This paper examines cyclical changes in comparative subclauses, showing how operators are reanalysed as complementisers via the general mechanism of the relative cycle, and how this is related to whether certain lexical elements have to be deleted at the left periphery. I also show that only operators appearing without a lexical XP can be grammaticalised, which follows from the nature of the formal features associated with the various operator elements. Though the main focus is on Hungarian historical data, the framework can be applied to other languages too, such as German and Italian, since the changes stem from general principles of economy.
My paper investigates the diachronic development of the Modern Hungarian finite declarative complementiser hogy 'COMP'. 1 In Old Hungarian, hogy could be combined with other complementisers, e.g. mint 'than/as', giving configurations like hogymint and minthogy, that is, complementiser combinations in general are attested both in the hogy+X and the reverse X+hogy orders, X standing for an unspecified complementiser. The rich variation of Old Hungarian complex complementisers is not fully reflected in Modern Hungarian: it is invariably only one of the orders that survived. I will show that it is always the one that fully grammaticalised into a single C head; this is ultimately tied to the original underlying order of hogy and X as separate C heads. I will also demonstrate that hogy came to be used as a general marker of finite subordination.(1) a. I don't remember who wrote the book.b. I don't remember what he wrote. c. I don't remember which book he wrote. d. I don't remember when he left. e. % John is taller than what Mary is. (Chomsky 1977: 87, ex. 51a)3 Note that Rizzi (1997) uses a more complex CP-system, in which other projections (TopP, FocP, etc.) may appear in between the two CPs hosting complementisers. This is not relevant for the present discussion, since I am largely concerned with the combinations of complementisers only, and Hungarian (both now and historically) rarely exhibits high topics, as topics are normally found below the CP-domain (and focus always is). 4 The reason why there are two distinct C heads is, according to Rizzi (1997), that they have different functions. The lower C head is responsible for defining the finiteness of the clause: while finite clauses contain a tensed verb (e.g. John is hungry is a finite clause), non-finite ones do not (e.g. the clause to go to Berlin is a non-finite one in a complex sentence such as I want to go to Berlin). Higher C heads, on the other hand, are responsible for defining the Force of the clause, i.e. whether it is declarative, interrogative, relative etc.
Clauses can fulfill various functions in discourse; in most cases, the form of the clause is indicative of its discourse function. The discourse functions (such as making statements or asking questions) are referred to as speech acts, while the grammatical counterparts are referred to as clause types (such as declarative or interrogative). Declarative clauses are canonical (that is, they are syntactically more basic than non-canonical ones): they are by default used to express statements, and they represent the most unmarked word order configuration(s) in a language. Other clause types, such as interrogatives, can be distinguished by various means, including changes in the intonation pattern, different (non-canonical) word orders, the use of morphosyntactic markers (such as interrogative words), as well as combinations of these, as can be observed across Germanic. The explicit marking of clause types is referred to as clause typing, and it affects both the syntactic component of the grammar and its interfaces. Apart from main clauses, which can correspond to complete utterances, there are also embedded clauses, which are contained within another clause, referred to as the matrix clause: matrix clauses can be either main clauses or embedded clauses. Embedded clauses may be argument clauses, in which case they are selected by a matrix element (such as a verb), but they can also be adjunct clauses, which modify some element in the matrix clause (or the entire matrix clause). Embedded clauses fall into various clause types. Some of these can also be main clauses, such as declarative clauses or interrogative clauses. Other embedded clause types do not occur as main clauses, as is the case for relative clauses or comparative clauses. Clause typing in embedded clauses has two major aspects: embedded clauses are distinguished from matrix clauses and from other embedded clause types. Main clauses can be typed in various—syntactic and non-syntactic—ways, but Germanic languages type embedded clauses by morphosyntactic means intonation plays little, if any, role. These morphosyntactic markers fall into various categories according to what roles they fulfill in the clause. Germanic languages show considerable variation in morphosyntactic markers, depending on the clause type and the variety, and in many cases, such markers can also co-occur, resulting in complex left peripheries.
The paper argues that structural case assignment properties of English and German reduced comparative subclauses arise from syntactic requirements as well as processes holding at the syntaxphonology interface. I show that constructions involving both an adjectival and a verbal predicate require the subject remnant of the adjectival predicate to be marked for the accusative case both in English and German, which cannot be explained by the notion of default accusative case, especially because German has no default accusative case. I argue that a phonologically defective subclause is reanalysed as part of the matrix clausal object, and hence receives accusative morphological case.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.