Date: Mon, 12 Aug 2002 12:18:10 -0700
From: Lee Fullerton
Subject: Abraham and Zwart, Issues in Formal German(ic) typology
Abraham, Werner, and C. Jan-Wouter Zwart, eds. (2002) Issues in Formal German(ic) Typology. John Benjamins Publishing Company, xviii+336pp, hardback ISBN 1588111024, USD 100.00, ISBN 9027227667, EUR 110.00, Linguistik Aktuell / Linguistics Today 45.
Lee Fullerton, Visiting Research Scientist, University of Arizona
This book is a collection of twelve papers with introduction and subject index. It presupposes familiarity with recent developments in generative grammar, including especially the Minimalist Program.
Werner Abraham, "Introduction", ties the following articles together on the basis of four distinctive aspects of Modern German: it arranges constituents according to thema and rhema; it passivizes without regard to transitivity or actionsart, making use of the expletive to preserve verb second; its definite and indefinite articles reflect discourse function; non-canceling multiple negation in certain dialects and aspects of ellipsis highlight the interface between formal predicate logic and language parsing. The remainder of the introduction is devoted to summarizing each article.
Werner Abraham and Laszlo Molnarfi, "German clause structure under discourse functional weight: Focus and antifocus", show that functions such as thema and rhema, which reflect the discourse context of the sentence, are inextricable from the grammar. This leads them to reject derivational functional categories of tense marking and agreement (AGR). An unmarked derivation begins with antifocus licensing of thematic (known or presupposed) elements such as personal pronouns and definite determiner-phrase (DP) objects within the verb phrase (VP), and continues by moving ("scrambling") these leftwards (upwards in the tree) out of VP. Sentence accent goes on the focus, a rhematic element such as an indefinite DP object which remains behind inside the VP. The movement and accent placement rules can be violated, but when so, the word order is marked and the accent is heard as contrastive. They conclude that SVO and SOV are typologically distinct and one should not try to derive the latter from the former. If one were to rewrite the latter as SXV, X would represent a large middle field to receive the thematic elements, something which SVO lacks.
Cedric Boeckx, "On the co-occurrence of expletives and definite subjects in Germanic", accounts for differences in the Definiteness Effect (DE)--an expletive requires an indefinite subject--between German (little DE), Icelandic (slightly greater DE), and Dutch and Swedish (greatest DE) in two ways. The first has multiple functional projections and a doubly filled comp filter. The second, the author's preferred solution, makes use of agreement patterns and the availability of multiple subjects/nominative objects.
Jocelyn Cohan, "Reconsidering identificational focus", confirms prior research to the effect that identificational focus relates to a set of contextually determined elements and can have scope over clause-mate operators, but she breaks new ground in other respects. Using data drawn from recordings of spontaneous speech, she shows that the exhaustiveness thought to hold for identificational focus is an implicature and can be violated when the discourse sets up alternative foci of the same semantic type. Further, the subtype known as contrastive focus turns out to be not a feature of the sentence, but rather of the sentence's context.
Christine Czinglar, "Decomposing existence", does semantic derivations for three Germanic existential constructions which use a semantically empty pronoun (not an expletive, in her view). In Mainland Scandinavian the verb is a historically mediopassive form of 'find'; in German it is 'give'; in the Alemannic dialect of German there is also a type with 'have'. This last derives from a structure with possessive 'have' in which the external argument (subject as possessor/owner) is "absorbed" by the pronoun and then recreated as a locative phrase. In contrast to this "locative" existential, the "pure" existentials with 'give' and 'find' derive by absorption of an agent/cause subject, which is not recreated. These two constructions contain an abstract generic quantifier which bars them from application to transitory states of existence or temporary characteristics of the existing entity. Thus while the 'give' type can co-occur with a locative phrase, this must be a habitat or habitual location. The author ends her paper with brief discussion of German 'give' examples which appear to violate this restriction.
Britta Jensen, "Polarity items in English and Danish", focuses on four such items in Danish which, when licensed by sentence negation, occur most naturally before the negation word. Given that the licenser must c-command the licensee, these items ought to follow the negation word. The author posits a strong plus-polarity feature for only these four items, with which--motivated by GREED--they raise into the negation phrase (NegP) for feature checking. Two of these items can also occur sentence finally after a relative clause. Unless the relative is also negative, the two are always interpreted as part of the matrix clause. Jensen's account posits a sigma phrase which subsumes NegP: when sentence final, the two polarity items are right adjoined to SigmaP for feature checking.
Wolfgang Klein, "The argument-time structure of recipient constructions in German", expresses dissatisfaction with notions such as subject, direct object, tense, aspect, actionsart, passive and tries to make some of these explicit using time intervals. The lexical content of a verb determines how many arguments it has and the time structure of each. Argument-time structures help determine case and the nature of the 'event' (e.g. accomplishment, achievement, activity, state). One reading of _H. bekam das Paket geoeffnet_ is 'H. received the package opened': the finite verb has a past time overlapping the two times before and after (the event); the subject H is not a possessor at the before time but is a possessor at the after time; the package has the same two times with respect to receiving, not possessed and possessed, but only one time with respect to opening: it is open beforehand and afterwards. In another reading, 'H got the package open', the argument-time structures are different.
Juergen Lenerz, "Scrambling and reference in German", seeks a theory of referential dependency to account for indefinite full noun phrase (NP) accusative objects in early positions, e.g. before sentence adverbials like _natuerlich_ 'naturally', or temporal adverbs like _immer_ 'always', or definite or indefinite full NP dative objects. He finds that the indefinite NP in its underlying position within the verb phrase (VP) is existential and referentially unambiguous when c-commanded by another NP. The same is true when the NP in question is topicalized, i.e. stands as first constituent before the finite verb. Otherwise the reference is ambiguous.
Enrique Mallen, "Attributive adjectives in Germanic and Romance", accounts for postnominal vs prenominal adjective position by assuming two base-generated structures having one NP embedded within another NP. In the one the adjective phrase (AP) is left of the noun of the higher NP, while in the other the AP is left of the noun of the lower NP, this lower NP being the rightmost branch of the NP above. The author posits raising of either structure into the specifier position of a function projection (FP) followed by checking of features called degree and temporal. Spanish data are treated in great detail, German data only briefly.
Laszlo Molnarfi, "Die Negationsklammer im Afrikaans: Mehfachnegation aus formaler und funktionaler Sicht" (Negation bracketing in Afrikaans: multiple negation from a formal and a functional perspective), accounts for two patterns of negation: written Afrikaans has two markers, one before the VP and one after the VP; spoken Afrikaans has these as well as a marker for every phrase in between. Plain negation resides in the head (not the specifier) of the NegP and spreads rightwards through its scope. Operators outside the (lowest) VP like _nooit_ 'never', _nerens_ 'nowhere' get raised into spec of NegP. The written language suppresses the spellout of negation between the first occurrence and the last. Both patterns mark the scope of negation for the parser; the spoken pattern redundantly marks rhematic elements.
Athina Sioupi, "On the syntax and semantics of verb-complement constructions that involve 'creation': A comparative study in Greek and German", considers a Greek construction in which a singular, count noun without determiner occurs as complement of a verb of creation: _htizo spiti_ '(I) build-1SG house' Rather than comparing the superficially similar German or English where an indefinite-article determiner must be present, one comes closer to the meaning with the indefinite plural construction, which also lacks a determiner: _Haeuser bauen, build houses_. The author argues that Greek belongs to the class of languages in which NP is [- argument, + predicate]; and that the singular noun is a predicate which has become an argument by taking on a zero determiner. This turns the accusative object into a kind and the construction into a process.
Wolfgang Sternefeld, "Wh-expletives and partial wh-movement: Two non-existing concepts?", examines two semantic accounts for the was-wh construction in German: _Was glaubst du, wen wir einladen sollen_ 'What do you think whom we ought to invite?' (Who do you think we ought to invite?). Only a few matrix verbs participate in this construction and they do not otherwise allow embedded questions. The first account, called direct dependency approach (DDA), proposes that _was_ is a meaningless expletive and a scope marker. The other account, called indirect dependency approach (IDA), proposes that _was_ is a logical operator/quantifier like _which_ in _which men_. While the latter selects for men, the former selects for propositions. After detailed semantic argumentation the author settles on a modified version of the IDA and buttresses his conclusion with data from Hungarian.
John te Velde, "Phases in the derivation of elliptical coordinate constructions in Germanic", proposes an alternative to the across-the-board treatment of conjoined clauses in which the second clause has either a subject gap or an object gap. His motivation are German coordinations in which the gap has a referent different from that of the antecedent in the first clause. The author's approach invokes Chomsky's notion of phases as syntactic objects and their use as probes for agreement. In this case the phase is the coordinating conjunction.
Like any anthology this collection is a mixed bag and every reader will consider important a different subset of articles in it. Even the four German-based themes identified by editor Abraham in his introduction reveal a particular interest in matters of discourse, an interest foregrounded in only three of the articles, namely those by him and Laszlo Molnarfi and that of Jocelyn Cohan. The articles on Afrikaans and Danish refer to German only occasionally. Two articles are not really about Germanic at all. Nonetheless, German is the language under scrutiny in most of the papers. As for the word "typology" in the title, it is meant to indicate a uniformly generative approach, rather than the search for universals. However, Wolfgang Klein's piece is not generative in any sense. Two disappointments are superficial: the book could have used better proofreading and an editor for English usage. In sum, urge your institution's library to buy this book; given its USD 100 price, you may not want it for your own library.