Refine
Year of publication
Document Type
- Conference Proceeding (1775) (remove)
Language
Has Fulltext
- yes (1775) (remove)
Keywords
- Germanistik (54)
- Japanisch (36)
- Bologna-Prozess (32)
- Exzellenzinitiative (32)
- Theorie (32)
- Zukunft (32)
- Informationsstruktur (30)
- Koreanisch (30)
- Deutsch (24)
- Computerlinguistik (23)
Institute
- Physik (249)
- Medizin (197)
- Rechtswissenschaft (119)
- Universitätsbibliothek (97)
- Extern (71)
- Informatik (41)
- Geschichtswissenschaften (31)
- Gesellschaftswissenschaften (30)
- Frankfurt Institute for Advanced Studies (FIAS) (27)
- Erziehungswissenschaften (25)
The paper presents an analysis of control switch in German and Norwegian, as exemplified in the German pair "Ich verspreche ihm zu kommen" 'I promise him to come' vs. "Ich verspreche ihm kommen zu dürfen" 'I promise him to be allowed to come'. The phenomenon is induced by deontic modals in the context of suasive verbs of communication. The analysis is cast both in LFG and HPSG framework, in both cases deploying a pronounced feature-based semantic component. Our core assumption is that a normative agent is computed on top of control relations.
Capturing word order asymmetries in English Left-Peripheral Constructions: A Domain-Based Approach
(2003)
Left-Peripheral Constructions: A Domain-Based Approach Even though the word order in English is rather straightforward, the distributional possibilities of left-peripheral elements like topic phrases, wh-phrases, and negative operators (introducing an SAI) are quite intriguing and complex. In particular, there seems to exist no straightforward way of capturing the linear order asymmetries of these elements in the main and embedded clauses. The prevailing analyses have resorted to movement processes with multiple functional projections. The goal of this paper is to explore an alternative analysis to such movement-based analyses. In particular, this paper adopts the notion of topological fields (DOMAIN) proposed by Kathol (2000, 2001) within the framework of HPSG. The paper shows that within this DOMAIN system, the distributional possibilities as well as the asymmetries we find in English left peripheral constructions can easily follow from the two traditional views: (i) a topic precedes a focus element, and (ii) in English a wh-phrase and a complementizer competes with each other for the same position.
Proceedings of the 9th International Conference on Head-driven Phrase Structure Grammar : 2003
(2003)
This paper, in the context of multilingual MT, proposes the use of ICONS (Individual CONstraintS) to add a representation of information structure to MRS. The value of ICONS is a list of objects of type info-str, each of which has the features CLAUSE and TARGET. The subtypes of info-str indicate which information structural role is played by the TARGET with respect to the CLAUSE. This proposal is designed to support both the calculation of focus projection from underspecified representations and the handling of multiclausal sentences.
Yatabe (2021) presents a theory according to which the meaning of a word like 'different' in a sentence like 'Anna and Bill like different films' contains the meaning of a reciprocal pronoun. Since the postulated reciprocal meaning inside the meaning of a word like 'different' requires the presence of a semantic antecedent, the theory entails that the apparent internal reading of a sentence like 'John saw and reviewed different films', which does not contain a plural DP that could serve as the semantic antecedent of the postulated reciprocal meaning, must be licensed in a way that is entirely different from the way in which the internal reading of a sentence like 'Anna and Bill like different films' is licensed. In the present paper, I adduce additional pieces of evidence for this theory. In order to enhance the plausibility of the proposed theory, I also show how the collective interpretation of reciprocals and the interaction of reciprocals and cumulative interpretation can be accounted for within the theory.
Semantic Representations become useful resources for various multilingual NLP applications such as Machine Translation, Multilingual Generation, cross Lingual QA, to name a few. No Semantic Representation, to our knowledge, adopts vivakṣā (Speaker's intention) as a guiding principle for the representation. This motivates us to develop a new Semantic Representation system – "Universal Semantic Representation (USR)" – following Indian Grammatical Tradition (IGT) and Paninian grammar. Since USR is designed to be language-independent, we have currently taken up the task of generating English, Hindi, Tamil and Bangla from the USR. For English generation, the USR is mapped to ERG meaning representation (Flickinger, D. 1999) which is couched in Minimal Recursion Semantics (MRS). We use an off-the-shelf ACE generator that uses ERG as a resource-grammar for generating English. While designing the transfer module from USR to ERG-based MRS, we came across various Abstract Predicates (APs) in MRS representation as described in ErgSemantics_Basic (Flickinger et al., 2014). These APs are used to represent the semantic contribution of grammatical constructions or more specialized lexical entries such as compounding or the comparative use of more and so on. This paper presents the strategy for postulating the APs from the information given in USR and then reports the implementation of the transfer module keeping the focus on the postulation of APs. We get around 95% accuracy in postulating APs from USRs.
Soranî Kurdish can reference up to two arguments morphologically, a subject agreement marker and an incorporated object pronoun. One of the argument referencing morphs is verb-bound and occurs in a fixed position in the verb template (after the stem), while the other is a mobile morph that can occur either verb-internally (in second or last position) or verb externally. Either the subject agreement marker or the object incorporated pronoun can be verb bound or mobile morphs, depending on the tense and presence of an NP complement. Previous literature has analyzed mobile morphs as (VP) endoclitics. We argue that this is not the case as verb-external mobile morphs occur at the end of the last word of the least oblique NP complement and cannot attach to the last word of VP-internal PPs. We provide an edge-feature based analysis of verb-external mobile morphs and show that the same realizational rules account for the exponents of mobile morph features whether they occur verb-internally or verb-externally. We furthermore suggest that the dissociation between paradigm class (verb-bound or mobile morph) and syntactic status (subject or object; agreement marker vs. incorporated pronoun) challenges views that treat morphological structure as isomorphic to syntactic structure.
The paper examines borrowed instances of what we call emphatic superlative ever (ES-ever) into two Germanic languages (Dutch and German) and two Romance languages (French and Spanish). We base our study on extensive corpus data. We model the data in three stages ranging from constructional borrowing (Stage-1: el coolest job ever 'the coolest job ever'), via diaconstructions (Stage-2: la mejor canción ever 'the best song ever'), up to lexical borrowing (Stage-3: las portadas más photoshopeadas ever 'the most photoshoped portals ever'). We extend an earlier approach to social meaning in HPSG to borrowing.
How things become red in Mandarin Chinese? A case study of deadjectival change of state predicates
(2023)
This paper provides an HPSG analysis for the morphosyntax and the semantics of deadjectival change of state (CoS) verbs in Mandarin Chinese. We first show that adjectives are a distinct word class from verbs in Mandarin Chinese and argue for the derivation of CoS verbs from property concept adjectives. We then model this derivation with a lexical rule. Finally, since CoS verbs can be combined with another verb to form a resultative verb compound (RVC) to express caused CoS, we also propose a lexical rule to account for RVCs.
Maliseet-Passamaquoddy (Algonquian, New Brunswick and Maine, MP) employs a set of enclitic particles to express tense, aspect, and various adverbial notions. These occupy second position in a clause: they follow either the first word in the clause or the first constituent. Johnson and Rosen (2015) propose an analysis of clitic placement in Menominee (Algonquian, Wisconsin) that takes clitics to occupy a functional head in the left periphery, postulating movement of one item into a specifier position to the left of this functional head, thus leaving the clitics in second position. Here I propose an alternative account for MP in the framework of Sign-Based Phrase Structure Grammar (Sag 2012) that makes no use of functional heads and postulates no movement operations. Instead, clitic positions are determined by a small number of maximally simple constructional statements.
Murrinh-Patha, a polysynthetic Non-Pama-Nyungan language of Australia features competition of subject and object agreement markers for a particular position (i.e. slot 2), meaning that certain subject agreement markers are realised in this position, unless already occupied by overt object agreement markers. In their typology of variable morphotactics, Crysmann & Bonami (2016) cite the case of Murrinh-Patha as an instance of misaligned, conditioned placement. I shall propose a formal account of this positional competition in Murrinh-Patha within Information-based Morphology. To this end, I shall generalise the "pivot" features previously proposed for placement relative to the stem (Italian; Crysmann & Bonami, 2016) or the edge (Soranî Kurdish; Bonami & Crysmann, 2013; Salehi & Koenig, 2023) and show how this will facilitate the treatment of conditioned placement in Murrinh-Patha.
A number of types of Welsh subordinate clause are introduced by what looks like the preposition 'i' 'to', 'for'. Earlier research has shown that there are three different lexemes here. It is not unusual for a language to have homophonous lexemes, but these lexemes share a variety of properties, and also share properties with the preposition 'i'. The similarities and the differences among these lexemes can be captured if they are grouped together as four different realisations of a single 'super-lexeme' within the hierarchical lexicon.
The theory of respectively interpretation proposed in Yatabe and Tam (2021) "In defense of an HPSG-based theory of non-constituent coordination" (Linguistics and Philosophy 44, pp. 1-77) entails that Binding Conditions A and B need to be formulated as constraints on the form of semantic representations. It is possible to formulate the two binding conditions as such constraints if anaphoric relations are encoded in semantic representations in a way analogous to the way they are encoded in Discourse Representation Theory.
How to be a ham sandwich or an eel: The English deferred equative and the Japanese eel sentence
(2022)
In some languages including English and Japanese, a nominal predicate construction (NPC; "NP1 is NP2") has a marked variety—"open-ended-relation NPCs" (ONPCs), to label it—where the referents of the subject NP and the predicate NP are understood to be in some pragmatically prominent relation other than identity or inclusion (e.g. I'm the ham sandwich 'I'm the customer who ordered the ham sandwich'). The Japanese ONPC has been called the "eel sentence (eel construction)", after an oft-cited example involving unagi 'eel' as its predicate NP. The English ONPC is discussed in good detail by Ward (2004; "Equatives and deferred reference", Language 80) under the rubric of the "deferred equative". The ONPCs in the two languages can be naturally used only under limited discourse configurations, with the English one being more severely constrained than the Japanese one. This work develops semantic analyses of the two ONPCs that improve on previous accounts.
This paper considers Chinese quantifier scope, an important, outstanding area of Chinese linguistics. In particular, there are two open questions on the subject: (1) the guiding principles that determine (a) the scopal readings of quantifiers and (b) the sometimes mandatory co-occurrence of the universal quantifier mei (every) and the universal adverb dou, and (2) the semantic functions of mei and dou and their connection to the co-occurrence of these words.
We reappraise three prior accounts of these subjects, reason through their consequences on some exemplary data, offer a new explanation based upon concord, a mechanism that is commonplace in many languages, and formulate it in lexical resource semantics (LRS). We use two principles adapted from Richter and Sailer's (2004) analysis of negative concord, expanded with a new quantifier order constraint to generate a coherent answer to the two aforementioned questions.
Computational Grammars can be adapted to detect ungrammatical sentences, effectively transforming them into error detection (or correction) systems. In this paper we provide a theoretical account of how to adapt implemented HPSG grammars for grammatical error detection. We discuss how a single ungrammatical input can be reconstructed in multiple ways and, in turn, be used to provide specific, high-quality feedback to language learners. We then move on to exemplify this with a few of the most common error classes made by learners of Mandarin Chinese. We conclude with some notes concerning the adaptation and implementation of the methods described here in ZHONG, an open-source HPSG grammar for Mandarin Chinese.
In this paper, we deal with register-driven variation from a probabilistic perspective, as proposed in Schäfer, Bildhauer, Pankratz, Müller (2022). We compare two approaches to analyse this variation within HPSG. On the one hand, we consider a multiple-grammar approach and combine it with the architecture proposed in the CoreGram project Müller (2015) - discussing its advantages and disadvantages. On the other hand, we take into account a single-grammar approach and argue that it appears to be superior due to its computational efficiency and cognitive plausibility.
This paper investigates the variation of resultative serial verb constructions in Benue-Kwa languages. The main claim is that the variation can be explained assuming three versions of general lexicon rules which turn main verbs into complex predicates selecting for a second verb and attracting its arguments. Each language has a language specific version of these lexicon rules, enriched with language specific peculiarities to account for the specific behaviour of verbal inflection. The fact that not all of the lexicon rules do operate in each languages is another source of variation.
In this paper, I shall discuss a peculiar coordination construction in German, where the shared subject of the two conjuncts is not found peripheral, but is contained within the first conjunct. Following Höhle (1983), this construction is called "Subject Gaps in Finite/Fronted" clauses (SGF). I shall discuss previous accounts, both symmetric coordination approaches (Frank, 2002; Kathol, 1999), as well as asymmetric adjunction approaches (Büring & Hartmann, 1998). The analysis I shall propose will treat the construction as coordination semantically, yet assume a head complement structure that combines the licensing first conjunct with an incomplete (=slashed) coordinate structure complement. I shall show how this addresses the ATB condition, permits straightforward licensing of the subject gap, and provides better control over the second conjunct, thereby improving over the adjunct analysis.
Welsh noun phrases have had much less attention than Welsh clauses, and there are unresolved issues about the nature of possessors, attributive adjectives, and the definite article and agreement clitics. There is evidence, especially from agreement, that possessors are complements, evidence that attributive adjectives are adjoined to a preceding [LEX+] nominal constituent, and evidence that the definite article and agreement clitics are specifiers. The last of these positions makes it fairly simple to capture the relation between the definite article and agreement clitics and possessors. It is not difficult to formalize these ideas within HPSG.
The paper looks at constraints on non-wh relatives in Sorani Kurdish (Iranian) and English (Germanic). We argue that some of them are grammatical, whereas others introduce social meaning. We present a basic, lexicalist syntactic analysis and expand it with social meaning constraints. We propose that classical sociolinguistic variables have the status of conventional implicatures and the overall assessment of a style is treated as a particularized conversational implicature.
This paper considers the role of nonlocal amalgamation in a system of analyses for typologically diverse languages. Nonlocal amalgamation (Bouma et al. 2001) was suggested in particular to get rid of extraction rules in Pollard and Sag's (1994) analysis of long-distance dependencies. However, in implemented projects like the English Resource Grammar (Flickinger, 2000, 2011) and the Grammar Matrix (Bender et al., 2002, 2010), the extraction rules have been maintained, while nonlocal amalgamation is used for the analysis of phenomena like the easy-adjectives. Zamaraeva and Emerson (2020) argue that, if extraction rules are kept, then supporting the English easy-adjectives may be an insufficient reason to maintain nonlocal amalgamation in a cross-linguistic system like the Grammar Matrix, as it complicates the analysis of multiple question word fronting with flexible word order (in languages such as Russian [rus]). However, I present here a case of morphological marking of questions (in languages like Makah [myh]) which further motivates nonlocal amalgamation, as the analysis is remarkably more simple with it than it is without it. An analysis of morphological marking of questions needs to be part of a cross-linguistic system such as the Grammar Matrix as well as an analysis of multiple fronting, which adds a new tension at the level of the Matrix "core" and provides concrete material for discussion of issues ranging from empirical implementation of theoretical ideas like nonlocal amalgamation to the big question of how much of typological space a single system of grammar is expected to cover.
It is a typologically well-attested generalization that simple personal pronouns are avoided when the purpose is to signal semantic identity between coarguments of a predicate (Faltz, 1985; Comrie, 1999; Levinson, 2000; Haspelmath, 2008, forthcoming; Volkova & Reuland, 2014). Many linguists assume what I call the Unified View, where these pronoun disjointness effects come out as a byproduct of a single syntactic constraint, generally known as Principle B of the Binding Theory (Chomsky, 1981, 1986; Pollard & Sag, 1994; Manning & Sag, 1998; Pollard, 2005; Müller, 2021). This paper argues that the Unified View is mistaken because it is both too weak and too strong. As an alternative, I propose that pronoun disjointness effects stem from a conspiracy of three distinct factors – none of which is a syntactic universal: (i) a preference for expressing identity with coindexation rather than anchoring distinct indices to the same referent (Reinhart, 1983); (ii) a language-specific variant of HPSG’s Principle B; and (iii) a constraint on the morphosyntactic encoding of reflexive relations (Faltz, 1985; Reinhart & Reuland, 1993; König & Siemund, 2000).
In this paper, we propose to extend the Przepiórkowski's 2000 analysis of Long Distance Genitive of Negation to the same phenomenon in Lithuanian. We discuss the features that have their origin in Categorial Grammar. We then develop a novel analysis of the case alternation in Categorial Grammar incorporating features of the HPSG analysis. The two accounts show a surprising convergence in basic assumptions and predictions.
The formal analysis of idioms has been oscillating between approaches that emphasize the unit-like character of idioms and approaches that focus on the autonomy of the idioms' parts. In this paper, we summarize the main arguments for and against these two positions to then propose an account that tries to capture and combine the insights and advantages of both types of analysis. The resulting theory is heavily influenced by the approach taken in Riehemann (2001).
Minimizer strong NPIs such as ''lift a finger'' are known to be more restricted in their occurrence than weak NPIs like ''ever''. Sedivy 1990 points to contexts with a ''negative side message'' in which ''lift a finger'' can occur but ''ever'' cannot. The paper provides a short overview over the relevant contexts and proposes an extension of a representational theory of NPI licensing with the following components: First, an utterance content is introduced that enriches the primary truth-conditional content by conventional implicatures and generalized conversational implicatures. Second, ''ever''-type NPIs can be licensed by weak NPI licensors, but only in the primary truth-conditional meaning of an utterance. ''Lift-a-finger''-type NPIs can only be licensed in the scope of negation, but the licensing can be checked at the representation of the enriched meaning of an utterance.
The aim of this paper is to propose three improvements to the HPSG model theory. The first is a solution to certain formal problems identified in Richter 2007. These problems are solved if HPSG models are rooted models of utterances and not exhaustive models of languages, as currently assumed. The proposed solution is compatible with all existing views on the nature of objects inhabiting models. The second improvement is a solution to "Höhle’s Problem", i.e., the problem of massive spurious ambiguities in models of utterances. The third is a formalisation of Yatabe's (2004) analysis of the coordination of unlike categories, one that requires a second-order extension of the language for stating HPSG grammars.
The aim of this paper is to provide a syntactico-semantic analysis of hybrid coordination, in which what is coordinated are phrases bearing different grammatical functions and different semantic roles. The proposed account improves on previous HPSG analyses by giving up the assumption that all conjuncts are dependents of the same head and, more importantly, by taking into account the syntax–semantics interface and providing semantic representations. This aspect of the analysis builds on and generalizes previous HPSG work on polyadic quantification.
Resultative phrases are generally believed to conform to the Direct Object Restriction: that is, they describe the direct object if verbs are transitive. However, some exceptions have occasionally been reported, and this paper investigates the problem by focusing on resultative phrases that occur with the valency alternation verbs in Japanese and Mandarin Chinese. Verbs that license the locative alternation and locatum-subject alternation describe events that involve two arguments, the location and the locatum, which are perceived to concurrently undergo a change of state. It will be shown that resultative phrases with a valency alternation verb can be predicated of either argument regardless of whether it is expressed as direct object. Furthermore, resultative verbal suffixes in Mandarin, interpreted as description of either the location or the locatum, give rise to the locative alternation while their interpretation remains the same. Thus, it is claimed that in Japanese and Mandarin, the predication relation of resultative phrases is not determined by the grammatical function of arguments as generally believed, but rather by the lexical semantics of the verbs.
This talk addresses the puzzle why there are different 'need' verbs in Germanic languages, which all are lexically polysemous and which all display some extent of negative polar behaviour. Whereas all the uses of Dutch 'hoeven' are negative polar, Modern Swedish 'behöva' is mostly distributionally unrestricted and only in its epistemic uses negative polar. Data suggest that this is a result of a gradual erosion of NPI-hood. The diverging behaviour of 'need'-verbs in Germanic languages can be most accurately managed assuming that lexical polysemy involves type hierarchy in which the different uses inherit from an abstract entry that defines semantics all these uses share. Moreover, it is concluded that if there is an NPI feature it is mandatorily inherited to all to its descendants. In languages such as Dutch this feature has scope over all uses, in languages such as Modern Swedish. it only bears scope over the epistemic uses.
As observed at various occasions, the usage of epistemic adverbs in information seeking questions is by far more restricted than the usage of epistemic adjectives. Starting from Lyons (1977) this contrast was motivated assuming that different types of epistemic operators come with different semantics and scope positions in the utterance, namely objective vs. subjective epistemic modality. However it is not possible to define clear classes of objective epistemic modal operators in terms of clear diagnostics. It will be shown here that the contrast of acceptability is more accurately explained in terms of locality and binding properties of the variable for the attitude holder rendering the epistemic judgement. If locally bound, epistemic modal operators can be embedded, if not, they are subject to much stricter conditions in order to be interpretable.
The current study presents an HPSG analysis for verbal reduplication in Mandarin Chinese. After discussing its interaction with Aktionsarten and aspect markers, we argue that it is a morphological rather than syntactic process. We put forward a lexical rule for verbal reduplication in Mandarin Chinese and the different forms of reduplication are captured in an inheritance hierarchy. The interaction between verbal reduplication and aspect marking is handled by multiple inheritance. This analysis covers all forms of verbal reduplication in Mandarin Chinese and has none of the shortcomings of the previous analyses.
Saying and shaking 'No'
(2021)
In many instances, the head shake can be used instead of or in addition to verbal 'No'. Based on previous work on negation in dialogue, we observe head shaking as answer particles and as responding to an implicit or an exophoric (i.e., real world situation) antecedent. Exophoric head shake, however, seems to come in two flavours: with positive and with negative emotional valuation of the antecedent situation. We provide semantic analyses for all three uses (and a head nod) within an HPSG version which is implemented in Type Theory with Records and the dialogue framewok KoS. In particular, we extend on previous work by grounding ''exophoric negation'' in positive or negative appraisal. Finally, we briefly speculate about differences between verbal 'No' and head shaking due to (the lack of) simultaneity.
The paper argues that there is compelling evidence for analyzing copy raising in English as a lexical rule that converts a subtype of perception verb with a stimulus subject (so-called "flip-perception" verbs) into a semantically bleached verb of mild evidentiary force, roughly equivalent to seem in some uses, which identifies the index of its external argument with the index of the pronominally expressed external argument of its complement.
Expletive negation refers to constructions where a negator in the complement of certain lexical items does not change the polarity of the complement proposition. Jin & Koenig (2021) show that expletive negation occurs rather widely in languages of the world and in very similar environments. They propose a language production model of why such apparently illogical uses of negation arise in language after language. But their study does not address the grammatical status and representation of expletive negation. In this paper, we argue that expletive negation is part of the lexical knowledge speakers have of their language and that the negator in expletive negation constructions contributes a negation to a non-at-issue content associated with expletive negation triggers. We provide a Lexical Resource Semantics analysis of how triggers combine in a non-standard manner with the standard semantic content of their complements: the negation (and in some cases an additional modal operator) of the content of their complement is part of the trigger's non-at-issue content while the scope of the negation is an argument of the trigger's MAIN content. Finally, we suggest that the expletive use of the French negator ne includes a lexical constraint that requires it to modify a verb that reverse selects for an expletive negation trigger.
The topic of this paper is the expression of negative directives in several Romance languages. The majority of Romance languages do not express negative directives by adding (pre-verbal) negation to the positive imperative form, but by using a different verb form (infinitive, subjunctive or something else), to which negation is attached. The present analysis shows that (some) directive verbal forms in Romance lost some hallmarks of their verbhood. The phenomenon is taken as witnessing different stages of de-verbalisation. De-verbalisation makes directive verb forms similar to interjections. The variation documented in the Romance imperatives with respect to the compatibility/incompatibility with negation may thus seen as tendencies of different degrees of the imperatives to come closer either to the verb, or to the interjection. In the context of these tendencies, the incompatibility between negation and imperatives may be explained through the concept of marking. Put briefly, imperatives require to be marked by negation but negation is or is not able to mark them.
The Dutch and German verbs wijsmaken/weismachen 'make wise' have an idiomatic interpretation as verbs of deception 'to fool'. As such, they have the unusual property of being contrafactive (presupposing the falsity of their complement). With second person or generic pronoun subjects, under negation and with future orientation, they are used to express disbelief on the part of the entity denoted by the indirect object. A corpus study shows this secondary use to be especially prominent in Dutch. It depends on the availability of the point of view of experiencer and is most common with first person dative objects.
For the past 20 years, the question of a ''creole prototype'' has been at the center of heated debates. Among features that are claimed to be typical of creole formation is the placement of negation, which usually appears preceding tense, aspect and mood markers (e.g. McWhorter 2018). In this paper, I examine diachronic data, in particular, French compound tenses and show that the position of negation in at least French-related creoles is nothing but the result of regular grammaticalization given input. As such the expression of negation typically exemplifies coalescence of the already grammaticalized negator pas into an inflectional exponent in the creoles.
This paper presents a study of so-called neg-phrases in Eton, a negative concord language spoken in Cameroon. These phrases strongly resemble negated noun phrases that consist of a negative determiner and a noun, however, I will show that Eton neg-phrases are built differently. Reconciling the non-negative approach to negative indefinites by Penka & Zeijlstra (2005) and the negative approach by Richter & Sailer (2004a,b, 2006), I will argue that Eton neg-phrases consist of an inherently negative modifier and a non-negative indefinite derived from a noun. Embedding the analysis in Lexical Resource Semantics, I will reveal the inherent negativity of Eton neg-phrases and account for their composition by using a lexical rule based on the semantic approach to noun phrases by Beavers (2003).
Middle Welsh is a VSO language with the verb before the subject in all kinds of finite clause. However, positive declarative main clauses normally show verb-second order with a constituent of some kind before the finite verb. There are questions about the nature of this restriction. There are also questions about subject-initial sentences, which show surprising agreement properties, whether the subject is a topic or a focused constituent. All these questions can be given plausible answers within HPSG.
We present an analysis of multiple question fronting in a restricted variant of the HPSG formalism (DELPH-IN) where unification is the only natively defined operation. Analysing multiple fronting in this formalism is challenging, because it requires carefully handling list appends, something that HPSG analyses of question fronting heavily rely on. Our analysis uses the append list type to address this challenge. We focus the testing of our analysis on Russian, although we also integrate it into the Grammar Matrix customization system where it serves as a basis for cross-linguistic modeling. In this context, we discuss the relationship of our analysis to lexical threading and conclude that, while lexical threading has its advantages, modeling multiple extraction cross-linguistically is easier without the lexical threading assumption.
The Smurf comics series is, among others, famous for the so-called "smurf language", in which words or parts of words can be replaced by "smurf". We will argue that this "smurfing" has the properties of placeholding. Based on data from German translations of Smurf comics, we will provide a formalization of smurfing in German which can be generalized to a theory of placeholder expressions.
This paper sketches an analysis in Lexical Resource Semantics of adverbial and adjectival modification in nominal projections which is extensible to modification of other syntactic categories. It combines insights into the syntax-semantics interface of recursive modification in HPSG with underspecified semantics and type-logical meaning representations in the tradition of Montague grammar. The analysis is phrased in such a way that it receives a direct implementation in the Constraint Language of Lexical Resource Semantics as part of the TRALE system.
The indigenous languages of North America have played a critical role in discussions of the universality of part-of-speech distinctions. In this paper, we show that Oneida does not include a grammatical distinction between nouns and verbs. Rather, Oneida inflecting lexical items are subject to two cross-cutting semantic classifications, one that concerns the sort of entities they describe, the other the sort of semantic relation they include in their content. Labels such as ‘noun' and ‘verb' can still be used for cross-linguistic comparison, as the semantic partition of lexical items corresponds to canonical nouns and verbs according to morphologists and some typologists. But the meta-grammatical status of these labels is quite distinct from the status of corresponding labels in Indo-European languages like English.
This paper presents an incremental approach to verb clusters in German which radically differs from standard HPSG accounts. While the common assumption is that the verbs in subordinate clauses form clusters and accumulate all their valence requirements on a SUBCAT list, the assumption in this paper is that the arguments in verb final clauses are encapsulated incrementally into syntactic and semantic structures before the verbs are attached. The proposed analysis is in line with psycholinguistic findings. A grammar fragment of German demonstrating an implementation of the analysis is presented.
This paper proposes an account for the four auxiliaries in Kazakh that express the imperfective aspect. The main factors – the auxiliary, the main verb, their inflections and the aspectual specifications reveal a complicated system, which can be captured with an appropriate monotonic, multiple inheritance type hierarchy using online-type construction with the implementation of Pāṇinian competition. This analysis sheds light to a very different auxiliary system that we find in Indo-European languages.
Unrealized arguments in SBCG
(2020)
In null instantiation (NI) an optionally unexpressed argument receives either anaphoric or existential interpretation (Fillmore, 1986; Mauner & Koenig, 2000; Kay, 2002; Ruppenhofer & Michaelis, 2010, 2014). Examples include Lexically licensed NI (Nixon resigned.), Contextual accessibility NI (Can I see?), Labelese (Contains alcohol), Diary NI (Got up, got out of bed, dragged a comb across my head), Generic-habitual NI (The police only arrest (people) when there's probable cause). We think of a predicator as having NI potential when one or more of its frame elements may remain unexpressed under certain conditions. While one cannot accurately predict a predicator's NI potential based either on semantic factors (e.g., Aktionsart class of the verb, as in Hovav & Levin (1998)) or pragmatic factors (e.g., relative discourse prominence of arguments, as in Goldberg (2006)), NI potential, while highly constrained, is not simply lexical idiosyncrasy, but is instead the product of both lexical and constructional licensing. In the latter case, a construction can endow a verb with NI potential that it would not otherwise have. Using representational tools of Sign Based Construction Grammar (Sag 2012, a.o), we offer a lexical treatment of null instantiation that covers both distinct patterns of construal of null instantiated arguments and the difference between listeme-based and contextually licensed, thus construction-based, null complementation
Research on unbounded dependency constructions (UDCs) has focused mainly on the properties that are shared by all UDCs, but a satisfactory theory of syntax also needs to capture the properties that distinguish specific UDCs and the properties that are shared by some but not all of them. Three Welsh unbounded dependency constructions – wh-interrogatives, free relatives, and cleft sentences – are of interest here because they show a challenging array of similarities and the differences. However, given a slightly expanded hierarchy of phrase types, HPSG can capture both the similarities and the differences in this area.
In this paper, we provide a novel account of French causatives that crucially derives the core properties of the construction inside-out from the downstairs lexical verb to the causative verb, rather than outside-in, as is commonly assumed by argument composition (Miller & Sag, 1997; Abeillé & Godard, 1997; Abeillé et al., 1998). We shall argue on the basis of clitic trapping (Miller & Sag, 1997), as well as marking of the downstairs subject (Koenig, 1998) that the downstairs verb assumes a more active role than what is suggested by an argument composition approach and, conversely, we shall show that argument composition leads to problems with coordination and with en-cliticisation. The analysis we are going to propose combines an inversion analysis of the downstairs subject as a downstairs complement, accounting for scrambling and case marking, with an analysis of clitic climbing in terms of inflectional periphrasis (Aguila-Multner & Crysmann 2020).
Dutch has four pronouns "er" which show an intriguing pattern of syntactic haplology when a finite verb has more than one "er" dependent. We present a theory that captures this pattern by relying on two central aspects of HPSG: (i) the distinction between ARG-ST and COMPS and (ii) the distinction between canonical and non-canonical synsem objects. No deletion rules of the kind used in transformational analyses of "er" are necessary.
The paper proposes a representational re-encoding of the scalar, pragmatic accounts of NPI licensing within the framework of Lexical Resource Semantics (LRS). The analysis focuses on a less researched distribution pattern: emphatic NPIs occurring in result clause constructions that receive an intensification reading. We will provide a scalar extension of a standard semantic account of result clauses to capture the high degree interpretations. Our investigation will also offer new insights on NPI licensing in embedded clauses. We will primarily consider Romanian data.
Progress toward distinguishing clearly between generative and model-theoretic syntactic frameworks has not been smooth or swift, and the obfuscatory term 'constraint-based' has not helped. This paper reviews some elementary subregular formal language theory relevant to comparing description languages for model-theoretic grammars, generalizes the results to trees, and points out that HPSG linguists have maintained an unacknowledged and perhaps unintended allegiance to the idea of strictly local description: unbounded dependencies, in particular, are still being conceptualized in terms of plugging together local tree parts annotated with the SLASH feature. Adopting a description language with quantifiers holds out the prospect of eliminating the need for the SLASH feature. We need to ask whether that would be a good idea. Binding domain phenomena might tell us. More work of both descriptive and mathematical sorts is needed before the answer is clear.
Gapping in Japanese, which is an SOV language, differs from gapping in SVO languages in that the conjuncts with the elided verbs appear in non-final position. In this paper I present an incremental approach to gapping in Japanese, where it is assumed that an argument structure type is constructed in the non-final clause(s) in the gapping construction. This type is unified with the construction type created by the final clause resulting in identical construction types for all conjuncts in the construction.
This paper presents a formalization of proportional analogy using typed feature structures, which retains all key elements of analogical models of morphology. With the Kasem number system as an example, I show that using this model it is possible to express partial analogies which are unified into complete analogies. The analysis presented is accompanied by a complete TRALE implementation.
The Welsh copula has a complex set of forms reflecting agreement, tense, polarity, the distinction between main and complement clauses, the presence of a gap as subject or complement, and the contrast between predicative and equative interpretations. An HPSG analysis of the full set of complexities is possible given a principle of blocking, whereby constraints with more specific antecedents take precedence over constraints with less specific antecedents, and a distinction between morphosyntactic features relevant to syntax and morphosyntactic features relevant to morphology.
We examine the fine structure of clausal right-node raising constructions in Japanese, and argue that there are sentences in which a tensed verb is right-node-raised out of coordinated tensed clauses as well as sentences in which a verb stem is right-node-raised out of coordinated tenseless phrases. In the latter case, the tense morpheme has to be assumed to take a tenseless complement clause, and we note that the existence of such a structure contradicts the so-called lexicalist hypothesis, according to which a verb stem and the tense morpheme immediately following it always form a morphosyntactic constituent.
In this paper we describe insights gained from building an extension to the LinGO Grammar Matrix customization system to cover adnominal possessive phrases. We show how the wide range of such constructions attested in the world's languages can be handled with the typical major phrase types used in HPSG and discuss the value of feature bundling in the multilingual grammar engineering context.
This study aims to analyze and develop a detailed model of syntax and semantics of passive sentences in standard Indonesian in the framework of Head-Driven Phrase Structure Grammar (HPSG) (Pollard & Sag, 1994; Sag et al., 2003) and Minimal Recursion Semantics (MRS) (Copestake et al., 2005), explicit enough to be interpreted by a computer, focusing on implementation rather than theory. There are two main types of passive in Indonesian, following Sneddon et al. (2010, pp. 256-260) and Alwi et al. (2014, pp. 352-356), called 'passive type 1' (P1) and 'passive type 2' (P2). Both types were analyzed and implemented in the Indonesian Resource Grammar (INDRA), a computational grammar for Indonesian (Moeljadi et al., 2015).
The copula construction in Hebrew has received much attention in the linguistic literature. Nevertheless, one non-canonical variant has been largely neglected. In this variant the copula, flanked by two NPs, exhibits agreement with the post-copular NP, contrary to the canonical variant, where the agreement controller is the initial NP. This phenomenon challenges the notion of subject and its relation to agreement. The current corpus-based study investigates the word order and agreement patterns exhibited by the Hebrew copular constructions and shows that their distribution is largely motivated by information structure considerations. The proposed analysis accounts for the syntactic symmetry and semantic asymmetry between the two NPs.
In this paper, we argue that by making a more detailed distinction of theta-roles, while at the same time investigating the correlation of case marking, theta-role assignment, and eventuality types, we can describe different psych-verb subclasses and explain their alignment patterns in Spanish and Korean. We propose a neo-Davidsonian treatment of psych-verbs in HPSG that allows us to account for the underspecification of theta-roles which are modeled in an inheritance hierarchy for semantic relations. By assuming linking properties modeled lexically, we can constrain the properties for psych-verbs which shows the mapping of semantic arguments (i.e. experiencer, stimulus-causer, subject matter and target) to the elements in the argument structure. The type hierarchy and lexical rules proposed here capture the alternation in case marking not only of the experiencer (as traditionally assumed in the literature), but also of the stimulus. This analysis leads us to a new fourfold classification of psych-verbs for both languages.
The paper shows how the plural semantic ideas of (Sternefeld, 1998) can be captured in Lexical Resource Semantics, a system of underspecied semantics. It is argued that Sternefeld's original approach, which allows for the unrestricted insertion of pluralisation into Logical Form, suffers from a problem originally pointed out by Lasersohn (1989) with respect to the analysis offered by Gillon (1987). The problem is shown to stem from repeated pluralisation of the same verbal argument and to be amenable to a simple solution in the proposed lexical analysis, which allows for restricting the pluralisations that can be inserted. The paper further develops an account of maximalisation of pluralities as needed to obtain the correct readings for sentences with quantiers that are not upward monotone. Such an account is absent in the orginal system in (Sternefeld, 1998). The present account makes crucial use of the possibility to have distinct constituents contribute identical semantic material offered by LRS and employs it in an analysis of maximalisation in terms of polyadic quantication.
We present an analysis of clausal nominalization developed in the context of the LinGO Grammar Matrix (Bender et al., 2002, 2010) to support the addition of subordinate clauses to the grammar customization framework. In particular, we examine the typological variation of nominalized clausal complements and nominalized clausal modifiers. To account for the range of variation in nominalized clauses across the world's languages and to support linguists in exploring alternative analyses, we propose a flexible library of analyses, allowing nominalization of the clause to occur at the V, VP or S level.
The paper addresses verbal agreement in German sign language from a constraint-based perspective. Based on Meir's Agreement Morphology Principles it presents an HPSG analysis of plain, regular and backwards agreement verbs that models the interaction between phonological (manual) features and syntactico-semantic relationships within a verbal sign by well-defined lexical restrictions. We argue that a sign-based declarative analysis can provide an elegant approach to agreement in sign language since it allows to exploit cross-modular constraints within grammar, and hence permits a direct manipulation of all relevant phonological features of a verb depending on its syntactic and semantic properties.
I describe an analysis of valence-changing verbal morphology implemented as a library extending the LinGO Grammar Matrix customization system. This analysis is based on decomposition of these operations into rule components, which in turn are expressed as lexical rule supertypes that implement specific, isolatable constraints. I also show how common variations of these constraints can be abstracted and parameterized by their axes of variation. I then demonstrate how these constraints can be recomposed in various combinations to provide broad coverage of the typological variation of valence change found in the world's languages. I evaluate the coverage of this library on five held-out world languages that exhibit these phenomena, achieving 79% coverage and 2% overgeneration.
In this paper, we study Old French declension, a system which exhibits the theoretically challenging phenomenon of morphological reversal (Baerman, 2007). Furthermore, the declension system of Old French only recognises a single exponent -s, which marks different case/number combinations in different paradigms, contrasting with the unmarked form. We show that reversal is only one of several syncretism patterns found in the language and propose that Old French declension is best understood in terms of two systematic syncretisms: a natural split between singular and plural for feminines, and a Paninian split for masculines that systematically marks the objective plural. Reversal, and other seemingly morphomic splits arise as a result of idiosyncrasy in the NOM.SG cell, comprising inflection class-specific s-marking, as well as stem alternation and overabundance. We provide a formal analysis in terms of Information-based Morphology (Crysmann & Bonami, 2016) that effortlessly captures the systematic splits, as well as the variation in the nominative singular. We suggest that the high degree of idiosyncrasy in this cell paired with the reduced frequency of overt nominative NPs when compared to objective NPs may serve to explain why the system was actually quite short-lived.
Dutch is well-known for the formation of verb clusters. A characteristic aspect of such constructions is that the order of the verbs may differ from the order in which they are selected. Across the Dutch language area verb clusters show different types of word order variation. This paper proposes a constructivist account of word order variation in Dutch verb clusters. Linearization is not modelled in terms of the GVOR feature, after Kathol (2000). Instead, it relies on the bidimensional phrase hierarchy initiated by Ginzburg & Sag (2000), which is extended for the analysis of constructions with verb clusters. This proposal accounts for the most common instances of word order variation in Dutch verb clusters, and it can be easily adapted to model a specific variety or dialect.
This paper discusses the syntactic properties of 'prepositional numeral constructions (PNCs)' in English, which is exemplified by 'about 250 babies' and 'over 16,000 animals'. In PNCs a preposition is followed by a numeral. Previous analyses have claimed that the preposition and the numeral make a prepositional phrase in PNCs, but we argue that this is not a satisfactory approach. In HPSG there are some possible analyses that might be proposed, but there are reasons for supposing that the best analysis is one in which the preposition is a functor, a non-head selecting a numeral head.
This paper investigates the syntax of the English "not only ... but also ..." construction, focusing on the linearization possibilities of "not only". Based on novel corpus data, I argue that the "not only ... but also ..." construction exhibits different properties from the "not ... but ..." construction or the adverbial "only". I propose that a linearization-based account, along with coordinate ellipsis, can explain the various linearization possibilities of "not only". I also propose that the construction as a whole is a subtype of the type correlative-coord-ph, which is a novel subtype of coord-ph. Finally, I argue that subject-auxiliary inversion triggered by the clause-initial not only is a new subtype of the type "negative-inversion-ph".
Partial inversion in English
(2017)
A typical finite clause in English has a single constituent that serves as subject. This constituent precedes the finite verb in non-inverted clauses like simple declarative clauses, follows the finite verb in inverted clauses like polar questions, agrees in person and number with the finite verb and with a tag subject when a tag is present, undergoes subject raising, and so on (Postal 2004). Five constructions violate these generalizations and in the literature have called into question the identity of the subject constituent. In each of these five constructions the finite verb agrees with a following constituent in a declarative clause despite the fact, among others, that the constituent preceding the verb exhibits subject behaviors of the kind identified by Keenan (1976). To the authors' knowledge, despite intensive analysis of several of these patterns, the group as a whole has not been subject to prior study. The constructions are: Presentational Inversion (e.g., On the porch stood marble pillars), Presentational there (e.g., "The earth was now dry, and there grew a tree in the middle of the earth"), Deictic Inversion (e.g., "Here comes the bus"), Existential there (e.g., There’s a big problem here) and Reversed Specificational be (e.g., "The only thing we’ve taken back recently are plants"). The approach of Sign-Based Construction Grammar (Sag 2012) enables us to establish precisely what all five patterns have in common and what is particular to each, revealing that a constructional, constraint-based approach can extract the correct grammatical generalizations, not only in 'core' areas of a grammar, but also in the hard cases, where concepts such as subject, which readily handle the more tractable facts, fail to fit the facts at hand. We see further that the five split-subject patterns, sometimes identified as clausal, yield to a strictly lexical analysis.
Against split morphology
(2017)
In this paper I present data from several Niger Congo languages, illustrating how the paradigms which make up the noun class systems of these languages are problematic to analyze within traditional morphosyntactic frameworks. I outline possible solutions to this problem, and argue for the introduction of an exemplar based Word and Paradigm (Blevins 2006) approach to morphology within SBCG. I then outline the consequences of this approach for the structure of the SBCG lexicon.
In this paper I present an incremental approach to gapping and conjunction reduction where it is assumed that the first sentence in these constructions is fully parsed before the second sentence with the elided verb is parsed. I will show that the two phenomena can be given a uniform analysis by letting the construction type of the first conjunct be carried over to the second conjunct. This construction type imposes constraints on the arguments that the second conjunct can have. The difference between gapping and conjunction reduction is captured by the already existing constructions for sentence and VP coordination. The analysis is implemented in an HPSG grammar of Norwegian.
This paper explores the conundrum posed by two different control constructions in Yucatec Maya, a Mayan language spoken by around 800,000 speakers in the Yucatán Peninsula and northern Belize. Basic syntactic structure of the language is introduced, and a general SBCG treatment of control in YM is presented, alongside with an example of motion verbs as control matrices. The unruly case of intransitive subjunctive control, where the controllee appears with an unexpected status (incompletive) and without set-A morphology, is discussed and a proposal to treat it as nominalization is evaluated. The nominalization proposal is rejected based on the following grounds: (1) nominalization tends to attract definitive morphology, which is absent from intransitive subjunctive control constructions, (2) nominalization does not truly explain the lack of set-A morphology if one desires to provide a unified account of set-A morphemes, (3) verbs bereft of otherwise expected set-A morphemes have an independent motivation in the form of agent focus constructions.
In this paper we discuss two contrasting views of exponence in inflectional morphology: the atomistic view, where content is associated individually with minimal segmentable morphs, and the holistic view, where the association is made for the whole word between complex content and constellations of morphs. On the basis of data from Estonian and Swahili, we argue that an adequate theory of inflection should be able to accomodate both views. We then show that the framework of Information-based Morphology (Crysmann and Bonami, 2016) is indeed compatible with both views, thanks to relying on realisation rules that associate m units of forms with n units of content.
Over the past few years, there has been renewed interest in the treatment of resumption in HPSG: despite areas of convergence, e.g. the recognition of resumptive dependencies as dependencies, as motivated by Across-the-Board (ATB) extraction, there is no unified theory to date, with differences pertaining, e.g., to the exact formulation of amalgamation (Ginzburg and Sag, 2000), or the place of island constraints in grammar. While Borsley (2010) and Alotaibi and Borsley (2013) relegate the difference in locality of gap and resumptive dependencies to the performance system, Crysmann (2012, 2016) captures insensitivity to strong islands as part of the grammar. Harmonising existing proposals becomes even more acute, if we consider the cross-linguistic similarity of the phenomenon, in particular, if we compare languages like Hausa and Arabic, which both feature island insensitivity to some degree, as well as bound pronominal resumptive objects and zero pronominal resumptive subjects, to name just a few of the parallels. In this paper, I shall reexamine resumption (and extraction) in Modern Standard Arabic (henceforth: MSA) and propose a reanalysis that improves on Alotaibi and Borsley (2013) in several areas: first, I shall argue that controlling the distribution of gaps and resumptives by means of case is not only empirically under-motivated but also leads to counter-intuitive constraint specifications in the majority of cases. Second, I shall show that the case-based account of Alotaibi and Borsley (2013) can be straightforwardly supplanted with the weight-based account I proposed in Crysmann (2016): in doing this, one does not only get a better alignment of case assignment constraints with overtly observable manifestations of case, but such an account is also general enough to scale from case languages, such as MSA, to languages without case, such as Hausa, or many Arabic vernaculars. Finally, I shall address case in ATB extraction and propose a refinement of the Coordination Constraint of Pollard and Sag (1994) that accounts for exactly the kind of mismatch observed in mixed gap/resumptive ATB extraction
Explanations and "engineering solutions"? Aspects of the relation between Minimalism and HPSG
(2017)
It is not simple to compare Minimalism and HPSG, but it is possible to identify a variety of differences, some not so important but others of considerable importance. Two of the latter are: (1) the fact that Minimalism is a very lexically-based approach whereas HPSG is more syntactically-based, and (2) the fact that Minimalism uses Internal Merge in the analysis of unbounded dependencies whereas HPSG employs the SLASH feature. In both cases the HPSG approach seems to offer a better account of the facts. Thus, in two important respects it seems preferable to Minimalism.
The Polynesian language Tongan appears to lack surface-oriented motivation for a VP constituent. Even so, adverbial elements appear in both a rightwards location and a leftwards location, superficially similar to the S-adverbs and VP-adverbs in well-studied western European languages. This paper explores how the Tongan ''VP-adverbs'' (as well as others) can be analyzed in HPSG without a VP for those adverbs to attach to. Several kinds of analyses, representing different strands of research on the syntax of adjuncts in HPSG, are explored: a Adjuncts-as-Valents analysis, a VAL-sensitive Adjuncts-as-Selectors analysis, and a WEIGHT-sensitive Adjuncts-as-Selectors analysis. All suggest that an analysis of the adverbs without a VP is possible; a WEIGHT-sensitive Adjuncts-as-Selectors seems to have the fewest issues.