Lexical Functional Grammar (LFG)

In this paper I set out to model the encoding of definiteness in Hungarian possessive DPs with particular attention to the (non)complementarity of definiteness markers and possessive markers. I develop a formal LFG account of the relevant phenomena. Its essence is that I propose that in Hungarian DPs the DEF feature is nonunifiable. Naturally, the definite article always encodes this feature, while some possessor types also encode it, and others do not. The latter can only require that this feature should be present in the possessive DP. It is in this way that I capture the (non)co-occurrence of the definite article and various possessors.

Andrews and Manning (1993, 1999) proposed to analyse a number of constructions, including complex predicates and modal adjectives, in a heavily modified version of LFG that could accomodate the fact that these constructions seem to involve more in the way of hierarchical structure than seemed to be supported by the LFG architecture. Here I show that the same results can be achieved by combining the 'hybrid objects' of the analysis of coordinate structures proposed by Dalrymple and Kaplan 2000 with a more extensive use of some of the capacities of LFG's glue semantics.

Download/Open Review at: http://langsci-press.org/catalog/book/25

This book introduces formal grammar theories that play a role in current linguistics or contributed tools that are relevant for current linguistic theorizing (Phrase Structure Grammar, Transformational Grammar/Government & Binding, Mimimalism, Generalized Phrase Structure Grammar, Lexical Functional Grammar, Categorial Grammar, Head-Driven Phrase Structure Grammar, Construction Grammar, Tree Adjoining Grammar, Dependency Grammar). The key assumptions are explained and it is shown how each theory treats arguments and adjuncts, the active/passive alternation, local reorderings, verb placement, and fronting of constituents over long distances. The analyses are explained with German as the object language.

In a final part of the book the approaches are compared with respect to their predictions regarding language acquisition and psycholinguistic plausibility. The nativism hypothesis that claims that humans posses genetically determined innate language-specific knowledge is examined critically and alternative models of language acquisition are discussed. In addition this more general part addresses issues that are discussed controversially in current theory building such as the question whether flat or binary branching structures are more appropriate, the question whether constructions should be treated on the phrasal or the lexical level, and the question whether abstract, non-visible entities should play a role in syntactic analyses. It is shown that the analyses that are suggested in the various frameworks are often translatable into each other. The book closes with a section that shows how properties that are common to all languages or to certain language classes can be captured.

This book is a contribution to the general discussion of the question whether argument structure constructions should be treated with reference to phrasal patterns as suggested by Goldberg (1995, 2006); Culicover and Jackendoff (2005) and others or whether lexical approaches like) are more appropriate. In the absence of any fully worked out formalized phrasal proposals in Construction Grammar many arguments against phrasal approaches had hypothetical character (see for instance Müller, 2006). This paper discusses recent approaches by Asudeh, Dalrymple, and Toivonen (2008, 2013), Christie (2010) and Asudeh, Giorgolo, and Toivonen (2014) in the framework of LFG, which can be seen as formalizations of phrasal constructionist approaches. The authors argue that certain arguments in resultative and benefactive constructions in English are licensed in phrasal constructions rather than lexically. Applying an old argument by Dowty and Bresnan to resultative constructions, I show that data involving deriva-tional morphology suggests that valence information is visible at the lexical level and hence should not be introduced at the phrasal level. The conclusion is that analyses like the classical lexical analysis of resultative constructions by Simpson (1983) are the only option for lexicalist theories like LFG and HPSG. A second part of the paper discusses active passive alternations in template-based approaches and points out that generalizations regarding c-structure are missing 1 in the same way as they are missing in simple phrase structure grammar. Such missing generalizations motivated Harris and Chomsky to introduce transformations , but transformations are not used in LFG and hence the generalizations regarding active/passive c-structure pairs can not be captured. Furthermore, I show that cross-linguistic generalizations cannot be captured with reference to phrasal configurations since languages differ in the way they actually realize resultative and benefactive constructions. It is shown that languages like German that allow much freer constituent order than English and partial verbal phrases are incompatible with phrasal views on argument structure. In a third part I develop a lexical account of German and English resultatives and benefactives in the framework of HPSG and show how this account captures the commonalities between German and English despite the superficial dissimilarities between the two languages.

A crossed/backward control structure in Indonesian is analysed within LFG as forming a complex SVC-like tight structure that allows argument sharing. It shows properties having intermediate status between a complex predicate and an ordinary complementation structure. The analysis is implemented and tested in the computational grammar of Indonesian, developed with the ParGram (Parallel Grammar) framework.

Verb agreement with disjunctively conjoined nouns results in either a singular or a plural verb. This variation is associated to an analysis of disjunction either as inclusive or or as exclusive or in truth-conditional semantics. Apart from the semantic implications, there are also syntactic issues involved in determining the number of the verb with disjunctively conjoined nouns and these are specific strategies used by native speakers, which motivate the choice of a singular or a plural agreeing verb.

The present research aims to investigate the effect of stress on Vafsi oblique clitic PAMs and accordingly determine their placement based on LFG analysis by using the prosody-syntax interface.
Vafsi is one of the Iranian languages belonging to the Tati group of Iranian North-Western languages which is spoken only in four villages of Vafs, Chehreghan, Gurchan, and Fark in Markazi province.
This research is library-based, but, in cases of need for more evidence, linguistic productions of the native middle and old-aged male residents of Chehreghan and Vafs are used.The theory in this research is bifurcated. First, the theory of clitics is introduced, and based on Zwicky & Pullum’s (1983) criteria for discriminating clitics from affixes, 8 types of Vafsi clitics are classified. Vafsi distinguishes between direct and oblique case person-agreement markers (PAM) (Stilo. 2010). While the PAMs representing the direct case (set 1) are suffixal to the verb, the oblique PAMs (set 2) are usually enclitic. Based on the transformational generative grammar, the last type of Vafsi clitics, which is Vafsi oblique PAMs create challenges in determining clitichood, as they act as affixes in some placements but as enclitics in others. Hence, based on the transformational generative grammar, earlier studies (Stilo, 2004a. 2004b; 2010), have classified Vafsi oblique PAMs (set 2) into two categories of “affixal form” and “clitic form”.

In the framework of a Modern Greek LFG/XLE grammar development project at ILSP/ " Athena " RC, we implemented a novel multilevel analysis of tense in main and na subordinated clauses. Existing analyses of tense and the subjunctive mood in Modern Greek do not cover the entirety of tenses available in this language, do not provide a unified analysis of the tense system and the subjunctive mood and do not encode facts of sequence of tenses in subordinated clauses with verbs in the subjunctive mood. Our proposal draws on Reichenbach's ideas and provides a unified analysis of a wide range of tense and subjunctive data. We rely on corpus data retrieved from the HNC (http://hnc.ilsp.gr/).

This paper discusses the interaction and the mismatch between a syntactically ergative system and morphologically symmetrical markings in Kulon-Pazeh (Austronesian language, ISO: UUN, Taiwan, henceforth Pazeh), specifically referring to the effect of voice morphology on the surface realisation of arguments. Based on the evidence from Pazeh, I argue that there is no one-to-one correspondence of symmetricality or correlation between syntax and morphology. At the morphology level, Pazeh demonstrates symmetricality for the markings on the verbs in actor voice, undergoer voice and instrumental voice, but at the syntactic level, the surface realisation of arguments shows that the system is, in fact, ergative, and hence deeply asymmetrical. This study shows that the operation of the Pazeh voice system contains a mismatch of symmetricality between syntax and morphology, involving co-present parallel structures which can be best illustrated using the Lexical Functional Grammar (LFG) framework. The finding contributes to the empirical understanding of voice alternations and the alignment system of Austronesian languages of Taiwan.

This study provides description and analysis of some verbal complementation patterns in Egyptian Colloquial Arabic (ECA), namely the Auxiliary /kaan/, Causative /xalla/, phasal verbs and modals. Each verb is represented by a set of sentences extracted from a 5 million word corpus of ECA online texts that was built for the purpose of the current study using the Sketch Engine tool. These verbal complements are described and analysed within the principles of LFG syntactic theory, and represented in a grammar fragment implemented using the XLE tool.
The analysis shows that both tense and aspect can be expressed on verbs in ECA, where in simple tense forms the verb carries tense only, while in compound tense, the main predicate marks tense and occupies I while the following lexical verb marks grammatical aspect and occupies V.
The bi- prefix marks present tense on verbs in I and imperfect aspect on verbs in V, as well as a HAB/PROG feature. The bare Imperfective verb form is treated as a non-finite verb in ECA, where it can not occupy I and is marked by VFORM=BARE. All of the verbal constructions analysed are bi-clausal structures, however, they show differences regarding the kind of control relation. Functional control was attested in constructions where the main predicate is the auxiliary /kaan/, the causative verb /xalla/, phasal verbs, as well as non-inflecting modals. Anaphoric
control was attested only with inflecting modals, with the modal /yi2dar/ ‘able’ showing a case of obligatory anaphoric control.
This is, to my knowledge, the first study which attempts to develop a grammar for ECA using the XLE platform. It provides an insight over the issues correlated with developing this grammar, which could be a step towards including ECA into the ParGram project in order to develop broad coverage grammars for a bigger number of languages.

In this paper we study obligatory object control (OOC) in Greek, in the sequence main + na-subordinate clause. Our aim is to provide a list of verbs categorized by the type of OOC they allow-exhaustive object control or partial control. To this purpose we annotated 14,000 sentences drawn from the Hellenic National Corpus in the BRAT Rapid Annotation Tool. These sentences contained verbs that in general are considered to take part in control constructions. Next we applied diagnostic tests on these verbs in order to decide which type of control they allow. We present a unified treatment on OOC phenomena in Greek and we model the results of this study in the LFG/XLE grammar that is being developed for Greek. 1. Εισαγωγή Σε αυτό το άρθρο παρουσιάζουμε τη μελέτη του υποχρεωτικού ελέγχου του αντικειμένου στα ελληνικά, βασισμένη σε σώμα κειμένων με ειδικό πεδίο μελέ-της την ακολουθία κύρια πρόταση + εξαρτημένη να-πρόταση. Το φαινόμενο του ελέγχου στα ελληνικά έχει συζητηθεί εκτενώς βιβλιογραφικά (). Ωστό-σο, στις υπάρχουσες μελέτες συζητούνται τα ρήματα που υποστηρίζουν τη δομή αυτή, χωρίς να δίνεται ιδιαίτερη έμφαση στο φαινόμενο του ελέγχου του αντι-κειμένου και χωρίς να μελετάται μεγάλο εύρος πραγματικών δεδομένων. Για τον λόγο αυτό, επιλέξαμε να δημιουργήσουμε ένα ειδικά επισημειωμένο σώμα κειμένων για το φαινόμενο της μελέτης μας, ώστε να μελετήσουμε το φαινόμε-νο σε ένα μεγάλο εύρος δεδομένων, τα οποία μπορούν να αξιοποιηθούν και σε Το παρόν άρθρο εντάσσεται στο πλαίσιο της ανάπτυξης μιας LFG/XLE γραμματικής για τα ελληνικά. Ευχαριστούμε τη Μαρίκα Λεκάκου για τις συζητήσεις που είχε μαζί μας, τους ανώνυμους κριτές για τις πολύτιμες παρατηρήσεις, καθώς και το κοινό της 39ης Συνάντησης Εργασίας του Το-μέα Γλωσσολογίας του ΑΠΘ. Για την K. Τζώρτζη: Το έργο συγχρηματοδοτείται από την Ελλάδα και την Ευρωπαϊκή Ένωση (Ευρωπαϊκό Κοινωνικό Ταμείο) μέσω του Επιχειρησιακού Προγράμματος "Ανάπτυξη Ανθρώπινου Δυναμικού, Εκπαίδευση και Διά Βίου Μάθηση", στο πλαίσιο της Πράξης "Ενίσχυση του ανθρώπι-νου ερευνητικού δυναμικού μέσω της υλοποίησης διδακτορικής έρευνας" (MIS-5000432), που υλο-ποιεί το Ίδρυμα Κρατικών Υποτροφιών (ΙΚΥ). Για την A. Φιωτάκη: Αυτή η έρευνα χρηματοδοτείται από τη Γενική Γραμματεία Έρευνας και Τεχνολογίας (ΓΓΕΤ) και από το Ελληνικό Ίδρυμα Έρευνας και Καινοτομίας (ΕΛΙΔΕΚ) (Κωδικός υπο-τροφίας: 1721).