On the Real Nature of Scrambling in Korean

Hong-Pin Im
2008 Macrolinguistics  
The putative scrambling in Korean has intrinsic difficulties due to the lack of driving force for the movement of constituents. The Case theory, EPP and Focus Feature, etc. does not support the movement of the scrambled constituent. The radical LF-reconstruction or the LF-lowering also raises serious questions about the semantics of the moved constituent. This paper claims that the putative scrambled constituent should be analyzed as a Word Order Topic, based on the Lexicalist Case Hypothesis
more » ... t Case Hypothesis which says that the case should be defined on the basis of lexical information, and the Hypothesis of Variable Medial Projection which says that the number of medial projections should be matched with the number of arguments of a predicate. As to the so-called Long Distance Scrambling (LDS), the typical cases considered hitherto to be LDS is not a real LDS, because the object in front of the matrix subject is not a constituent of the embedded clause, but that of the.matrix clause. As to the embedded constituent in front of the matrix subject, this paper introduces the principles of Constituent Affiliation Principle which says that a constituent should be in its own clause, or should be construed to be as if it were in its own clause, and Transparent Verb Principle (TVP) which says that a thinking-or saying-verb functions as if it were nonexistent in certain discoursesyntactic contexts. The typical object of the so-called LDS that appears in the thinking-or saying-verb construction is on a par with that of the Short Distance Scrambling (SDS) in virtual structure. In virtual structure, the matrix clause can be seen nonexistent by the TVP and the matrix subject can be seen to function as a parenthetical phrase. Hong-Pin Im Macrolinguistics 2 ( 2008) 57 Introduction This paper intends to make clear the real nature of scrambling and the semantic effects of seemingly scrambling in Korean. In general, scrambling has been taken to be the operation of rearranging the order of constituents placed at given positions in a sentence. However, from the start, scrambling runs into unexpected difficulties. One is concerned with the basic word order. In the Minimalist framework (Chomsky, 1995) , the basic order of the constituents in a sentence is not fixed or not in a stable state, because the internal merge can be applied even to a structure which is not full-fledged. The order of constituents is largely dependent on factors like Head parameter and/or Linear Correspondence Axiom by Kayne (1994) , and the like. The other is concerned with "Last Resort." Obviously it is not clear what is the motive or driving force for a constituent to scramble to another position. In this paper it is claimed that there is no such thing as scrambling in Korean, as a variety of surface syntactic structures are analyzed in terms of the Hypothesis of Variable Medial Projections and Case Lexicalist Hypothesis. Тhere are only word order variations as given, which seemingly look like scrambling or rearrangement of constituents. If so, naturally the question of how the multiple word order variations in surface structures can be explained arises. The answer is not so difficult: The alleged scrambled constituent is given at that position. The semantic and/or pragmatic function of the putative scrambled constituent is that of topic, which is called "word order topic" in this paper. Dichotomy in approaches to scrambling The current dichotomy in the approaches to the problem of scrambling seems to be that between the movement approach and the base generation (or non-movement) approach. The former sees that a variety of word orders is the result of movements, whereas the latter sees that it is not the result of movements, but that it is given as such (Cf. Macrolinguistics 2 (2008) 58 (1) a. Harry-ga Mary-eui son-eul jab-ass-da. ① NK GK hand-AK hold-PST-DF "Harry held Mary"s hand." b. Mary-eui son-eul Harry-ga jab-ass-da. (Scrambled version of (1a)) GK hand-AK NK hold-PST-DF "Mary"s hand, Harry held." ② c. Mary-eui son-eul Harry-ga t jab-ass-da. (Derivation from (1a) to (1b)) The example (1a) exhibits the normal, basic, or canonical word order of a typical transitive sentence in Korean. In (1b), it seems that the object of (1a) moves to the sentence initial position. The movement approach to scrambling (or "derivationalist" hypothesis) sees that the structure of (1b) is derived from the canonical structure like (1a) by way of a movement as postulated by the arrow in (1c). As to the A and/or A-bar distinction of the landing site of scrambling, various proposals are put forth, depending on the grammatical theories and versions (Cf. By contrast, the base generation (or non-movement) approach to scrambling sees that the structure of (1b) is not derived from a structure like (1a). Instead, as postulated in Y.-S. is set up. The putative scrambled constituent is lowered into its θ-role position, so that its θ-role is checked. Granted that LF lowering can be motivated for the θrole checking, there still remains the question of how the putative scrambled structure is to be generated. ③ Movement approach and lack of motivation Derivationalists take it granted that there is a real movement like (1c) between (1a) and (1b) . However, the derivationalists should answer the question of what is the driving force involved in the movement as depicted in (1c). It is well known that the derivationalist posits underlying word order for a putative scrambled syntactic surface structure. If the operation of merge in Minimalist Program generates a surface structure similar to the scrambled ① This paper adopts the Korean Official Romanization System for Korean data. ② The English annotations in this paper reflect the author"s point of view on scrambling. ③ As to the counter-attack done to LF lowering, see
doi:10.26478/ja2008.2.2.4 fatcat:mkvdt72f7fab5p4bvdoea35bpy