ZAS, Jägerstr. 10/11, 10117 Berlin
Room 522
Map, see ZAS for route description

Organizers:
Penka Stateva
Barbara Schlücker
Sabine Krämer

Contact:
barbara.schluecker@rz.hu-berlin.de

Abstracts

Andreas Bulk: Pronomonal Clitics in Arabic

The focus of the talk will be the morphophonological behaviour of the so called pronominal clitics of standard arabic (forthcomming SA).
Traditionally SA is described as a language which manifests its grammatical relations via case. But this might be too shortsighted, because in spoken language case morphology is not normaly expressed. That means that grammatical relations in SA are expressed via verb morphology, i.e. agreement and clitics, and not via case. In other words, SA is not a dependent marking language, as it resonates in the traditional perception and in most recent considerations, but a head marking one.
To support this assumption it is necessary to explore the formatives that can serve as these kind of markers. Exactly this is the motivation behind my research interest in pronominal clitics. The aim of this talk is to detect the morphophonological and the morphosyntactic facets of the formatives that can potentially realise these grammatical relations. That is agreement morphology for subject relation and clitics for direct and indirect object relation:

(1)     ?aÀt~~aiÖta=niù=hiù
         gave-3M.SG=me=it
         „You gave it to me.”
 

For this task it is necessary to delimit the formatives from the agreement morphology on the one hand and from free pronouns on the other. Further it should be investigated whether the object clitics are in some sense on the way to being agreement morphemes.
This talk will check out the common criteria for clitics, developed by Zwicky an others, on above mentioned formatives and give some outlook on how SA could be described as a predominantly head marking language.
 

Literatur

Stephen S. Anderson & Arnold M. Zwicky (to appear). Clitics. An Overview. [from th second edition of the Oxford International Encyclopedia of Linguistic, to appear].

T. Alan Hall (1999). The Phonological Word: A Review. In: T.A. Hall & U. Kleinhenz (eds.). Studies on the Phonological Word. Amsterdam: Benjamins.

Aaron L. Halpern (1998). Clitics. In: The Handbook of Morphology. Ed. by Andrew Spencer and Arnold M. Zwicky. S. 101 - 123.

Johanna Nichols (1986). Head-Marking and Dependent-Marking Grammar. In: Language Vol. 62, No. 1, S. 56 - 119.

Arnold M. Zwicky (1985). Clitics and Particles. In: Language, Vol. 61, No. 2, S. 283 - 305

Arnold M. Zwicky & Geoffrey K. Pullum (1983). Cliticization vs. Inflection. In: Language 59, No. 3, S. 502 - 513.
 
 

Ljudmila Geist:
Predication and Equation in Copula Sentences: Syntax-Semantics Interface.

The paper investigates the mapping of syntax to semantics in three types of copular sentences with two nominal phrases based on the classification suggested by Higgins (1979):

I.  Predicational: (What is John?) John is a teacher.
II.  Equative:  Mark Twain is Samuel Clemens.
III.  Specificational:  (Who is the winner?) The winner is John.

The difference between these types consists in the underlying semantic relation predication in type I vs. identity in type II, and combination of both in type III. There is a great deal of controversy in the contemporary syntactic and semantic literature concerning explanation of the fact that the same copula BE can convey both the predication and the identity relation.

English is a language in which, at least superficially, predicational, equative and specificational copula sentences resemble in surface morpho-syntax. The puzzles concerning ambiguity of English copula sentences come out more clearly when comparing them with a language which appears to use radically different morpho-syntactic means for expressing predication and equation in copular sentences. We will take Russian. Russian is a language with a rich morphology but without articles. With respect to copular sentences, Russian displays at least three peculiarities:

a)  defective paradigm: the copula verb byt' has no form for present tense, but only a past and a future form, cf. (1)
b)  overt equation marker: equative sentences require the addition of a morphological element, a pronoun, and have a complex syntactic structure, cf. (2)
c)  case alternation: in predicational sentences, post-copular noun phrases can occur in the Nominative or in the Instrumental and so can the initial noun in specificational sentences, cf. (3).

We will examine three types of copular sentences in Russian in comparison with English. The main issues we want to tackle are these:

1) What is the invariable semantics of equation and specification?
2) How is this semantics mapped to the syntax in each language?
3) What are the reasons for surface differences in the three types of copular sentences between Russian and English?

We will show that all three types of copula sentences are semantically related and can be derived form the predicational type by type-shifting rules as suggested by Partee (1986, 1987) and Chierchia (1988). We will argue in favour of the view that – despite of being morphologically non-overt – such type-shifting operations can be justified on a structural basis. The way to do this is to examine the Russian counterparts of the types of copula sentences and exploit the morpho-syntactic distinctions they show to re-analyse the less transparent English data. We will attempt to find out how these differences can be derived from language specific parameters.

(1)       Ivan     u?itel'
           Ivan  ?Cop.Pres teacherNOM.
           Ivan is  a theacher.
(2)       Mark Tvein       – *(éto)                  Samuel Clemens.
           Mark TwainNOM      this     ?Cop.Pres.Samuel ClemensNOM
           'Mark Twain is Samuel Clemens'
 (3)      Ivan        byl u?itel'          / u?itelem.
            IvanNOM was teacherNOM / teacherINSTR
            ‘Ivan was a teacher.'
 
 

Barbara Schlücker: Ambiguities of German and Dutch bleiben /blijven

The talk deals with semantic ambiguities of the German copular verb bleiben. When bleiben appears together with an infinite verb of position the resulting construction is – in a null context – ambiguos and can either mean that the subject of the construction remains in the position indicated by the infinite verb or else that there is a change of state which results in the position denoted by that verb:

(1) Peter bleibt stehen                                                                                            [Peter remains/becomes standing]

(2) Alle setzten sich hin, nur Peter blieb stehen                                [Everybody sat down but Peter remained standing]

(3) Plötzlich blieb Peter stehen                                                                                              [Suddenly Peter stopped]

"Bleiben + infinite verb of position" can obviously either denote a state or a change of state. In this second reading, bleiben seems to be equivalent to werden.
The general assumption about the REMAIN-reading of bleiben is that bleiben asserts some state P at an interval I and presupposes P at an interval I' which precedes I immediately. But on a closer look there are a lot of data where this alleged presupposition doesn't hold. In these cases, bleiben often seems to have a BECOME-reading, too.

(4) a. Im Dorf mit den niedrigen, weissgetünchten Häusern war eine Gewehrsalve zu hören. Dann blieb es still.
                        [In the village with the small, whitewashed houses a volley was heard. Then it remained/became quiet.]
                                                                                                                                                                                                                   [COSMAS-Korpus]
b. Wo ist das Buch geblieben?
c. (Zu jemandem, der sichtlich nervös hin- und herrennt): "Jetzt bleib mal ganz ruhig!"
                                                                           [(Talking to somebody who is extremely nervous):  "Please calm down!"]

In my talk I want to claim that there are two groups of bleiben-constructions:
(1) Constructions which denote states and whose meanings are obtained compositionally. They consist of three meaning components: the assertion of some state P, the reference to a counterstate ? P in a possible world and the presupposition of a non-specified state at an interval immediately before the interval of the assertion. Here, bleiben can either have REMAIN- or BECOME-reading.
Especially, I want to argue that the event structure of a BECOME-reading like in example (4a) is due to the internal structure of the verb, and not only the result of contextual enrichment, like for example in (5), where only the context tells us that the state of being rich is the result of some change of state.

(5) Plötzlich war Maria reich    [Suddenly, Maria was rich.]

(2) The second group only consists of the construction "bleiben + infinite verb of position" in the BECOME-reading. Here, bleiben doesn't denote a state but rather an event. This meaning cannot be obtained compositionally. I assume that this construction is – in contrast to group (1) – the result of a process of grammaticalization.
Temporal modification serves – among others – as a diagnostics for the contrast between these two groups.
 

Olav Mueller-Reichau: Contingent States

Story 1 (Krifka et al. 1995): NL sentences fall into two classes, depending on the semantics of their predicate, namely into particular/episodic sentences as opposed to generic/characterising sentences. The former refer to events, due to an event slot in the argument structure of the predicate. The latter, however, do not, because of a generic operator quantifying over the event argument in a certain way.

Story 2 (Maienborn 2001): NL sentences fall into two classes, depending on the semantics of their predicate, namely into eventive sentences as opposed to noneventive sentences. According to Maienborn, there is a grammatically relevant ontological contrast between accomplishments, achievements, activities and dynamic states ("d(avidsonian)-states" in her terminology) on the one side as opposed to static states ("k(im)-states") like know, contain, weigh, love, resemble etc. and copula-predicative constructions on the other side. Whereas predicates giving rise the former do contain a Davidsonian event variable in their argument structure, k-state predicates and copula predicative complexes do not.

Story 3 (Carlson 1977, Kratzer 1988): NL sentences fall into two classes, depending on the semantics of their predicate, namely into SLPredications as opposed to ILPredications. Typical SLPs are, among others, copula predicative complexes such as be hungry, be tired, be naked etc., whereas typical ILPs are be intelligent, be left-handed, be a bird etc. According to Kratzer, SLPs contain a Davidsonian spatio-temporal argument in their semantics, but ILPs do not.

Story 4 (Chierchia 1995): The similarity between characterising sentences and ILPredications is all too obvious and calls out for a uniform semantic treatment.

The problem: All four stories cannot be true at the same time. If ILPs are to be treated as inherent generics (story 4), then SLPs must be nongeneric. If SLPs are nongeneric, they must denote events (story 1). This conclusion is in line with story 3. However, it is in contradiction to story 2! If a sentence like Kira is hungry is nongeneric (story 4 + story 3) and noneventive (story 2), what is it?

The misunderstanding: Kratzer characterises the Davidsonian argument she assumes to be present in the semantics of SLPs to be a spatio-temporal argument. However, statio-temporal location is only a necessary, not a sufficient condition for an argument to count as Davidsonian. Therefore, a Kratzerian argument must not be called Davidsonian. The distinction Carlson 1977 aimed at is not the distinction between eventive and noneventive eventualities, as Kratzer´s terminology seems to suggest.

The solution: NL sentences fall into three (!) classes, depending on the semantics of their predicate, namely into generic sentences, eventive sentences and noneventive, nongeneric sentences. Good candidates for the latter class (which I want to call "contingent states") are the following:

(1) The flat stone over there is my table
(2) The mobile phone is in the kitchen
(3) Kira is hungry
(4) This table wobbles
(5) John works in the supermarket this summer
 

Cornelia Endriss & Stefan Hinterwimmer: „Ihr seid doch meistens Tiermediziner“ – Quantificational Variability and
Individual Level Predicates

We argue against the hypothesis that Q-adverbs such as usually, always, etc. are unselective binders in the sense of (Heim 82), (Diesing 92), (Kratzer 95), capable of binding situation/event variables and individual varables. In contrast to these approaches, we claim that Q-adverbs can only quantify over tense variables.
(Kratzer 95) argues that an otherwise ungrammatical sentence containing an individual level predicate can be rescued by the presence of an indefinite which introduces a free individual variable. The following example shows that this is not true:

(1) *A mobile phone that rings right now is usually expensive.

In (1), the Q-adverb cannot be applied to the matrix predicate, because expensive is an individual level predicate and as such does not introduce a variable which can be bound by the Q-adverb. But despite the fact that there is an indefinite in subject position, (1) is still ungrammatical. We claim that the reason for (1)‘s ungrammaticality is the temporal boundedness of the subject noun phrase due to the temporal adverb now in the relative clause. This is evidenced by the grammaticality of the following example sentences:

(2) A mobile phone that is sold here is usually expensive.
(3) Mobile phones (that are sold here) are usually expensive.
(4) Die Handys, die hier verkauft werden, sind meistens  teuer.
(The mobile phones that are sold here are usually expensive.)

In (2)-(4), the respective relative clauses are temporally unbounded. (4)‘s grammaticality shows that it is not the variable introduced by the indefinite  that makes the sentence acceptable, as plural definites are usually not assumed to introduce free variables. What seems to be crucial is the availability of a free tense variable.

The following example shows that it really has to be a tense variable, not an event variable that has to be bound by the Q-adverb:

(5) *People who are jumping in the air now are usually hippies. (Imagine a situation in which the rumour is spread that jumping in the air at a certain fixed time – now – will make the world a better place.)

Despite the fact that in (5) a plurality of events is introduced, the sentence is ungrammatical. This contrasts with:

(6) People who jump in the air are usually hippies.

Whereas sentences containing individual level predicates and Q-adverbs only become grammatical if the subject is not temporally bound, sentences with stage level predicates  are ambiguous under these circumstances:

(7) A horse that Peter likes usually eats grass.
(8) Horses that Peter likes usually eat grass.
(9) Die Pferde, die Peter mag, essen meistens Gras.
(The horses that Peter likes usually eat grass.)

Setting specific readings aside, the sentences in (5)-(7) have two readings each. In one reading, the predicates are interpreted as individual level predicates and the Q-adverb binds the free tense variable of the relative clause modifying the subject noun phrase. In the second reading, the Q-adverb binds the free temporal varable introduced by the matrix (stage level) predicate.

Diesing, Molly (1992): Indefinites. The MIT Press.
Heim, Irene (1982): The Semantics of Definite and Indefinite Noun Phrases. PhD-Thesis,
University of Massachusetts, Amherst.
Kratzer, Angelika (1995): Stage-Level and Individual-Level Predicates. In: G. N. Carlson & F. J. Pelletier
(eds.), The Generic Book, Chicago: The University of Chicago Press, 125-175.

Thomas Weskott: Pronominal Epithets and Informativity

It has been noted long ago that so-called "Pronominal Epithets" (PEs) like 'the cretin' in (1) behave both like pronouns and like names (or R-expressions) with respect to Binding Theory (cf. Akmajian & Jackendoff, 1970; Jackendoff, 1972; Lasnik, 1989; Hornstein & Weinberg, 1990).

(1) Mary slugged Bill_i  before [the cretin]_i  could make stupid excuses for himself(_i).

(2) * Bill_i  thinks I admire [the idiot]_i .

(Examples from Lasnik 1989; coindexation is to be understood as coreference.)

To say that PEs exhibit properties of pronouns means that they are marked as being anaphoric, hence as lacking descriptive
information about the referent they denote; this anaphoric character is witnessed by the fact that they can serve as anaphors that can be interpreted as being coreferential with a suitable antecedent, like 'Bill' in (1).

Strangely enough, PEs also share properties of R-expressions insofar as -- QUA being definite descriptions -- they do contain descriptive information about the referent they denote (e.g. that the referent named "Bill" in (1) is a cretin). It is this property that is taken to explain their obeying to Principle C of Binding Theory; hence the unacceptability of (2) above.

As far as I know, the accounts of PEs given in the literature have concentrated on their binding-theoretical status, neglecting their
strange relation to information structural notions like "new" and "old", both of which they seem to fall under at the same time. After giving a short overview of the syntactic behaviour of PEs, I will discuss their apparently self-contradictory information structural
status. I will propose to draw a close parallel between Topics and PEs, and discuss this claim in the light of a few real-life text samples.
Some speculation about the processing properties of PEs conclude the talk.

References:

Akmajian, Adrian & Ray Jackendoff (1970): Coreferentiality and Stress.
        LI, 1.1, 124-126.
Hornstein, Norbert & Amy Weinberg (1990): The Necessity of LF. The
        Linguistic Review,7, 129-167.
Jackendoff, Ray (1972): Semantic Interpretation in Generative Grammar.
        Cambridge, MA:  MIT Press.
Lasnik, Howard (1989): Essays on Anaphora. Dordrecht: Kluwer.