Linguistik
Refine
Year of publication
Document Type
- Article (1269)
- Part of a Book (784)
- Conference Proceeding (646)
- Working Paper (254)
- Review (181)
- Preprint (122)
- Book (109)
- Part of Periodical (64)
- Report (58)
- Doctoral Thesis (23)
Language
- English (1931)
- German (1055)
- Croatian (298)
- Portuguese (120)
- Turkish (43)
- Multiple languages (24)
- French (21)
- mis (16)
- Spanish (7)
- Polish (4)
Has Fulltext
- yes (3528) (remove)
Keywords
- Deutsch (436)
- Syntax (152)
- Linguistik (127)
- Englisch (123)
- Semantik (112)
- Spracherwerb (101)
- Phonologie (86)
- Rezension (77)
- Fremdsprachenlernen (69)
- Kroatisch (68)
Institute
- Extern (438)
- Institut für Deutsche Sprache (IDS) Mannheim (113)
- Neuere Philologien (43)
- Sprachwissenschaften (43)
- Universitätsbibliothek (5)
- Sprach- und Kulturwissenschaften (3)
- Gesellschaftswissenschaften (2)
- Medizin (2)
- Präsidium (2)
- SFB 268 (2)
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.
This paper is the third in a series of papers dedicated to the investigation of subjunctive complement clauses in Modern Standard Arabic. It began with Arad Greshler et al.'s (2016) search for obligatory control predicates in the language and continued with Arad Greshler et al.'s (2017) empirical and theoretical investigation of the backward control construction. In this paper we show that Arad Greshler et al.'s (2017) findings and ultimate analysis, which is cast in a transformational framework, can be straightforwardly formalized using the existing principles and tools of HPSG. Our proposed analysis accounts for all the patterns attested with subjunctive complement clauses in Modern Standard Arabic, including instances of control and no-control.
This paper investigates the structure and agreement of coordinated binominals in the form Det N1 et N2 in French. We provide corpus data and experimental data to show that different strategies exist, depending on their readings: singular Det for joint reading (mon collègue et ami, 'my.MSG colleague.MSG and friend.MSG'), plural Det agreement (mes frère et soeur 'my.PL brother.MSG and sister.FSG') or closest conjunct agreement (mon nom et prénom, 'my.MSG surname.MSG and first name.MSG') for split reading. These results challenge previous syntactic analyses of binominals (Le Bruyn and de Swart, 2014), stating that Det combines with N1, forming a DP and the later coordinates with N2. We then propose an HPSG analysis to account for French binominals.
Modern Standard Arabic (MSA) has simple and complex comparatives, which look rather like their counterparts in many other languages. MSA simple comparatives are indeed like those of other languages, but MSA complex comparatives are quite different. They involve an adjective with a nominal complement, which may be an adjectival noun or an ordinary noun, and are rather like so-called 'adjectival constructs'. Simple comparatives, complex comparatives, and adjectival constructs can all be analysed with lexical rules within HPSG.
In this paper, it is demonstrated that there is a phenomenon that can be viewed as a mirror image of medial right-node raising and thus might be designated as medial left-node raising, and it is argued that the properties of this phenomenon are consistent with the predictions of the HPSG-based theory of non-constituent coordination first proposed in Yatabe (2001) and modified in later works such as Yatabe (2015).
Right-node raising is usually set apart from other elliptical constructions for imposing a strict identity condition between the omitted and the peripheral elements. Since Pullum & Zwicky (1986), it is assumed that only syncretic forms may resolve a feature conflict between the two conjuncts (I certainly will and you already have set the record straight.). We present an empirical study of RNR with final verb in English and French that shows that verb mismatch does occur in corpora with and without syncretic forms, i.e. that syncretism does not appear to play a role. We present an acceptability judgement task on French that confirms this hypothesis. We therefore propose a new HPSG analysis of RNR that is based on sharing LID features and not morphophonological forms.
We look at definite marking in Esperanto, Papiamentu, and Yiddish considering three semantically definite contexts: the referential use of proper names and unique nouns, as well as anaphoric definites. We argue for a three-dimensional analysis of definiteness: an individual denotation, an existence presupposition, and a uniqueness conventional implicature. We present an HPSG encoding of this system and model the central aspects of the definite marking systems of our three object languages.
This paper outlines a new analysis of the syntactic structure and discourse function of a ‘prominent internal possessor construction' (PIPC) in Chimane (unclassified, Bolivia) and compares it with an existing analysis of a different kind of PIPC found in Maithili (Indo- Aryan, India/Nepal). PIPCs in Chimane and Maithili involve an apparently non-local agreement relation between verbs and possessors which are internal to possessive NPs. In Chimane, it is argued that internal possessors are able to control object agreement via a clause-level 'proxy' of the internal possessor – see also Ritchie (under review). The paper goes on to compare this construction with PIPCs in Maithili, and shows that speakers use PIPCs in discourse to indicate the information structure role of the internal possessor. In the case of Chimane, it seems that internal possessors which bear the secondary topic role are more likely to control object agreement, while in Maithili, other semantic and information structural features of internal possessors are at play. The contributions of the various levels of sentence structure are modelled using the LFG architecture developed in Dalrymple & Nikolaeva (2005; 2011).
Early work on quantification in natural languages showed that sentences like 'Every ape picked different berries', on the reading that the sets of berries picked by any two apes are not the same, can be logically represented with a single polyadic quantifier for the two nominal phrases. However, since that quantifier cannot be decomposed into two quantifiers for the two nominal phrases, a compositional semantic analysis of this reading is not possible under standard assumptions about syntax and semantics. This paper shows how a constraint-based semantics with Lexical Resource Semantics can define a systematic syntax-semantics interface which captures the reading in question with a polyadic quantifier.
This paper concerns the argument structure analysis of raising-to-subject with passive predicates in Swedish and other Germanic languages. Support is given for the analysis in which the raising-to-subject construction constitutes a regular passive, the passive counterpart of active raising-to-object. The fact that there does not seem to be an active counterpart for certain predicates, such as the predicate say, as well as the fact that raising-to-subject does not seem to be possible with the periphrastic passive in Swedish is attributed to certain semantic restrictions on the raising-to-object construction and the periphrastic passive construction, respectively.
The paper briefly reexamines arguments for the argument–adjunct dichotomy, commonly assumed in contemporary linguistics, showing that they do not stand up to scrutiny. It demonstrates that – perhaps surprisingly – LFG currently only assumes this dichotomy in its f-structure feature geometry, and does not rely on it in any crucial way. Building on this observation, the paper presents a way of getting rid of this dichotomy altogether.
The aim of this paper is to reexamine the rich repertoire of grammatical functions assumed in LFG and provide novel arguments for the claim, voiced earlier for example in Alsina et al. 2005, that most of them are redundant. We also demonstrate that a textbook LFG test for the sameness of grammatical functions of different predicates fails on closer scrutiny. Constructively, we propose a more constrained approach to grammatical functions, which, however, has the advantage of formalising the grammatical function hierarchy, assumed in LFG analyses of diverse phenomena but apparently not previously formalised.
This paper presents a new analysis of quirky subjects according to which quirky subjects bear multiple grammatical relations and hence differ syntactically from regular subjects. This contrasts with the standard analysis of quirky subjects according to which quirky subjects are regular subjects bearing lexical case and therefore differ only morphologically from regular subjects. Based on the behavior of quirky subjects in Faroese and German, I argue that the syntactic account is superior. Faroese shows that the case borne by a quirky subject is not lexical, whereas German shows that quirky subjects are not regular subjects to begin with. The behavior of quirky subjects in Icelandic, on which the standard analysis is based, is argued to be the result of a morphosyntactic peculiarity of Icelandic.
Quantifiers canonically attach to nouns or noun phrases as modifiers to specify the amount or number of the entity expressed by the noun. However, it has been observed that quantifiers can be positioned outside of the noun phrase. These so-called floating quantifiers (FQs) exhibit intriguing syntactic and semantic characteristics. On the one hand, they appear to have a closerelationship with a noun; semantically they quantify a noun in the same way as non-floating quantifiers, and quite often they exhibit agreement with the noun. On the other hand, their phrase structure distribution is very similar to that of VP-adverbs. In this paper, we argue that the distribution of FQs is constrained not purely by syntax, but also by information structure. We show that FQs play a focus role whereas modified nouns are reference-oriented topic expressions. Building upon Dalrymple and Nikolaeva's (2011) recent proposal, we formulate the interaction between syntactic, semantic and information structure features of FQs within LFG's projection architecture.
This paper discusses recent LFG proposals on resultative and benefactive constructions. I show that neither resultative nor benefactive constructions are fully fixed and that this flexibility requires traces or a stipulation of constructional templates at several unrelated places in the grammar, something that is not necessary in lexical approaches. A second part of the paper deals with the active/passive alternation and shows that language-internal generalizations are missed if constraints are assumed to be contributed by phrase structure rules. A third part examines the parallel constructions in German and shows that cross-linguistic generalizations are not captured by phrasal approaches.
We want to show how basic copula clauses in Indonesian can be dealt with within the framework of Head Driven Phrase Structure Grammar (HPSG) (Pollard & Sag, 1994). We analyzed three types of basic copula clauses in Indonesian: copula clauses with noun phrase complements (NP) expressing the notions of 'proper inclusion' and 'equation', adjective phrases (AP) expressing 'attribution', and prepositional phrases (PP) expressing relationships such as 'location'. Our analysis is implemented in the Indonesian Resource Grammar (INDRA), a computational grammar for Indonesian (Moeljadi et al., 2015).
A singular countable noun in English normally requires a determiner and they should agree in number. However, there is a type of noun phrase, such as those thousand teachers, which does not conform to this generalisation. As a subtype of singular countable noun, thousand requires a determiner, but the determiner has number agreement with the head noun teachers. The standard HPSG treatment, in which the determiner requirement and the determiner-noun agreement are both represented in the SPR specifications of the head noun, cannot capture this special agreement pattern. Our analysis, in which the determiner requirement and the determiner-noun agreement are dissociated from each other, can provide a straightforward account of the data.
The phenomenon of so-called 'mixed' categories, whereby a word heads a phrase which appears to display some features of one lexical category, and some features of another, raises questions regarding the criteria used for distinguishing syntactic categories. In this paper I critically assess some recent work in LFG which provides 'mixed category' analyses. I show that three types of evidence are typically utilized in analyses of supposed mixed category phenomena, and I argue that two of these are not, in fact, crucial for determining category status. I show that two distinct phenomena have become conflated under the 'mixed category' heading, and argue that the term ‘mixed category’ should be reserved for only one of these.
Verbal present participles in Norwegian: Controlled complements or parts of complex predicates
(2016)
Norwegian has a limited option for verbal present participles. These participles only exist with a small number of verbs, and they are selected by a handful of predicates. The analysis of sentences with these participles raises some challenges. Taking the analysis of Thurén (2008) as my point of departure, I argue that verbal present participles have two possible analyses, as controlled complements, or as parts of complex predicates. The presentational focus construction gives important evidence for this analysis.
This paper points out certain flaws in the semantics for lexical rule specifications developed in Meurers (2001). Under certain circumstances, certain words may not be licit inputs to a rule according to this semantics while one would expect them to be from inspecting the specification of the rule. The reasons for this are shown to be that whether properties of paths should be transferred from the input of a rule to its output is decided considering only the respective paths and their properties in isolation, ignoring the ‘non-local’ effects that transferring their properties can have. Furthermore, the semantics is insensitive to the possible shapes of inputs to the rule, which also makes it possible that inputs of certain shapes are unexpectedly not accepted. An alternative semantics is developed that does not suffer from these deficits.
This paper desribes four areas in which grammar engineers and theoretical linguists can interact. These include: using grammar engineering to confirm linguistic hypotheses; linguistic issues highlighted by grammar engineering; implementation capabilities guiding theoretical analyses; and insights into architecture issues. It is my hope that we will see more work in these areas in the future and more collaboration among grammar engineers and theoretical linguists. This is an area in which HPSG and LFG as a distinct advantage, given the strong communities and resources available.
The paper considers a phenomenon in Korean where ambiguity in the written language is resolved prosodically. An LFG analysis is provided which extends the proposals of Mycock and Lowe (2013) to Korean, based on experimental evidence on the prosodic expression of focus in Korean which challenges the phrase-boundary based account of Jun and Oh (1996), and suggests that considering expanded pitch range may give a more robust account of focus expression.
In this paper we discuss second position clitics in Ancient Greek, which show a remarkable ability to break up syntactic constituents. We argue against attempts to capture such data in terms of a mismatch between c-structure yield and surface string and instead propose to enrich c-structure by using a multiple context free grammar with explicit yield functions rather than an ordinary CFG.
In this paper we propose an LFG/XLE treatment of Exhaustive Object Control (EOC) constructions in Greek na clauses. We draw on data retrieved from the Hellenic National Corpus (HNC) in order to define the verbs that allow EOC. We treat EOC using anaphoric control. We take the subject of the subordinate na clause (controllee) to be a PRO marked with nominative case that is anaphorically related to the object of the matrix clause (controller). We implement this analysis in our LFG/XLE Grammar by adding the new feature ANAPH_C_BY.
The aim of this paper is to provide an adequate analysis in LFG of the prepositional passive, e.g. That problem has been dealt with, My pen has been written with. This construction has been examined in LFG before by Bresnan (1982), Lødrup (1991), and Alsina (2009), but empirical and theoretical problems, some well-documented, some new, mean that such proposals cannot be maintained. Instead, I offer an account couched in recent work on the mapping between grammatical functions and arguments (Asudeh et al., 2014; Findlay, 2014a) that treats the defining characteristic of the prepositional passive not as purely syntactic, but rather as being located at the interface between syntax and semantics.
The aim of this paper is to tease apart two available views of the VP in Persian. The prevailing view of the Persian VP initially suggested in generative studies assumes a hierarchical structure with two object positions, mainly motivated by the existence of differential object marking in Persian. Building on quantitative studies, we revisit this hierarchical view and show that it is not born out by the data. A flat structure view of the VP, on the contrary, is in line with the data.
Previous accounts of the perfect tense-aspect in the K'ichee'an languages have concluded that the category or part-of-speech of the perfect is a verb, or less often, a participle. We believe otherwise. Empirical support is presented for the hypothesis that the perfect is expressed using either a deverbal participial adjective or a deverbal possessed nominal in the form of a detransitivized non-verbal predicate. We show that the perfect always consists of a one-place intransitive but that it, nonetheless, retains the capacity to express two argument roles. Further, we argue that the perfect is, in fact, a perfect. We present the various semantic types of perfect, including the perfect of result and the experiential perfect, and also show the temporal restrictions that constrain the perfect. The analyses are implemented using the syntactic architecture of LFG.
Within recent work on the treatment of resumption in HPSG, there is growing consensus that resumptive unbounded dependency constructions (=UDCs) should be modelled on a par with gap-type UDCs (Alotaibi and Borsley, 2013; Borsley, 2010; Crysmann, 2012b; Taghvaipour, 2005), using a single feature for both types of dependencies, rather than separate features, as proposed by Vaillette (2001a,b). Yet, authors disagree as to where exactly in the grammar the resumptive function of pronominals should be established: while Crysmann (2012b, 2015) advances an ambiguity approach that has pronominal synsem objects being ambiguous between a resumptive and an ordinary pronoun use, Borsley (2010); Alotaibi and Borsley (2013), by contrast, treat all pronominals, resumptive or not, as ordinary pronouns and effect their resumptive use by means of tailoring the amalgamation principle to potentially include pronominal indices. While their decision provides a straightforward account of McCloskey’s generalisation that resumptives always look like the ordinary pronouns of the language, it fails to capture the difference in semantics between ordinary pronominal and resumptive uses. In this paper, I shall reexamine the evidence from Hausa and propose to synthesise the approaches put forth by Alotaibi and Borsley (2013) and Crysmann (2012b), and propose that the potential for pronominal and resumptive function (including their difference w.r.t. semantics and non-local features) is captured by means of underspecification, yet the decision as to canonical vs. non-canonical use is made at the level of the governing head (Borsley, 2010; Alotaibi and Borsley, 2013). I shall argue that this division of labour is sufficient to derive the correct gap-like semantics for resumptives, maintains standard deterministic amalgamation, and, finally, provides an answer to McCloskey’s generalisation.
The Free Linguistic Environment (FLE) project focuses on the development of an open and free library of natural language processing functions and a grammar engineering platform for Lexical Functional Grammar (LFG) and related grammar frameworks. In its present state the code-base of FLE contains basic essential elements for LFG-parsing. It uses finite-state-based morphological analyzers and syntactic unification parsers to generate parse-trees and related functional representations for input sentences based on a grammar. It can process a variety of grammar formalisms, which can be used independently or serve as backbones for the LFG parser. Among the supported formalisms are Context-free Grammars (CFG), Probabilistic Contextfree Grammars (PCFG), and all formal grammar components of the XLEgrammar formalism. The current implementation of the LFG-parser includes the possibility to use a PCFG backbone to model probabilistic c-structures. It also includes f-structure representations that allow for the specification or calculation of probabilities for complete f-structure representations, as well as for sub-paths in f-structure trees. Given these design features, FLE enables various forms of probabilistic modeling of c-structures and f-structures for input or output sentences that go beyond the capabilities of other technologies based on the LFG framework.
We discuss agreeing adverbs in Urdu, Sindhi and Punjabi. We adduce crosslinguistic evidence that is based mainly on similar patterns in Romance and posit that there is a close connection between resultatives and so-called pseudo-resultatives, which the agreeing adverbs appear to instantiate. We propose a diachronic relationship by which the originally predicative part of a resultative is reinterpreted as an adjunct that modifies the overall event predication, not just the result.
In Libyan Arabic, the preposition fi 'in' has developed into a marker of continuous or habitual aspect. While structurally remaining a preposition which marks the objects of the non-tensed forms of dynamic transitive verbs, it serves to attribute an aspectual interpretation to the clause as a whole. We argue that this aspectual object marking is naturally modeled by an inside-out functional designator, and provide arguments that the aspectual value contributed by aspectual fi is best treated as an f-structure feature.
Dargwa languages have two types of agreement at clause level: gender and person agreement. In the general case, person agreement is hierarchical (speech act participants prefered to 3rd persons), while gender agreement is with the absolutive (S/P) argument. Two exceptions to this pattern have been observed in some dialects: first, some auxiliary verbs have a gender agreement slot which can be controlled by both ergative and absolutive arguments; second, adverbials agreeing in gender can agree with either ergative or absolutive if they are located at clause edges. A proposed explanation of this behaviour is through effectively splitting each clause into two layers, with the top layer having its own zero absolutive position, coreferential with either the subject or the direct object of the lower layer. In this way, the general rule that gender agreement is with the absolutive can be preserved. In this paper, I argue that the data of Ashti Dargwa do not support the Backward Control theory. Peripheral adverb agreement and auxiliary gender agreement are independent phenomena, while auxiliary agreement can be explained by splitting the 3rd person based on topicality, as in proximateobviative systems. This allows us to preserve the conventional account of clause structure while framing the data of Dargwa in a wider typological context.
Languages differ in how they employ finite and non-finite clauses. Welsh finite and non-finite clauses have a similar distribution to their counterparts in English. However, it doesn’t look like this because Welsh has certain finite clauses which look rather like non-finite clauses. We examine two types of pseudo-non-finite clauses: finite "bod" clauses and finite "i" clauses. We argue that both cases are instances of a mismatch between syntax and morphology, while the latter only involves periphrasis. We provide an HPSG analysis capturing similarities and differences between these two constructions and canonical finite and nonfinite clauses.
This paper discusses relative clauses (RCs) in Marori, showing that this language unusually has almost all of relative clause types, from headed/headless, externally/internally headed, single-/double-headed, to pre-/post-head, to attached/detached RCs. Special attention is given to internally headed relative clauses (IHRC). It is argued that Marori IHRCs are of the restrictive or non-maximalising type, which accounts for certain intriguing properties, such as their indefiniteness constraints and the possibility for RC stacking.
In this paper we investigate the status of control constructions in Modern Standard Arabic (MSA). MSA has several embedded clause constructions, some of which resemble control in English (and other languages). However, these constructions exhibit some notable differences. Chief among them is the fact that the embedded verb carries agreement features that can indicate both coreference and disjoint reference between a matrix argument and the understood subject of the complement clause. We conducted a thorough corpus-based investigation of such constructions, with a special focus on a search for obligatory control in the language. We show that our findings contradict accepted generalizations (and predictions) proposed by state-of-the-art theories of control, as they indicate that there are no "real" control predicates in MSA. We outline an HPSG analysis that accounts for the MSA data.
The A-NOT-A structure is one way to express polar questions in Mandarin Chinese. The present study provides a constraint-based analysis of A-NOT-A questions in Mandarin Chinese within the framework of HPSG (Pollard & Sag, 1994) and MRS (Copestake et al., 2005). We propose two possible approaches to analysing the A-NOT-A structure — a morphological/lexical approach as well as a syntactic approach — and illustrate their implementation, as well as their respective strengths and weaknesses.
This paper provides an analysis of the Cantonese post-verbal particle "can1". We argue that can1 is a resultative particle encoding the meaning of 'a small degree'. It is only compatible with (i) verbs that entail a specific resulted state of the theme argument and (ii) verbs that encode a potential change of the theme argument (Beavers, 2011, 2013). Assuming that change of state verbs involve a property scale (Hay et al., 1999), we propose that "can1" makes the property scale bounded by providing an end-point. This endpoint, however, is not precise. It consists of a range of values on the lower end of the scale.
This paper addresses some Japanese constructions where the predicate heading a subordinate clause – specifically, a suspensive form of IU 'say', OMOU 'think' or SURU 'do' – appears to be elided. I will discuss that these elliptic constructions are subject to certain syntactic and interpretative constraints which do not apply to their non-elliptic counterparts, and develop an SBCG-analysis that aims to model these constraints without postulating a covert element in the place of the missing verb.
A singular countable noun in English normally needs a determiner and they should agree in number. However, there is a type of noun phrase, such as 'these sort of skills', which does not conform to this generalisation. As a singular countable common noun, the noun 'sort' requires a determiner, but there is an agreement mismat ch here: 'sort' is singular but the determiner is plural. Rather, the determiner agrees with the NP after the preposition 'of'. There are several po ssible analyses that might be proposed, but the best analysis is the one in which 'sort' and the preposition 'of' are 'functors', non-heads selecting heads.
This paper deals with the encoding of affectedness in Abui, a Papuan language of Indonesia. Abui is a head-marking language of the rare type where the verbs are marked for their undergoer arguments (So, O) formally split into several subtypes. This marking has been previously analyzed as a type of semantic alignment sensitive among others to affectedness. Affectedness is understood here as a scalar property delimiting the predicate (following Tenny 1987 and Beavers 2011). The paper explores the structure of the affectedness scale for Abui, comparing the functions and meaning of three types of person prefix paradigms. We show that verbs with similar meaning, encoding the same type of change (in Beavers’ terms) can differ in their entailments. We also show that there may be additional dimensions in which affectedness can be measured, such as affected agents, and that the interpretation of the degree on the affectedness scale interacts with instigator’s (source of force) status on the referential hierarchy. While human agents in some cases allow lower degrees of affectedness, the inanimate forces select the maximal degree reading. We conclude, that despite a considerable amount of fluidity of marking (Fedden et al. 2013, 2014), the shifts in degree of affectedness can be predicted as lowering of the degree stipulated for the predicate.
Function words like prepositions, adverbs, particles, and complementizers may be assigned more than one category due to the different functions they can have. In this paper I present an approach that assumes unique lexical entries for words that are assigned more than one category. I will focus on prepositions and how they may function as heads of modifying PPs, selected prepositions, or as particles.
Multi-word expressions (MWEs) are challenging for grammatical theories and grammar development since they blur the traditional distinction between the lexicon and the grammar, and vary in the degree of idiosyncrasy with respect to their semantic, syntactic, and morphological behavior. Nevertheless, the need to incorporate MWEs into grammars is unquestionable, especially in light of estimates claiming that MWEs account for approximately half of the entries in the lexicon. In this study we focus on verbal MWEs in Modern Hebrew: we consider different types of this class of MWEs, and propose an analysis in the framework of HPSG. Moreover, we incorporate this analysis into HeGram, a deep linguistic processing grammar of Modern Hebrew.
This paper describes some of our attempts in extending Zhong, a Chinese HPSG shared-grammar. New analyses for two Chinese specific phenomena, reduplication and the SUO-DE structure, are introduced. The analysis of reduplication uses lexical rules to capture both the syntactic and semantic properties (amplification in adjectives and diminishing in verbs). Words showing non-productive reduplication are entered in the lexicon, and the semantic relations will be captured in an external resource (the Chinese Open Wordnet). The SUO-DE structure constrains the meanings of relative clauses to a gapped-object interpretation.
Standard accounts of HPSG assume a distinction between morphology and syntax. However, despite decades of research, no cross-linguistically valid definition of 'word' has emerged (Haspelmath, 2010), suggesting that no sharp distinction is justified. Under such a view, the basic units are morphemes, rather than words, but it has been argued this raises problems when analysing phenomena such as zero inflection, syncretism, stem alternations, and extended exponence. We argue that with existing HPSG machinery, a morpheme-based approach can in fact deal with such issues. To illustrate this, we consider Slovene nominal declension and Georgian verb agreement, which have both been used to argue against constructive morpheme-based approaches. We overcome these concerns through use of a type hierarchy, and give a morpheme-based analysis which is simpler than the alternatives. Furthermore, we can recast notions from Word-and-Paradigm morphology, such as 'rule of referral' and 'stem space', in our framework. We conclude that using HPSG as a unified morphosyntactic theory is not only feasible, but also yields fruitful insights.
This paper describes an analysis for possessive idioms in English (e.g. 'I twiddle my thumbs' ''I am idle''). The analysis relies on matching at the semantic level, to allow for syntactic variation. It has been implemented in the English Resource Grammar, and tested by parsing a subset of the British National Corpus. In addition to the syntactic analysis, we have linked the idioms to entries in the Princeton Wordnet, to allow for further lexical semantic analysis.