preliminary version of the talk schedule (pdf-file)

Thu, 4th May Fri, 5th May Sat, 6th May
09:15am Welcome
09:30am Joachim Jacobs
Main Clause Phenomena in German Relative Clauses
09:30am Magdalena Kaufmann & Adrian Stegovec
Imperatives - relatively Slovenian
09:30am Manfred Krifka
Conditional Assertions in Commitment Space Semantics
10:30am Coffee Break 10:30am Coffee Break 10:30am Poster Session
11:00am LIP
On Feature Mismatches in Relative Clauses
11:00am INT
Anaphoric priority modals in attitude contexts
11:45am VER
Relative Clauses as target for Long Extraction?
11:45am CON
Idioms and Nonrestrictive Relative Clauses
11:45am Christian Lehmann
Der Relativsatz des Cabecar
12:30pm Lunch Break 12:30pm Lunch Break 12:45pm Lunch
02:00pm Valentina Bianchi
Relative Clause Extraposition in Italian
02:00pm Doug Arnold
"Mismatch" Relatives
03:00pm ERW
On the Semantics and Syntax of Relative Clauses in Acquisition
03:00pm ROM
Relativizers are not complementizers: evidence from Romance
03:45pm Coffee Break 03:45pm Coffee Break
04:15pm NRR
Non-at-issueness and the appositive-continuative distinction
04:15pm Richard Kayne
English for as a wh-phrase


"Mismatch" Relatives | handout
Doug Arnold

Normal English Relative Clauses agree in number with the nominal they modify (e.g. "a problem that bothers/*bother me" "problems that *bothers/bother me"). However, this pattern is violated in a relatively common, but hitherto little studied English construction, as e.g. "this is [one of those problems that really bothers me". It appears that this construction will pose a serious challenge for any theory of agreement that tries to deal with both morphosyntactic and semantic aspects of agreement. A detailed description of the construction is provided, and a formal analysis is presented. The analysis is formalised in the framework of HPSG.

go back up

On relative clause extraposition: Italian vs. English | handout
Valentina Bianchi

In this talk I discuss a recent approach to relative clause extraposition proposed by Fox & Johnson (2016), in terms of attachment of the extraposed RC to the copy of the relative “head” created by Quantifier Raising. I show that this analysis cannot account for two constraints on RC extraposition in Italian, whereby the “head” must be a non-presuppositional weak NP. I then elaborate a base-generation account of extraposition in Italian, based on the assumption that (a) the weak NP “head” sitting in the argument position leaves this argument position open, and (b) the RC is adjoined to the predicative nucleus, and is semantically integrated via generalized conjunction. I then discuss an elaboration of this approach to account for RC extraposition from strongly quantified NPs in English. I adopt and modify Fox & Johnson’s proposal that the actual quantifier is adjoined to the predicative nucleus, and I finally sketch a unified analysis that can account for both Italian and English, doing away with rightward adjunction.

go back up

Main Clause Phenomena in German Relative Clauses
Joachim Jacobs

Based on a distinction between two types of main clause phenomena (MCP) the talk explores the conditions under which German relative clauses (RC) can host MCP. It starts with a comparison of two theoretical approaches to MCP: (i) Theories based on the assumption that MCP are licensed by illocutionary elements have fundamental problems. (ii) These problems are solved or avoided by a theory that assumes that restrictions on MCP follow from two semantic/pragmatic context conditions which reflect the status of MCP as expressives. But RC remain problematic. The two context conditions do not account for all RC that exclude MCP. Therefore a third condition, ANTI-RESTRICTIVITY, is added, which not only covers RC but also several types of adverbial clauses. Moreover, it explains unexpected occurrences of MCP in integrated non-appositive RC. Finally, possible explanations for ANTI-RESTRICTIVITY are discussed. RC restricting strong quantifiers probably prohibit MCP because they are referentially variable and this runs counter to a presupposition of the speaker attitudes expressed by MCP. The fact that some kinds of MCP are also excluded from RC restricting (in)definite or weakly quantifying arguments might be due to conflicts between the semantic function of such RC with the discourse functions typically associated with the relevant MCP, but the details are not clear.

go back up

Imperatives - relatively Slovenian
Magdalena Kaufmann & Adrian Stegovec

go back up

English for as a wh-phrase | handout
Richard Kayne

go back up

Conditional Assertions in Commitment Space Semantics
Manfred Krifka

There are two fundamentally different ways to understand conditional clauses: as assertions of conditional propositions (it is asserted that if p, then q), or as conditional assertions of propositions (if p, then it is asserted that q). The first approach is prominent in linguistic semantics (Stalnaker, Lewis, Kratzer, von Fintel), and sophisticated theories have been developed to represent the meaning of conditionals compositionally. But there are compelling arguments for the second account (e.g., Dudman, Barker, Edgington; also Stalnaker, 2009), like the constraints for embedding conditionals under negation and disjunction, or the existence of non-assertive conditionals. In this talk I will propose an analysis of conditionals within dynamic semantics (cf. Heim, Veltman, Kaufmann, Starr for such approaches), in particular within the framework of Commitment Space Semantics (Cohen & Krifka 2014, Krifka 2015). CSS has a notion of common ground (called commitment space) that does not only assume a body of shared information, but also has a projective component indicating the various ways how the common ground may develop in the future. An indicative conditional does not affect the currently shared information in the narrow sense, but also its future development; the assertion of the conditional if p then q changes a commitment space so that whenever p holds in future developments, q is asserted. I will show that this modelling of conditionals predicts that conditionals cannot be negated or disjoined in a straightforward way, and that it allows for non-assertive conditionals, like conditional questions. I will also discuss counterfactual conditionals and argue that their “fake” past tense (Iatridou) can be understood as the process of going back to a hypothetical past information state in which the protasis p could still be consistently assumed.

go back up

Der Relativsatz des Cabecar
Christian Lehmann

The Cabecar relative clause is a circumnominal relative clause without any marking of subordination, attribution or nucleus formation. It is recognizable as such by structural criteria only if its syntactic function in the matrix is signaled grammatically, and otherwise only by semantic and prosodic criteria. Elegibility of an internal nominal expression as nucleus of the relative construction follows a complex decision hierarchy of grammatical and semantic conditions. Diachronically, the relative construction is derivable from an asyndetic combination of two independent declarative clauses.

go back up

Research Group Talks

Idioms and Nonrestrictive Relative Clauses | handout
CON with Doug Arnold

Vergnaud (1974) and Fabb (1990) have established the view that an idiom cannot be split up by a nonrestrictive relative clause (NRC). (1) *The headway, [NRC which the students made last week], was phenomenal. (Fabb 1990:71) We will discuss data that seem to represent exceptions to this generalization. These are data with decomposable, syntactically highly flexible idioms such as pull strings in constellations as in (2a) (Arnold & Bargmann 2016), which are slightly different from (1), and data with kinegrams such as prick one's ears, which are even more permissive with NRCs, see (2b). (2) a. The strings that were pulled for you before, [NRC which I hereby promise I will pull for you again], will get you that job. (Arnold & Bargmann 2016) b. Hans hat echt große Ohren, [NRC die er in meinem Unterricht übrigens auch ruhig mal spitzen könnte.] 'Hans has really big ears, which by the way he might also want to prick in my class every now and then.' We will argue that these data support approaches such as Arnold (2007) and AnderBois et al. (2015), which allow for an interaction between the matrix clause and the NRC on a syntactic and semantic level.

go back up

On the Semantics and Syntax of Relative Clauses in Acquisition

In this talk we investigate two central aspects of the acquisition of relative clauses (RCs) in German: the acquisition of different RC interpretations and the acquisition of word order variation in RCs.
Focusing on the two well-studied types of RCs, restrictive and appositive, we first address the question as to how and when children acquire the semantics of restrictive and appositive RCs. While the exact semantic nature of both interpretations remains debated, according to some accounts appositive RCs involve more semantic operations than restrictive RCs (e.g., del Gobbo 2008). As for acquisition, this may suggest that restrictive RCs are mastered before appositive RCs. Two novel comprehension experiments were developed, using picture-selection and acceptability judgment, to test the interpretation of restrictive and appositive RCs. As prosody and discourse adverbs have been argued to tease appositive and restrictive RCs apart, we also considered whether prosody (integrated or non-integrated) and presence of discourse adverbs (ja ‘right’) were used as cues for the different reading. A total of 116 children (ages 3-6) and 20 adults participated. 3-year-old children strongly preferred the appositive over the restrictive reading of RCs. 4- to 6-year-olds, in contrast, exhibited a strong preference for the restrictive reading. Interestingly, across ages we found no influence of prosody and discourse particle on the interpretation of RCs. We argue that in general acquisition of RC semantics proceeds in a step-wise fashion with mastery of restrictive RCs starting at age 4. The 3-year-olds’s preference of appositive RCs we take to reflect their interpretation as independent clauses rather than as ‘real’ appositive RCs.
Focusing on verb placement, we examine the acquisition of word order variation. In German RCs usually the verb is in final (V-final) position, but under specific conditions V2 is licensed (Gärtner 1998, 2001). We address the question of how and in which order children acquire these word orders, and what these results reveal about the syntax of embedded V2. Using the delayed imitation method, we designed a novel production experiment testing children’s repetitions of V2 and V-final structures. A total 97 children aged 3 to 6 and 21 adults participated. In general, the results reveal a strong preference for V-final RCs over their V2 counterparts, especially in the group of 3-year-olds. Children produced sig. more V-final RCs than V2 structures. Moreover, they changed the verb position from V2 to V-final sig. more often than from V-final to V2. We conclude that (i) V-final is the default placement in RCs; (ii) the acquisition of V-final precedes that of embedded V2. We argue that this preference for V-final is in line with economy-based strategies. In case of variation in the primary linguistic data children choose the underspecified value, because its licensing conditions form a superset of those of the other variant (V2).

go back up

Anaphoric priority modals in attitude contexts | handout

This talk investigates a semantic puzzle in German that features a sentence containing an intensional verb and a modalized relative clause. In an unspecific reading, this sentence not only has ambiguous features concerning the status of the relative clause but the modal inside the relative clause also seems to lack semantic contribution to the sentence meaning. In this talk I am going to discuss similar data to show that the puzzle raises misleading questions, and that it rather is the result of several independent aspects at interplay. It will turn out that we can solve the puzzle by a careful examination and comparison of different types of data. Additionally, I will present the true features of the phenomenon in question, and propose a simple analysis that can handle the newfound characteristics of the data. I will argue that an anaphoric relation between the two intensional verbs involved in the sentence is responsible for the reading to appear, which in some cases like in the original puzzle leads to the impression that the modal is semantically empty.

go back up

On Feature Mismatches in Relative Clauses | presentation

The topic of this talk are non-restrictive relative clauses (NRRCs) headed by a 1st/2nd person singular/plural pronoun as e.g. in (1)-(4):
(1) Ich (1sg.NOM), der (3sg.masc.NOM) schläft (3sg) / ?schlafe (1sg) | Ich, der ich schlafe (resumptive)
(2) Du (2sg.NOM), der (3sg.masc.NOM) schläft (3sg) / ?*schläfst (2sg) | Du, der du schläfst (resumptive)
(3) Wir (1pl.NOM), die (3pl.NOM) schlafen (1=3pl) | Wir, die wir schlafen (resumptive)
(4) Ihr (2pl.NOM), die (3pl.NOM) schlaft (2pl) / *schlafen (3pl) | Ihr, die ihr schlaft (resumptive)

pronoun-1/2sg/pl, who-sg/pl (resumptive-1/2sg/pl) sleep-,...

After providing an (experimentally founded) empirical basis for this RC type (in German) that substantiates the above judgements, namely that subject headed subject RCs with singular head nouns show 3rd person agreement whereas such with plural head nouns show (1st/)2nd person agreement (Trutkowski & Weiß 2016, contra Ito & Mester 2000, Vogel 2007/8), I will argue that this (and other) variation concerning agreement patterns of the finite verb within the RC occurs due to the phi/case feature inventory/specification of head noun and relative pronoun/complementiser. The findings can be explained within an approach that treats the RC as a second (agreeing) complement of the D head (cf. Sternefeld 2006).

go back up

Non-at-issueness and the appositive - continuative distinction

According to standard assumptions (McCawley 1982, Potts 2005, Simons et. al 2010), non-restrictive relative clauses (NRRCs) (i) always take widest scope with respect of matrix clause operators and (ii) are inherently non-at-issue. In this talk, we present two sets of experiments that challenge this view: one testing the non-at-issue status of German NRRCs, the other testing the embeddability of German NRRCs. In particular, we will argue that German NRRCs can be at-issue and can be embedded. As we will discuss, both, at-issueness and embeddability, are heavily affected by the position of the NRRC in the embedding sentence (sentence-internal vs. sentence-final) and the discourse-relation holding between the NRRC and the clause embedding it (coordinating vs. subordinating). For German, Holler (2005) assumes that there are two structurally different types of NRRCs, appositive and continuative NRRCs, which differ in position and discourse structure. We, thus, will raise the question, if the at-issueness and the embeddability of German NRRCs are affected by the type of the NRRC (appositive vs. continuative).

go back up

Relativizers are not complementizers: evidence from Romance | handout

In this talk we investigate the syntactic status of relative complementizers and relative pronouns in Old and Modern Italian, Italian varieties and Portuguese and we propose a unified analysis for both types of items. Based on recent proposals by Kato & Nunes (2009), Kayne (2010), Manzini & Savoia (2011) among many others, who question the strict dichotomy between relative complementizers and relative pronouns, we aim at showing that what we call complementizers are not C0 heads, as commonly assumed. Instead, we propose that both relative “complementizers” and “pronouns” have the same categorial status, i.e. they are nominal functional categories parallel to wh-determiners. The empirical arguments we bring for this conclusion are based on empirical studies in a diachronic and synchronic perspective of standard Italian and Italian dialects and Portuguese. We will show that what are usually dubbed as complementizers can actually display inflectional properties typical of pronouns and what are usually considered as pronouns can fail to do so. In addition, relative complementizers can show sensitivity to animacy and they can also occur as complements of prepositions in some Romance varieties. With respect to the diachronic development of relativizers, we will question the perspective that relative complementizers have developed out of relative pronouns. We argue that the diachronic pathway can be rephrased in terms of feature loss/or independent parallel development and does not necessarily imply a categorial change. We will also show that the link between interrogatives and relatives is rather strong, since the forms of [what/what N] occurring in wh-interrogatives are exactly the same found in relative clauses in over two-hundred Italian dialects and Portuguese, which shows that the generalization is rather robust. We will also discuss other theoretical arguments to the advantage of considering relativizers as nominal functional elements parallel to wh-determiners and not as the counterpart of declarative complementizers.

go back up

Relative Clauses as target for Long Extraction?

Based on findings from Andersson & Kvam (1984), long extraction into relative clauses has been considered to be ungrammatical in German (see Lühr, 1988). However, this claim has never been tested with more rigorous experimental methods. We ran a series of acceptability experiments investigating long extraction into relative clauses as well as long extraction targeting main clause and embedded clause wh-questions. The results show that long extraction into relative clauses is as acceptable as long extraction targeting wh-questions. An additional finding is that the individual differences with regard to the acceptability of long extraction correlate with those of an unrelated structure that is not grammatical in Standard German (verb clusters with the auxiliary in second position). This suggests that the individual variability concerning the acceptability of long extraction is partially due to individual differences concerning the use of non-standard structures.

go back up

Poster Session

Split-antecedent Relative Clauses and the Symmetry of Predicates (CON+NRR)

Verb placement and prosody in the acquisition of German relative clauses - Evidence from an elicited imitation experiment (ERW)

Ordinary Semantics (ERW+NRR)

Bare Plurals, prosody and the interpretation of German relative clauses (ERW)

go back up