Syntactic word formation in Northern Sámi By Marit Julien (review)

480 LANGUAGE, VOLUME 76, NUMBER 2 (2000) suggests that much human conceptualization consists of function-like and rather abstract predicates whose relationships with their typical arguments are spelled out in relatively fixed phrases or collocates. For example , 'do, carry out' can be lexically instantiated by a number of verbs, but, in the case of a particular thing which is being carried out, typically only one verb will do. If a blow is being carried out then one talks typically of receiving a blow not of undergoing a blow. As some of the other authors show, this is a problem for those learning a foreign language. Many of these studies show just how common phrasal lexical items are in speech and writing, suggesting that linguistic performance is often a matter of stitching relatively fixed phrases, rather than words, together. They also showjust how very many phrasal lexical items are in a native speaker's repertoire and how subtle the speaker's judgments are as to what their import is and where it is appropriate to use each one in discourse. From a discipline-based perspective, it is also notable how much terminology there is in the field of phraseology, little of it used consistently across the field. [Koenraad Kuiper, University of Canterbury, New Zealand.] Syntactic word formation in Northern Sámi. By Marit Julien. (Troms0 studies in linguistics 16.) Oslo: Novus Press, 1996. Pp. 295. In this revised version of her 1995 thesis, Marit Tulien seeks to refute the lexicalist hypothesis and to amend Mark C. Baker's incorporation approach to the morphosyntax of causatives and passives. In addressing the lexicalist hypothesis, J argues that appropriate analyses of causatives and passives require both syntax and morphology and an intermingling of the two to a certain extent. In particular, she maintains that causatives and passives are due primarily to syntactic processes (including a unification of verbs from different clauses) and that morphology is a sideeffect of such syntax. That is, morphological affixation is a result of syntactic head movement, and morphotactic sequencing often reflects that head movement (à la Baker's mirror principle). J assumes that surface differences among languages are due to parameters and that there should be an essential unity to the underlying structures of passives and of causatives . Thus in establishing a unified class oftrue passives , T distinguishes them from other passive-like constructions. Her critique of the lexicalist hypothesis occupies the first 80 pages of the book, and she chooses to restrict her focus to A. M. DiSciullo and E. Williams 's On the definition of word (Cambridge, MA: The MIT Press, 1987) and S. R. Anderson's A-morphous morphology (Cambridge: Cambridge University Press, 1992). The remainder ofthe book presents a more explicit treatment of the derivations of causatives (Ch. 4) and passives (Ch. 5) within a Chomskyan minimalist framework, along with residual questions (Ch. 6). A two-and-a-halfpage concluding chapter summarizes her findings and abandons the lexicalist hypothesis. In her treatment of causatives, J follows Baker by deriving causative verbs through the joining of two verbal heads (a base verb and an affixal causative verb). The case marking patterns in causatives are then a consequence of the use of appropriate licensing heads. In this chapter J rejects any special case insertion rule or argument fusion, surveys several treatments of double objects, and probes specific problems raised in Chichewa, Kinyarwanda, and Chamorro. Although she maintains a potential separation of direct and indirect causation, she proposes that there is no clear-cut distinction between Baker's Type 1 and Type 2 causatives—she calls for some sort of continuum between the two. In light of the VP-internal subject hypothesis, J argues that passive is a property of VP plus a pro subject with 'properties of a clitic' (219). That is, a (passive) verbal head is to be incorporated as morphology in another verb. Syntactic movement is then determined and restricted by the morphological needs of the incorporated verb. Object case and promotion of the underlying object are then effects of the passive head. The 'clitic' treatment oípro is less satisfying, however, as clitics in her framework require certain mysterious properties (failing to move to subject...