Arnold D J (doug@essex.ac.uk)
Sun, 18 Dec 94 13:38:42 GMT
Mike_Maxwell wonders about why ?My aunt is difficult to try to persuade anyone to visit. is odd, suggests it could be something semantic, and wonders who has talked about it before. Various people have proposed the related idea that tough-constructions are some kind of complex predicate: Nanni, in her thesis, and the paper below (the idea is also taken up in Chomsky '81 (LGB)), Keenan and Timberlake, Montalbeti et al, and no doubt others. At the end of this message is some stuff I wrote about the idea it a few years ago, when I did not think much of the idea. Now I am not so sure (Louisa Sadler and I have a paper the most recent Journal of Linguistics which argues that prenominal adjective form sort of lexical constructions with the head nouns (i.e. "tall person" is sort of lexical, rather than being an N', or whatever), and one characteristic we suggest may be related is that the A+N must be some sort of `natural' predicate. Of course it is characteristic of tough adjectives that they can appear prenominally with their complements (as in "a hard to please person"), which most adjectives do not do. So I would probably have to rethink my position on all this (but life is short!)). I hope this is useful Doug Arnold Anyway, here are the references: %A M. Montalbetti %A M. Saito %A L. Travis %D 1982 %T Three Ways to Get Tough %J Papers from the Regional Meeting of the Chicago Linguistic Society %V 18 %P 348-366 %A E.L. Keenan %A A. Timberlake %D 1988 %T Natural Language Motivations for Extending Categorial Grammar %E R.T. Oehrle, E. Bach and D. Wheeler %B Categorial Grammars and Natural Language Structures %I D. Reidel Publishing Co %C Dordrecht %P 265-295 %A D. Nanni %D 1978 %T The \fIeasy\fP Class of Adjectives in English %I PhD Thesis, Dept of Linguistics, UMass %C Amherst, Mass %A D. Nanni %D 1980 %T On the Surface Syntax of Constructions with \fIEasy\fP-type Adjectives %J Language %V 56 %P 568-581 And here is some discussion I wrote a few years ago: (TC = tough construction) The `complex predicate' view has been proposed in Nanni (1978, 1980; for at least some kind of TC), Chomsky (1981), Montalbeti et al (1982), and Keenan & Timberlake (1988). The idea is that there is some sort of lexical or restructuring process which takes an expression such as hard to pronounce, and forms a complex predicate out of it. Evidence in favour of it includes the following. Nanni notes that tough predi- cates can appear pre- nominally, as in (0a), a position which is otherwise reserved for adjectives (i.e. without complements, as opposed to APs). (0) a hard to pronounce name. Nanni explicitly limits her predicate formation process to TCs where the propositional complement does not itself con- tain further embedding. Chomsky (1981:308-314), and follow- ing work (e.g. Montalbetti et al 1982) has generalised this to account for properties of the TC in general, such as the exclusion of overt subjects from the propositional comple- ment in TCs, discussed above, and the alternation tough predicates allow between expletive It subjects (in the It- Extraposition structures), and the full NP subjects of TC (in GB terms this alternation suggests that subject of tough adjectives both is, and is not a theta position, which is a contradiction). Chomsky also suggests that this approach provides an account for the following data. (1) a. It is hard to play these sonatas on this violin. b. These sonatas are hard to play [] on this violin. c. This violin is hard to play theseisonatas on [] . b'. Which violin are these sonatas hard to play [] on [] ? c'. *Which sonatas is this violin ihard to play [] ion [] . i j i j (1a) is a non-TC, as a reference to clarify what is going on. (1b,c) indicate that either of the complements of play may undergo Tough Movement (i.e. may become the subject of a TC), without any difference in acceptability. However, if a further extraction occurs from the complement of hard, then a difference emerges. Tough Movement of these sonatas does not prevent further extraction (1b'), but Tough Movement of this violin does (1c'). It seems that in TCs, only items that are `more peripheral' than the understood position can be extracted. Chomsky's suggestion is that this is because TCs involve complex predicates, so non-peripheral extrac- tions are blocked in the same way as exctractions from ordi- nary lexical items. Roughly, (1c') is bad because it involves a derivation like the following: (2) this violin is hard-to-play-these-sonatas-on [] . j ------------ j <--------------------------| However, there are several problems with the `complex predi- cate' approach. First, there are reasons for thinking that expressions like hard to pronounce are phrasal, not lexical in any way. For example: (i) however to is analysed, its presence is reminiscent of phrasal, rather than lexical structure. (ii) If expressions like hard to pronounce are adjectives, one would expect them to take part in lexical processes such as adverb formation. However, there are no adverbs like easy to pronounce+ly (cf easily).[NOTE1] (iii) The adjectival `Specifier' Enough always follows the adjective it modifies. If hard-to-pronounce is an adjective, one would expect to get the ungrammatical (3a), rather than (3b). (3) a. *a hard-to-pronounce enough name. b. a hard enough to pronounce name. Second, if all TCs are to be regarded in this way, then the apparent unboundednes of `Tough Movement', is problematic both technically and conceptually, since it raises the pros- pect of predicates of unbounded length. Third, and finally, the problems that the approach is intended to solve can be solved in other ways. For example, in the case of the res- triction on `peripheral extraction' in TCs, an alternative view is that it is `crossing dependencies' that are excluded: when the dependencies `nest', as in (4a)=(1b') the result is acceptable, where they cross, or intersect, as in (4b)=(1c') the result is unacceptable (cf Fodor 1978, Kaplan & Bresnan 1982): (4) a. Which violin are these sonatas hard to play [] on [] ? | j | i |i |j | ------------------- | ----------------------------------------------- b. *Which sonatas is this violin hard to play [] on [] . | i | j | |j | -------------------|----- ----------------------------------------- [NOTE1] Such a adverb would mean, perhaps, `in an easy to pro- nounce manner', as in `He handled the difficult name in an easy to pronounce manner'. Nanni suggests (1980:575) that adjectives like easy_to_pronounce do take part in some lexical processes, such as the formation of com- pound verbs: alongside [ [ interesting] [ sounding ]], we also get hard to pronounce sounding. VThe only account I am able to give of these would parallel that of `compounds' such as: (i) (He gave me) an I-am-so-angry-that-I-could-cheerfully-strangle-you sort of look. An additional problem for Nanni's formulation is the possibility of `phrasal' verbs in the `Hard Nut' con- struction: (ii) a difficult to talk about problem. This is problematic because Nanni's formulation means that the complex predicate formation rule is blocked if there is anything after the verb. Notice that in a case like (ii) it is not possible to argue that verb and preposition form a constituent (a complex V, perhaps), since in talk about about forms a constituent with the NP: ------------------------------------------------------------------------------ Doug Arnold, Doug Arnold, doug@uk.ac.essex (Janet) Dept. of Language & Linguistics, doug@essex.ac.uk University of Essex, doug%essex.ac.uk@ean-relay.ac.uk (ean) Wivenhoe Park, Tel: +44 206 872084 (direct) Colchester, CO4 3SQ, UK. Fax: +44 206 872085 ------------------------------------------------------------------------------
This archive was generated by hypermail 2.0b3 on Fri Dec 18 1998 - 20:31:40 PST