|
|
|
Workshop on Comparative and Theoretical Syntax:
“When and why do constituents move?”
December 14-16, 2004
Department of English
Institute of Language, Literature & Culture,
University of Aarhus
Building 467, Room 415 (4th floor)
Jens Chr. Skous Vej 7, DK-8000 Aarhus C
The theme of this workshop is constraints on (and in) syntactic movement or displacement – potentially one of the main sources of word order variation both within and across languages. The following are some of the questions to be discussed: What is syntactic movement? Which types of constraints are involved, e.g. violable vs. inviolable constraints, principles vs. parameters, based on considerations to do with parsing or with something else? What causes syntactic movement, e.g. interface conditions such as scope interpretation, feature checking, or prosody? What are the implications for the architecture of the language faculty? Although many of the contributions come from the perspective of generative grammar, there will also be a number of talks from other formal and functional perspectives.
Workshop Program:
|
Tuesday, December 14 | ||
10:00-11:00 |
Introduction |
Theories of Dependencies. |
11:15-12:15 |
Øystein Vangsnes
|
12:15-13:30 |
lunch |
13:30-14:30 |
Ken Ramshøj Christensen
|
|
14:45-15:45 |
Henrik Jørgensen
|
The Concept of Movement in Paul Diderichsen's Sentence Scheme. |
16:00-17:00 |
Sten Vikner
|
|
|
Wednesday, December 15 | ||
10:00-11:00 |
Ralf Vogel
|
|
11:15-12:15 |
Maia Andréasson
|
12:15-13:30 |
lunch |
13:30-14:30 |
Johanna Wood
|
|
14:45-15:45 |
Hans Götzsche
|
|
16:00-17:00 |
Torben Thrane
|
|
|
Thursday, December 16 | ||
10:00-11:00 |
Øystein Vangsnes
|
|
11:15-12:15 |
Gunnar Hrafn Hrafnbjargarson
|
Is It Really Moving? No V°-to-I° and the Stylistic Fronting Paradox. |
12:15-13:30 |
lunch |
13:30-14:30 |
Ralf Vogel
|
|
14:45-15:45 |
Ole Togeby
|
Is Movement Psychologically Real or Just a Pedagogical Shortcut? |
16:00-17:00 |
Ken Ramshøj Christensen
|
Quantifier Movement and Derivation by Phase - Now You See It, Now You Don't. |
[Printer friendly .pdf version]
Organizers:
Sten Vikner and
Ken Ramshøj Christensen
Financed by Sprogvidenskabelig Forskerskole Nord (SFN).
Registration:
The workshop is open to everyone who is interested.
If you are planning to attend, it would be helpful to us to know about it,
so please send an email to Ken Ramshøj Christensen
(engkrc@hum.au.dk)
Abstracts:
Maia Andréasson
(Department of Swedish, University of Göteborg)
Let's Assume They Just Don't Move...
– on Dealing with Word Order Differences in Lexical Functional Grammar.
In Lexical Functional Grammar the general assumption is that elements are base generated – they just stay where they're born. Instead of using the notion of movement of elements in a binary branching tree as a linking from a deep structure to a surface structure and onwards to logic and fonological structures, LFG lets different parts of grammar be illustrated by separate dimensions and links them together by mapping principles.
In this talk I will briefly outline the underlying reasons for why LFG doesn't "do movement" and present an OT-LFG-analysis of word-order differences in the Swedish mid field, an area of the clause that may very well be analyzed as a flat structure, similar to the mid field of Diderichsen's field schema.
Ken Ramshøj Christensen
(Department of English, University of Aarhus)
Visible Movement: Syntax and the Brain
Numerous aphasia studies have shown that what is impaired with regard to comprehension in agrammatism, associated with Broca's aphasia, is the interpretation of theta-related syntactic movement. Furthermore, Yosef Grodzinsky and his colleagues have shown that the computation of moved constituents, more specifically XPs, activates Broca’s area in normal subjects. I will present data from a study on Danish sentence comprehension that supports the correlation between the computation of syntactic chains and neural activation in Broca’s area. However, the correlation may be dependent not so much on changes in the order of theta roles as the target of movement. Dative shift doesn't increase activation in Broca's area, nor does NEG-shift. Phrasal movement rather increases activation in the cerebellum. Crucially, this shows that movement is not just a theoretical construct but has psychological as well as neurological reality.
Ken Ramshøj Christensen
(Department of English, University of Aarhus)
Quantifier Movement and Derivation by Phase - Now You See It, Now You Don't.
In this talk I apply Chomsky's (2001) phase-based derivational approach to quantified objects. In the traditional Government and Binding approach scope ambiguities of quantifiers (e.g. what, all, some, any, no), are accounted for in terms of covert movement. In "Derivation by phase" Chomsky explicitly dispenses with covert movement. However, I propose a that operators may strand their phonological features - a very constrained version of covert movement. This approach provides a unified account of the empirical data from a variety of languages including Danish, Icelandic, English, and French.
Hans Götzsche
(Centre for Linguistics, University of Aalborg)
What is movement in Epi-Formal Theory?
Epi-Formal Theory is the overall term for theories and models based on Epi-Formal Systems, i.e. formal systems with one or more external requirements, forming combinatorial systems by the use of a formal language, and the Epi-Formal Analysis in Syntax (EFA(X)) is able to produce calculi as formal descriptions of syntactic constructions. From this perspective the notion of movement is rather odd, even weird, because one may justifiably ask a number of challenging questions about movement: what is movement as an ontological category? Do things move in a formal system? Do things move in linguistic constructions? Is movement a process in a cognitive (Chomskyan computational) system, i.e. in a synchronical approach, or is movement a part of a historical process, i.e. in a diachronical approach? If you think it over there seem to be far more questions than answers, so, accordingly, I will do my very best to reduce the number of questions.
Gunnar Hrafn Hrafnbjargarson
(Department of Linguistics, University of Oslo)
Is It Really Moving? No V°-to-I° and the Stylistic Fronting Paradox
When it comes to word order, Icelandic differs from the other Scandinavian languages in various ways, the most obvious difference being the position of the finite verb in embedded clauses. Not all embedded clauses are different. When Stylistic Fronting has taken place, Icelandic has the same word order as Danish or Faroese. When Vº-to-Iº movement has changed the base-generated word order from adverb-verb to verb-adverb, Stylistic Fronting reestablishes the base generated word order, i.e. it changes the order from verb-adverb to adverb-verb. This makes up the paradox of Stylistic Fronting, i.e. that it is invisible in most of the cases.
Because of the Stylistic Fronting Paradox, I will argue that younger speakers of present day Icelandic do not have Stylistic Fronting of adverbs (as opposed to Stylistic Fronting of participles and particles). Furthermore, I will follow Angantýsson (2001) in assuming that the Mainland Scandinavian/Faroese word order that can be found in embedded clauses in Icelandic is rather due to shorter movement of the verb to T (as opposed to Agr) than to Stylistic Fronting.
Henrik Jørgensen
(The Nordic Department, University of Aarhus)
The Concept of Movement in Paul Diderichsen's Sentence Scheme
Paul Diderichsen's description of Danish syntax (Diderichsen 1936, 1941, 1946, 1966) contains a series of controlled choices between different places for parts of the sentence with comparable relational character. These choices deal with the front place ('forfelt' or 'fundamentfelt in Danish'), enclitic pronouns, presentational subjects in object positions, extraposition, free adverbs, and backwards movements of heavy NP's. Diderichsen's own chosen formulations do not point to any clear-cut theory or to any heavy formalism. Instead his wording is always carefully vague and seems to try to evade closer scrutiny of the linguistic factors related to movements by simply using a number of semantically incoherent metaphors. Nevertheless the words in use point to a general picture of what Diderichsen thinks movements to be. My paper shall try to give a complete picture of this.
Bibliography:
Diderichsen, Paul: "Prolegomena til en metodisk dansk Syntax." In Spang-Hanssen, E.,
Viggo Brøndal & Johannes Brøndum-Nielsen (Hgg.): Forhandlinger paa det ottende nordiske
Filologmøde i København den 12-14 August 1935. København: J. H. Schultz Forlag 1936 S. 41 - 46.
Diderichsen, Paul: Sætningsbygningen i skaanske Lov. = Universitets-Jubilæets danske
Samfunds Skrifter Nr. 327. København: Ejnar Munksgaard 1941.
Diderichsen, Paul: Elementær dansk Grammatik. København: Gyldendal 1946
Diderichsen, Paul: Helhed og struktur. København: G. E. C. Gads Forlag 1966.
Torben Thrane
(Faculty of Language and Business Communication, Aarhus School of Business)
The Explanatory Range of Movement
Drawing a distinction between systemic and functional explanations of movement in general, I shall argue that the Chomskyan view of movement in language is originally functional. With the advent of the Minimimalist Program, however, it has become systemic, but no argument for this change has been forthcoming. I’ll then present data (from Danish) to sustain the view that only functional type explanations of movement can be empirically motivated, and these only if movement is reinterpreted as transition states between representations of different kinds.
Ole Togeby
(The Nordic Department, University of Aarhus)
Is Movement Psychologically Real or Just a Pedagogical Shortcut?
In Diderichsen's tradition the concept of 'movement' has never been anything but a pedagogical way of generalizing about word order. Simon Dik completely rejects talking about movement but has something that resembles it nonetheless; what is that? And in Chomsky's various models there has always been movements of different kinds; and when the theory claims to be a theory about biology, does movement have a psychological reality, and if so, which?
Øystein Vangsnes
(Department of Linguistics, University of Potsdam)
Definiteness across Scandinavian
Apart from Western Jutlandic all Scandinavian varieties possess an enclitic definite article which attaches to the noun. A notionally definite noun phrase not containing any adnominal modifiers will thus in the major subset of the varieties have the form N+DEF. In my first talk I will start out by pointing out some of the well-known differences that arise across the major subset once modifiers are added, such as the appearance of a preadjectival/prenumeral lexical article in Mainland Scandinavian and Faroese (but not in Icelandic), the incompatibility of the lexical and enclitic article in Danish, and the phenomenon of adjective incorporation in definite noun phrases in Northern Swedish and Central Norwegian dialects. I will then turn to some lesser known facts concerning definiteness marking in Scandinavian varieties. For one thing I will discuss the fact that in Icelandic definite noun phrases with numerals, a phrasal constituent containing the enclitic definite morpheme is displaced to a prenumeral position (cf. the abstract for my second talk). I will also discuss the partitive use of the enclitic morpheme in Northern and Österbotten Swedish dialects, i.e. "definite forms" found on singular mass and non-specific plural noun phrases as well as in generic noun phrases. The Österbotten dialects are especially interesting since they have adjective incorporation as an option alongside the regular Swedish pattern with a preadjectival lexical article (and no A-incorporation). Interestingly, the preadjectival article seems to be incompatible with 'partitivity' (in the sense alluded to above).
The talk will to a large extent be descriptive and it will prepare the ground for my second talk where I will discuss some of the theoretical issues that arise if one wants to give a unified microcomparative analysis of the structure of the Scandinavian noun phrase. One very general issue concerns the fairly robust empirical generalization that the enclitic article appears to be associated with a low/rightward position whereas the lexical article is associated with a high/leftward position. The question is what the relation, if any, is between the two positions.
Øystein Vangsnes
(Department of Linguistics, University of Potsdam)
A phrasal roll-up analysis of the Scandinavian DP
My second talk will build on some of the empirical issues laid out in my first talk on definiteness across Scandinavian. In particular I will focus on the displacement seen in Icelandic definites containing a numeral. Compare the examples in (1) with those in (2), and notice specifically the ungrammaticality of (1b)(with a non-partitive meaning).
(1) a. frægu bækurnar mínar fjórarAn important fact concerning this displacement phenomenon is that it involves a phrasal constituent and can thus not be analyzed in terms of head movement of, say, the noun. Moreover, the phenomenon clearly shows that the enclitic definite article must originate within the moving constituent - the movement cannot be an effect of attraction of the XP in question by the enclitic definite article (in a prenumeral position), witnessed by the fact that the enclitic article attaches to the noun and not to the possessive (the latter marking the right boundary of the moving XP in (1)). Further complications arise, however: whereas possessives do move along past the numeral, phrasal possessors, relative clauses, and complement PPs do not and are instead stranded in a postnominal position.
In my analysis I will entertain an idea whereby the moving XP not only undergoes XP-movement past the numeral, but also that it itself is formed by phrasal roll-up movements in the low area of the noun phrase (notably a little nP phase). This will shed some light on why the "bigger" adnominal modifiers do not move along, whereas possessives do. A question that arises is why the other Scandinavian varieties do not employ this strategy but instead inserts a prenumeral lexical definite article. I will argue that this pertains to the featural properties of the XP eligable for movement, and that intervention effects (Relativized Minimality) bars the movement in the non-Icelandic varieties.
Sten Vikner
(Department of English, University of Aarhus)
Syntactic Movement and Optimality Theory
Optimality Theory (OT) is a very general framework, in fact it is more like a theory about the form of linguistic theories. Whether or not syntactic movement is found in an given implementation of OT depends completely on which kind of constraints are employed within the OT framework. This again means that OT is not incompatible with syntactic movement and that OT can be used to explain and/or motivate it.
This will be illustrated by an OT-based account of verb movement differences between Middle English, (modern) English, and (modern) Danish. The analysis will also be an example of how OT may offer an alternative to the view based on binary parameters, in that the options here involve more alternatives concerning finite verbs and I° than for a language either just to have V°-to-I° movement (like Middle English and e.g. French) or not to have it (like Danish). I will be arguing that in modern English, the finite verbs that surface in I° (or higher) are base-generated in I°.
Based on an analysis of the different positions of the finite main verb in English and Danish (and some additional related languages) in terms of V°-to-I° movement, the paper starts by linking these positional differences to the presence (or absence) of inflection for person in all verbal tenses (cf. Vikner 1997). It is then shown that modern English is not just different from Danish but actually unique in that there are two different types of finite verbs with different syntax, and also that the two verb types should be taken to be thematic and non-thematic verbs, rather than main and auxiliary verbs (cf. Roberts 1985).
The rest of the paper will give a comprehensive analysis of the complex syntax of these two types of English finite verbs (also as compared to finite verbs in Danish and in Middle English) in terms of violable (and potentially conflicting) constraints, in particular the following three:
Ralf Vogel
(Department of Linguistics, University of Potsdam)
For many generative syntacticians, object shift seems to be an exotic property of the Scandinavian languages. This could be due to the way the descriptive generalisation is stated. If we state it differently then we observe that German and English have something very similar.
The descriptive generalisation with broader coverage is: "Weak function words in the Germanic languages tend to escape the edges of larger prosodic domains." This tendency can be found with weak pronouns in Scandinavian languages, English, German and Dutch, but you can also find it with auxiliaries, e.g., in German dialects.
A purely phonological account fails to integrate the syntactic licensing conditions for object shift in an appropriate way. The standard semantico-syntactic accounts of object shift, on the other hand, fail to explain why it is only weak pronouns that undergo object shift in Mainland Scandinavian.
I will present an Optimality theoretic model of the phonology-syntax interaction in the spirit of Selkirk (1996), Truckenbrodt (1999) and Samek-Lodovici (2002). In ths model, linear order is determined by the interaction of (abstract) syntactic and prosodic factors. The account can successfully be applied to pronominal object shift and further related phenomena in English and German.
Ralf Vogel
(Department of Linguistics, University of Potsdam)
Minimal Link Effects in a correspondence based OT model
I will explore an optimality theoretic architecture for syntax that is guided by the concept of correspondence: the grammar regulates the "translation" of an underlying representation into a surface form. This surface form, called PF in generative syntax, is, nevertheless, quite complex.
The means by which underlying relations and properties are encoded at the surface are words, their inflection, their relative order (precedence and adjacency), and prosodic structure. Information that is not encoded in one of these ways remains unexpressed, and gets lost unless it is recoverable via the context. Different kinds of information are often expressed by the same means. The resulting conflicts are resolved by the relative ranking of the relevant correspondence constraints.
Both semantic and abstract syntactic information are reflected by the surface form. The empirical domain where this architecture is tested are minimal link effects, especially in the case of wh-movement. The OT constraints require the surface form to reflect the underlying semantic and syntactic representations as maximally as possible. This way, it is possible to treat word order freezing (the impossibility of object-subject order under surface ambiguity of the involved NPs) as another instance of minimal link phenomena (like, e.g., superiority in English multiple questions).
Though this model is strictly representational, it has an abstract syntactic part that may or may not entail syntactic movement. By focusing on the surface form as the most important object of syntactic study, we shift our attention away from the mechanisms underlying sentence generation to the well-formedness conditions for outputs.
Johanna L. Wood
(Department of English, University of Aarhus)
Degree Modifiers and Movement in the DP
Typically, adjectives in Germanic languages immediately follow the article and precede the noun. However, certain adjectives may precede the indefinite article if they are premodified by a degree adverb:
(1) So inspiring a speaker
(2) *Inspiring a speaker
The degree adverbs involved in this construction form a restricted set. In present-day English they are: so, too, as, how, that and this. Similarly, such and what also precede the indefinite article, but without an accompanying adjective.
(3) Such an inspiring speaker
(4) *Such inspiring a speaker
While it is clear that the operations involved are DP internal, it is not so clear how these constructions are derived. The degree phrase could be base generated preceding the article (e.g. Delsing 1993) or could be generated following the indefinite article and subsequently moved (e.g. Bolinger 1972, Matushansky 2002). On the basis of modern corpus data and historical data I argue that fronted degree modifiers are derived through XP movement but that such behaves more like a head.
Last updated December 12, 2004
Back to the program