (3) are as biclausal in meaning as their English glosses, even though they look monoclausal morphologically. In this sense, the verb forms in (3) "do the work" of two verbs, thereby presenting another case of apparent mismatch between morphology and syntax. This is the mOf-phological causative construction, the most famous of such mismatches. Unlike noun incorporation, this topic has been subject to long and complex discussion in generative linguistics. I The guiding assumptions set down in chapter 2 determine the heart of an analysis for this construction. For concreteness, let us focus on (3a). Here it is the waterpot that breaks, and the girl who is responsible for that event taking place. Thus, the same theta role assignments occur in (3a) as in (2a). The Uniformity of Theta Assignment Hypothesis therefore says that (3a) and (2a) should have parallel D-structures. This implies a D-structure approximately like (4) (details omitted): (4) 4.1 Causative Constructions as Verb Incorporation s ~ NP VP I~ girl V s I A -its NP VP 'make' I I waterpot V I -gw-'fall' 149 ~ext, the causative affix -its and the verb root -gw- clearly combine into a ~mgle word at some stage. Thus we are led to an analysis in which a lexical Item undergoes syntactic movement to combine with another lexical item in the structure. By the Projection Principle, this movement may not de-stroy thematically relevant structure; hence, the moved verb root must leave a trace to allow theta role assignment to the "stranded" subject and to head the embedded clausal complement which the causative morpheme lexically selects. The S-structure of (3a) must therefore be approximately: (5) s ~ NP VP I~ ~rl V s A A V V NP VP I I I I gWj -its waterpot V I tj Thus, I claim that morphological causatives are (at this level of abstrac-tio.n) exactly like Noun Incorporation, except for the category of the word bemg mov~d. Morphological causatives are Verb Incorporation. The claIm that m~rphological causatives are derived by movement may seem less controverSial to some when I point out its strong similarities to !he claim that "subject-to-subject raising" is derived by movement, famil-Iar fr~m Chomsky (1981). Raising verbs like seem systematically appear in two dIfferent S-structure configurations:150 Verb Incorporation (6) a. It seems that Sara adores Brussels sprouts. b. Sara seems to adore Brussels sprouts. Since these two sentences are "thematic· paraphrases," in that the same