Predicate-argument Structure Research Articles

СЛОВООБРАЗУВАТЕЛНАТА КАТЕГОРИЯ АГЕНС ПРИ СЪЩЕСТВИТЕЛНИТЕ, ОБРАЗУВАНИ С НАСТАВКА -ИК, В РАМКИТЕ НА ПРЕДИКАТНО-АРГУМЕНТНАТА ТЕОРИЯ

The paper presents an analysis of the derived nouns, formed by the suffix -ик, which are in the derivational category Agent. The theory of the predicate-argument structure is applied. It is assumed that the suffix -ик in these nouns expresses the argument Agent and the stem expresses a predicate or an argument from their semantic structure. The analysis is based on the way the theory of predicate-argument structure is applied in four researches: the two volumes of the Polish academic grammar from 1984, which are devoted to syntax and morphology, and the fifth and the ninth volume of the Bulgarian-Polish Contrastive Grammar, which are devoted to syntax and word-formation.

Assessing the Cross-linguistic Utility of Abstract Meaning Representation

Abstract Semantic representations capture the meaning of a text. Abstract Meaning Representation (AMR), a type of semantic representation, focuses on predicate-argument structure and abstracts away from surface form. Though AMR was developed initially for English, it has now been adapted to a multitude of languages in the form of non-English annotation schemas, cross-lingual text-to-AMR parsing, and AMR-to-(non-English) text generation. We advance prior work on cross-lingual AMR by thoroughly investigating the amount, types, and causes of differences that appear in AMRs of different languages. Further, we compare how AMR captures meaning in cross-lingual pairs versus strings, and show that AMR graphs are able to draw out fine-grained differences between parallel sentences. We explore three primary research questions: (1) What are the types and causes of differences in parallel AMRs? (2) How can we measure the amount of difference between AMR pairs in different languages? (3) Given that AMR structure is affected by language and exhibits cross-lingual differences, how do cross-lingual AMR pairs compare to string-based representations of cross-lingual sentence pairs? We find that the source language itself does have a measurable impact on AMR structure, and that translation divergences and annotator choices also lead to differences in cross-lingual AMR pairs. We explore the implications of this finding throughout our study, concluding that, although AMR is useful to capture meaning across languages, evaluations need to take into account source language influences if they are to paint an accurate picture of system output, and meaning generally.

More Than Syntaxes: Investigating Semantics to Zero-shot Cross-lingual Relation Extraction and Event Argument Role Labelling

Syntactic dependency structures are commonly utilized as language-agnostic features to solve the word order difference issues in zero-shot cross-lingual relation and event extraction tasks. However, while sentences in multiple forms can be employed to express the same meaning, the syntactic structure may vary considerably in specific scenarios. To fix this problem, we find semantics are rarely considered, which could provide a more consistent semantic analysis of sentences and be served as another bridge between different languages. Therefore, in this article, we introduce Syntax and Semantic Driven Network (SSDN) to equip syntax and semantic knowledge across languages simultaneously. Specifically, predicate–argument structures from semantic role labelling are explicitly incorporated into word representations. Then, a semantic-aware relational graph convolutional network and a transformer-based encoder are utilized to model both semantic dependency and syntactic dependency structures, respectively. Finally, a fusion module is introduced to integrate output representations adaptively. We conduct experiments on the widely used Automatic Content Extraction 2005 English, Chinese, and Arabic datasets. The evaluation results demonstrate that the proposed method achieves the state-of-the-art performance. Further study also indicates SSDN could produce robust representations that facilitate the transfer operations across languages.

On Emotional States and their Syntactic Representation

The study focuses on the analyses of syntactic structure of some predicates and state predicatives denoting positive emotion: veselya (to make someone happy, in a good moon), veselya se (to frolic, to party), vessel sam (to be happy), veselo (mi) e (to be in a good mood), veselba e (it is fun). Our aim is to provide a syntactic model representing argument structure of predicates and predicatives from this thematic group. The syntactic realization of arguments, semantic role hierarchy, and specifics concerning the referentiality of the nominal phrases in argument position are examined. The analyses show differences both in the semantic structure of predicates and predicatives and in the syntactic realization of their arguments. In contrast to other emotional predicates, veselya and veselya se have only nominal phrases as complements. While the analyzed verbs represent causative and dynamic predicates, respectively, the predicatives vessel, vesselo denote emotional states. Nominal phrases, as well as complement clauses can be used in complement positions. In this respect, the syntactic realization of predicatives is similar to the one of other emotional state predicatives, such as radosten sam (to be happy), radostno (mi) e (it is rejoicing).

Surprisal From Language Models Can Predict ERPs in Processing Predicate-Argument Structures Only if Enriched by an Agent Preference Principle.

Language models based on artificial neural networks increasingly capture key aspects of how humans process sentences. Most notably, model-based surprisals predict event-related potentials such as N400 amplitudes during parsing. Assuming that these models represent realistic estimates of human linguistic experience, their success in modeling language processing raises the possibility that the human processing system relies on no other principles than the general architecture of language models and on sufficient linguistic input. Here, we test this hypothesis on N400 effects observed during the processing of verb-final sentences in German, Basque, and Hindi. By stacking Bayesian generalised additive models, we show that, in each language, N400 amplitudes and topographies in the region of the verb are best predicted when model-based surprisals are complemented by an Agent Preference principle that transiently interprets initial role-ambiguous noun phrases as agents, leading to reanalysis when this interpretation fails. Our findings demonstrate the need for this principle independently of usage frequencies and structural differences between languages. The principle has an unequal force, however. Compared to surprisal, its effect is weakest in German, stronger in Hindi, and still stronger in Basque. This gradient is correlated with the extent to which grammars allow unmarked NPs to be patients, a structural feature that boosts reanalysis effects. We conclude that language models gain more neurobiological plausibility by incorporating an Agent Preference. Conversely, theories of human processing profit from incorporating surprisal estimates in addition to principles like the Agent Preference, which arguably have distinct evolutionary roots.

ДІЄСЛОВО В МОВНІЙ ТА КОНЦЕПТУАЛЬНІЙ КАРТИНАХ СВІТУ НОСІЇВ АНГЛІЙСЬКОЇ МОВИ

The paper studies the peculiarities of mental representations of the verb in the conceptual worldview and their reflection in the linguistic worldview of English language native speakers. Special attention is paid to the etymology and meaning of the word “verb” in the English language. In particular, its connection with the meaning “word” in general, which is manifested in the corresponding derivatives, is clarified. A tendency to define the verb through its syntactic function in scientific literature has been revealed. The key role of the predicate-argument structure, which reflects the linguistic worldview of English language native speakers, has been established. At the same time, the conceptual structure of the verb meaning and its mental representations shed light on the place of the verb in the conceptual worldview of English language native speakers. In the naive worldview of the English speakers, a tendency to associate the verb with an action or an act can be observed. High productivity of verbification in present-day English, as well as the overall high saturation of verbs in the English language corpora, estimated at 10.86 per cent, testify to the significance of the verb in English. The corpus analysis has been conducted based on the Corpus of Contemporary American English (COCA), the British National Corpus (BNC), the Strathy Corpus of Canadian English, and iWeb Corpus of modern web communication. The total distribution of verbs across the aforementioned corpora does not reveal any substantial variation. Nonetheless, the corpus data provide illuminating insights into the worldview of the native speakers of English. The prospects of further research in this direction lie in the construction of a mental model of the English verb, which would integrate the approaches to verb semantics within different modern frameworks, employing both experimental findings and empirical corpus data.

Contact Linguistics of Interlanguage

Different from traditional definitions of “interlanguage” and descriptions of interlanguage production, this study defines interlanguage in terms of its outstanding linguistic characteristics at the level of abstract lexical structure, including its three sublevels: lexical-conceptual structure, predicate-argument structure, and surface morphological realization patterns (Myers-Scotton and Jake 1995). Also different from most previous studies of interlanguage, this study views interlanguage or any interlanguage system as an outcome of a language-contact phenomenon. Adopting the Matrix Language Frame model (Myers-Scotton 1993[1997]) of bilingual intrasentential codeswitching in particular and contact linguistics in general, this study identifies the “matrix language” or the grammatical frame which structures interlanguage and reviews such a matrix language as a composite (Jake 1998; Wei 2009c). By adopting the Bilingual Lemma Activation model (Wei 2002, 2015, 2020), this study explores the nature and activity of the bilingual mental lexicon during interlanguage development. Accordingly, it proposes a particular approach to the nature and sources of second language learner errors or first language transfer in second language learning. It concludes that any interlanguage system must be driven by an incompletely acquired target language system in general and by an incompletely acquired target language abstract lexical structure in particular.

Relationship between Predicate-Argument Structure and Emotionality in Text

Relationship between Predicate-Argument Structure and Emotionality in Text

The Composite Abstract Lexical Structure of Interlanguage and Its Implications for Second Language Acquisition

This paper explores the nature of interlanguage with a focus on its lexical structure in relation to second language acquisition. The lexical structure of any language is assumed to be ‘abstract’ in that the mental lexicon contains ‘lemmas’, which are pieces of information about individual lexemes at three abstract levels: lexical-conceptual structure, predicate-argument structure, and morphological realization patterns. The abstract lexical structure of IL is assumed to be ‘composite’ in that during the process of second language acquisition several linguistic systems are in contact, each of which contributes different amounts to interlanguage (i.e., the developing linguistic system). This study claims that lemmas are language-specific and the bilingual mental lexicon contains cross-linguistic lemmas at each of these abstract levels. It further claims that bilingual lemmas are in contact in interlanguage production, and it is cross-lemma variations in the composite abstract lexical structure of interlanguage which induce learner errors. Naturally occurring interlanguage production data for the study include several first and second language pairs. Based on the research findings, this study concludes that interlanguage variations are driven by an incompletely acquired abstract lexical structure of a target language and offers some implications for second language acquisition.

Coordination and referential dependencies: a dependency grammar account in terms of predicate-valent structures

Abstract The present paper explores the nexus of coordination and referential dependencies, focusing on sentences such as ‘Max and Lucie talked about him’, in which it is hardly possible for Max and him to be co-valued. Previous accounts claim that such referential dependencies are in fact possible on a collective reading. It will be demonstrated here, however, that informants consistently judge the co-valued reading in such cases to be either strongly marginal or unavailable, regardless of whether a collective or distributive reading obtains. It is claimed that the manner in which referential dependencies interact with coordinate structures can be addressed insightfully in terms of ‘predicate-valent’ structures (as opposed to ‘predicate-argument structures’). A constraint in this regard will be proposed: the ‘Conjunct Referential Dependency Constraint’ (CRDC). The CRDC is couched in a Dependency Grammar (DG) approach to syntax.

High and Low Applicatives of Unaccusatives: Dependent Case and the Phase

The principal objective of this article is to establish a direct relationship between the structural height of the base position of the applied argument and the case and promotion-to-subject patterns observed in applicative constructions, with particular reference to applicatives of unaccusatives. The article achieves this through an approach exploiting dependent case, with the domains relevant for dependent case assignment being identified as phases, defined as (a) complete predicate-argument structures and (b) propositions. By making argument structure a defining ingredient of the delineation of phases, the article distills precise and accurate predictions about the interaction between the base-generation site of the applied object and the case patterns of unaccusative constructions featuring such an object, improving on the efficacy of previous accounts. In the process, the article reexamines the syntactic status of constituents located on the edge of a phase.

Bilingual Complex Abstract Lexical Structure and Its Relevance to Interlanguage Studies

This study adopts two assumptions about abstract lexical structure. One is that lexical structure is modular: lexical information is organized into subsystems pointing to different levels of linguistic structure, and parts of lexical structure can be split and recombined. The other concerns the sources of morphemes actually occurring in surface strings. As commonly recognized, abstract lexical structure contains three levels: lexical-conceptual structure, predicate-argument structure, and morphological realization patterns. This study argues that abstract lexical structure becomes “complex” because the bilingual mental lexicon contains lemmas (i.e., abstract entries about lexemes) from different sources, such as learners’ first language, their target language, and their interlanguage. By adopting a multi-layered speech production model (Levelt, 1989; Myers-Scotton & Jake, 2001; Wei, 2002, 2015, 2020), it further argues that different types of morphemes are accessed at different levels of language production, resulting in different degrees of learning difficulty (an implicational hierarchy of second language morpheme acquisition) and different types of learner errors in interlanguage production. The typical instances of learner errors for the study are collected from interlanguage performance by adult second language learners with various first language backgrounds. This study aims to explore the nature of the bilingual mental lexicon and mechanisms of interlanguage development.

A Canonical Context-Preserving Representation for Open IE: Extracting Semantically Typed Relational Tuples from Complex Sentences

Modern systems that deal with inference in texts need automatized methods to extract meaning representations (MRs) from texts at scale. Open Information Extraction (IE) is a prominent way of extracting all potential relations from a given text in a comprehensive manner. Previous work in this area has mainly focused on the extraction of isolated relational tuples. Ignoring the cohesive nature of texts where important contextual information is spread across clauses or sentences, state-of-the-art Open IE approaches are thus prone to generating a loose arrangement of tuples that lack the expressiveness needed to infer the true meaning of complex assertions.To overcome this limitation, we present a method that allows existing Open IE systems to enrich their output with additional meta information. By leveraging the semantic hierarchy of minimal propositions generated by the discourse-aware Text Simplification (TS) approach presented in Niklaus et al. (2019), we propose a mechanism to extract semantically typed relational tuples from complex source sentences. Based on this novel type of output, we introduce a lightweight semantic representation for Open IE in the form of normalized and context-preserving relational tuples. It extends the shallow semantic representation of state-of-the-art approaches in the form of predicate-argument structures by capturing intra-sentential rhetorical structures and hierarchical relationships between the relational tuples. In that way, the semantic context of the extracted tuples is preserved, resulting in more informative and coherent predicate-argument structures which are easier to interpret.In addition, in a comparative analysis, we show that the semantic hierarchy of minimal propositions benefits Open IE approaches in a second dimension: the canonical structure of the simplified sentences is easier to process and analyze, and thus facilitates the extraction of relational tuples, resulting in an improved precision (up to 32%) and recall (up to 30%) of the extracted relations on a large benchmark corpus.

Linguistic Codeswitching as a Cross-linguistic Lexical Bridge in Bilingual Communication

Most studies of linguistic codeswitching (CS) focus on what language items can be switched and how such switched items are intrasententially configurated at the surface sentence level. This study investigates linguistic CS at a rather abstract level by adopting the Bilingual Lemma Activation (BLA) Model (Wei, 2020). This model claims that lemmas (i.e., abstract entries in the mental lexicon about lexemes) are language-specific, language-specific lemmas are in contact in bilingual speech involving CS, and thus CS is cross-linguistically and lexical-conceptually driven in bilingual communication. In support of such a claim, this study provides evidence that bilinguals perform CS as a communicative strategy to make their intended meanings realized in terms of language-specific lemmas activated for the current exchange. Some typical instances of naturally occurring CS as observed in various language pairs involved in CS are described and explained at two levels of abstract lexical structure: lexical-conceptual structure and predicate-argument structure. This study offers some explanations of linguistic CS from a particular perspective and aims to explore the nature and activity of the bilingual mental lexicon during CS.

An electrophysiological exploration of modifiers embedded in different hierarchies: Insight from short passives in Mandarin

This event-related potentials study explored the processing of Mandarin short passives. We used a syntactic violation paradigm to compare the processing of two auxiliary phrases (i.e., a verb-modifier di-phrase and a noun-modifier de-phrase). In the control condition, the syntactic hierarchy of the di-phrase was lower than that of the de-phrase. In the violation condition, the low-level violation was created by replacing di with de in the auxiliary phrase, while the high-level violation was created by replacing de with di in the auxiliary phrase. The ERP data showed that the noun-modifier elicited the greater left anterior negativity (LAN) and the P600 than the verb-modifier, in both the control condition and the violation condition. We also observed that the LAN induced by the verb-modifier phrase was greater in the control condition than that in the violation condition, while the LAN induced by the noun-modifier was greater in the violation condition than that in the control condition. These results suggested that the greater cortical LAN-P600 might differentiate the high-level hierarchy from the low-level hierarchy. In addition, we tentatively claimed that given the same predicate argument structure, long passives might be the default representational mode of short passives (generally, a constructional alternation might be activated during the processing of the target structure).

The best game in town: The reemergence of the language-of-thought hypothesis across the cognitive sciences.

Mental representations remain the central posits of psychology after many decades of scrutiny. However, there is no consensus about the representational format(s) of biological cognition. This paper provides a survey of evidence from computational cognitive psychology, perceptual psychology, developmental psychology, comparative psychology, and social psychology, and concludes that one type of format that routinely crops up is the language-of-thought (LoT). We outline six core properties of LoTs: (i) discrete constituents; (ii) role-filler independence; (iii) predicate-argument structure; (iv) logical operators; (v) inferential promiscuity; and (vi) abstract content. These properties cluster together throughout cognitive science. Bayesian computational modeling, compositional features of object perception, complex infant and animal reasoning, and automatic, intuitive cognition in adults all implicate LoT-like structures. Instead of regarding LoT as a relic of the previous century, researchers in cognitive science and philosophy-of-mind must take seriously the explanatory breadth of LoT-based architectures. We grant that the mind may harbor many formats and architectures, including iconic and associative structures as well as deep-neural-network-like architectures. However, as computational/representational approaches to the mind continue to advance, classical compositional symbolic structures - that is, LoTs - only prove more flexible and well-supported over time.

Transition-based semantic role labeling with pointer networks

Semantic role labeling (SRL) focuses on recognizing the predicate–argument structure of a sentence and plays a critical role in many natural language processing tasks such as machine translation and question answering. Practically all available methods do not perform full SRL, since they rely on pre-identified predicates, and most of them follow a pipeline strategy, using specific models for undertaking one or several SRL subtasks. In addition, previous approaches have a strong dependence on syntactic information to achieve state-of-the-art performance, despite being syntactic trees equally hard to produce. These simplifications and requirements make the majority of SRL systems impractical for real-world applications. In this article, we propose the first transition-based SRL approach that is capable of completely processing an input sentence in a single left-to-right pass, with neither leveraging syntactic information nor resorting to additional modules. Thanks to our implementation based on Pointer Networks, full SRL can be accurately and efficiently done in O(n2), achieving the best performance to date on the majority of languages from the CoNLL-2009 shared task.

Towards a Marine Accident Frame

This paper presents the results of a study that combines information extraction by means of word sketches with corpus-based analysis. In the analysis, terminological information was extracted from a corpus of investigation reports by retrieving the word sketches of the selected verbs. The aim of the study is to demonstrate how to identify predicate-argument structures in a specific domain corpus by adopting the principles of Frame-based Terminology (FBT). Thus, this research shows how FBT can be applied to different scenarios such as a marine accident event, which is the object of this study. The long-term goal is to build a marine accident frame by utilising the predicate-argument structures identified in this study. The predicate-argument structures and the evolving frames can be used to enhance the representation and understanding of marine accident events. Thus, the predicate-argument structures are presented for consultation to future end users of the results, namely the stakeholders in the field of maritime safety, including maritime English teachers, as well as LSP teachers and terminologists.

Semantic Representation of Context for Description of Named Rivers in a Terminological Knowledge Base.

The description of named entities in terminological knowledge bases has never been addressed in any depth in terminology. Firm preconceptions, rooted in philosophy, about the only referential function of proper names have presumably led to disparage their inclusion in terminology resources, despite the relevance of named entities having been highlighted by prominent figures in the discipline of terminology. Scholars from different branches of linguistics depart from the conservative stance on proper names and have foregrounded the need for a novel approach, more linguistic than philosophical, to describing proper names. Therefore, this paper proposed a linguistic and terminological approach to the study of named entities when used in scientific discourse, with the purpose of representing them in EcoLexicon, an environmental knowledge base designed according to the premises of Frame-based Terminology. We focused more specifically on named rivers (or potamonyms) mentioned in a coastal engineering corpus. Inclusion of named entities in terminological knowledge bases requires analyzing the context that surrounds them in specialized texts because these contexts convey specialized knowledge about named entities. For the semantic representation of context, this paper thus analyzed the local syntactic and semantic contexts that surrounded potamonyms in coastal engineering texts and described the semantic annotation of the predicate-argument structure of sentences where a potamonym was mentioned. The semantic variables annotated were the following: (1) semantic category of the arguments; (2) semantic role of the arguments; (3) semantic relation between the arguments; and (4) lexical domain of the verbs. This method yielded valuable insight into the different semantic roles that named rivers played, the entities and processes that participated in the events educed by potamonyms through verbs, and how they all interacted. Furthermore, since arguments are specialized terms and verbs are relational constructs, the analysis of argument structure led to the construction of semantic networks that depicted specialized knowledge about named rivers. These conceptual networks were then used to craft the thematic description of potamonyms. Accordingly, the semantic network and the thematic description not only constituted the representation of a potamonym in EcoLexicon, but also allowed the geographic contextualization of specialized concepts in the terminological resource.

Ad astra or astray: Exploring linguistic knowledge of multilingual BERT through NLI task

AbstractRecent research has reported that standard fine-tuning approaches can be unstable due to being prone to various sources of randomness, including but not limited to weight initialization, training data order, and hardware. Such brittleness can lead to different evaluation results, prediction confidences, and generalization inconsistency of the same models independently fine-tuned under the same experimental setup. Our paper explores this problem in natural language inference, a common task in benchmarking practices, and extends the ongoing research to the multilingual setting. We propose six novel textual entailment and broad-coverage diagnostic datasets for French, German, and Swedish. Our key findings are that the mBERT model demonstrates fine-tuning instability for categories that involve lexical semantics, logic, and predicate-argument structure and struggles to learn monotonicity, negation, numeracy, and symmetry. We also observe that using extra training data only in English can enhance the generalization performance and fine-tuning stability, which we attribute to the cross-lingual transfer capabilities. However, the ratio of particular features in the additional training data might rather hurt the performance for model instances. We are publicly releasing the datasets, hoping to foster the diagnostic investigation of language models (LMs) in a cross-lingual scenario, particularly in terms of benchmarking, which might promote a more holistic understanding of multilingualism in LMs and cross-lingual knowledge transfer.