Syntactic Representation Research Articles

DP-Parse: Finding Word Boundaries from Raw Speech with an Instance Lexicon

AbstractFinding word boundaries in continuous speech is challenging as there is little or no equivalent of a ‘space’ delimiter between words. Popular Bayesian non-parametric models for text segmentation (Goldwater et al., 2006, 2009) use a Dirichlet process to jointly segment sentences and build a lexicon of word types. We introduce DP-Parse, which uses similar principles but only relies on an instance lexicon of word tokens, avoiding the clustering errors that arise with a lexicon of word types. On the Zero Resource Speech Benchmark 2017, our model sets a new speech segmentation state-of-the-art in 5 languages. The algorithm monotonically improves with better input representations, achieving yet higher scores when fed with weakly supervised inputs. Despite lacking a type lexicon, DP-Parse can be pipelined to a language model and learn semantic and syntactic representations as assessed by a new spoken word embedding benchmark. 1

Syntactic Theory Perception on Language Acquisition

This paper focuses on identifying the syntactic theory perception on acquiring first language. The paper highlights the role and function of language syntactic structure in childhood on acquiring first language from birth and using it in communicative situations in reality later. The paper, however, emphasizes the impact of grammatical development, focusing mainly on nativist and usage-based, on acquiring first language. The paper bases its hypotheses and assumptions on Chomsky’s universal grammar and transformation grammar theories. The study, moreover, exposes language as a symbolic system and how it is formed in mind from birth, focusing mainly on the difference between human language and animal sounds. A brief idea about linguistic and communicative competence is presented to associate them to first language acquisition. the study concludes that grammar, which is part of the syntactic theory, is concerned with discovering the principles and sets of rules used for governing the structures of utterances and sentences. These principles and sets of rules constitute what is called linguistic competence of language. This linguistic competence is what helps children acquire their first language.

Mapping out‐ argument structure

AbstractThis article explores a phenomenon of English in which out‐ combines with a predicate to form a complex predicate (e.g., outsing, outdo, outrun, outsmart, …), here called “out‐pred.” A thorough investigation uncovers several new generalizations, which can be summarized as follows. (i) Out‐pred formation is productive and syntactic, building upon the structure for the pred. (ii) Out‐ is the core of the out‐pred clause's extended verbal projection. These findings are derived via an analysis in which out‐ merges with the pred before any argument(s) can merge. This analysis is then further supported by exploring domains in which out‐pred is unavailable; though these are seemingly unrelated, they share deep derivational properties that are incompatible with the derivation of out‐pred. The findings of this article have implications for the syntactic representation of argument structure more generally, supporting analyses where all arguments of a verb are syntactically severed from it.

Raising out of Finite Clauses (Hyperraising)

Classical syntactic theory was designed to permit the type of movement called raising to proceed out of infinitival, but not finite, clauses—a positive result for languages such as English. But crosslinguistic investigation reveals that many languages actually allow raising out of finite clauses (hyperraising), challenging certain commonly assumed locality constraints on movement. This article reviews three types of Minimalist analyses of hyperraising and how they address these challenges, noting the strengths and shortcomings of each. Defectiveness/nonphase analyses commendably tie a clause's ability to launch hyperraising to independent observables, but such analyses struggle to derive the former from the latter. Deactivation analyses boast empirical successes but do not straightforwardly rule out hyperraising in English. Phase-edge analyses also boast empirical successes but face empirical and/or conceptual problems (to which a solution is sketched out) and open questions about learnability. These evaluations are intended to spur syntacticians to develop stronger versions of all three types of analyses, bringing us closer to fully understanding the factors regulating movement, a subcase of the fundamental structure-building operation Merge.

The Syntactic Argumentation of the Word / le: š / in Jordanian Arabic

This study aims at investigating the syntactic status of the word /le:š / in question formation in Jordanian Arabic (JA). It is argued that this word has two syntactic functions in JA questions. The first of which is that it works as interrogative operator, meaning ‘why’, which moved to the [Spec C] position in one type of wh-question. However, the second role of this word is found in other types of questions, wh-questions and yes-no questions. In these types of questions /le:š/ is realized as an interrogative complementiser formed within the C position and giving the interrogative force for the question. To this end, empirical evidence is given and discussed to distinguish between the two syntactic representations of this word especially in ambiguous cases of having two different types of questions sharing the same surface structure.

Filler–gap dependencies and the remnant–correlate dependency in backward sprouting: Sensitivity to distance and islands

The relationship between the wh-remnant and the null correlate in the type of ellipsis known as backward sprouting is superficially almost identical to the relation between a wh-filler and a gap in a wh-question. In both cases, there is a dependency between the wh-phrase and a later null element. We conduct a sentence acceptability experiment to test whether the remnant–correlate dependency in backward sprouting exhibits two well-known properties of a filler–gap dependency in wh-questions: sensitivity to clause boundaries (distance) and sensitivity to islands. The results show that both dependency types are sensitive to clause boundaries, although the effect is larger in the case of filler–gap dependencies, but that only filler–gap dependencies are sensitive to islands. These results present a challenge to analyses of sprouting in which the ellipsis site contains a full representation of the structure of the antecedent clause, since such analyses predict island-sensitivity for remnant–correlate dependencies. The results also suggest that island-sensitivity cannot be reduced to simple processing demands without regard to the syntactic representation of the dependency, since such a view would predict greater similarity between filler–gap dependencies and remnant–correlate dependencies than is found.

The whole picture: Disentangling locality, logophoricity and subjecthood in English picture noun anaphora

This article provides a solution to the long-standing puzzle of English anaphors within so-called picture noun phrases, which superficially exhibit an exceptional binding behavior. In particular, picture noun anaphors seem, under certain conditions, to escape the locality conditions imposed by Condition A of Binding Theory. Previous proposals attribute such apparently exceptional behavior to various sources: the classical Binding Theory appeals to the possible presence of covert agents within NPs; predicate-based theories introduce the possibility of exemption from Condition A; others capitalize on possible homophony with (logophoric) pronouns. While all of these proposals provide valuable insight into some aspect of the puzzle, we show that they all fail to capture the full empirical picture. Based on a detailed examination of their behavior in various syntactic and interpretive conditions, we instead propose that English picture noun anaphors, like any other anaphor, systematically obey Condition A. Their apparent exemption from it in some cases derives from the possible implicitness of some binders, in particular, logophoric pronouns or nominal subjects. Furthermore, the availability of such covert binders is crucially affected by a binding-independent competition principle between weaker and stronger forms. Thus, the apparently irregular behavior of English picture noun anaphors results from the interaction between several factors (syntactic representation of logophoricity, syntactic projection of subjects in nouns, pronominal competition), which is responsible for the illusion that Condition A does not apply systematically. By disentangling these factors, we propose a solution that integrates previous insights without compromising on empirical adequacy or analytical parsimony.

Subjects are not all alike: Eye-tracking the agent preference in Spanish.

Experimental research on argument structure has reported mixed results regarding the processing of unaccusative and unergative predicates. Using eye tracking in the visual world paradigm, this study seeks to fill a gap in the literature by presenting new evidence of the processing distinction between agent and theme subjects. We considered two hypotheses. First, the Unaccusative Hypothesis states that unaccusative (theme) subjects involve a more complex syntactic representation than unergative (agent) subjects. It predicts a delayed reactivation of unaccusative subjects compared to unergatives after the presentation of the verb. Second, the Agent First Hypothesis states that the first ambiguous NP of a sentence will preferably be interpreted as an agent due to an attentional preference to agents over themes. It predicts a larger reactivation of agent subjects than themes. We monitored the time course of gaze fixations of 44 native speakers across a visual display while processing sentences with unaccusative, unergative and transitive verbs. One of the pictures in the visual display was semantically related to the sentential subject. We analyzed fixation patterns in three different time frames: the verb frame, the post-verb frame, and the global post-verbal frame. Results indicated that sentential subjects across the three conditions were significantly activated when participants heard the verb; this is compatible with observing a post-verbal reactivation effect. Time course and magnitude of the gaze-fixation patterns are fully compatible with the predictions made by the Agent First Hypothesis. Thus, we report new evidence for (a) a processing distinction between unaccusative and unergative predicates in sentence comprehension, and (b) an attentional preference towards agents over themes, reflected by a larger reactivation effect in agent subjects.

Person and honorification: Features and interactions in Magahi

The Magahi language is rich in honorification morphology: it has three versions of ‘you’ (nonhonorific, honorific, and high honorific) and makes honorific distinctions for third person nominals as well as second person nominals. This work uses these empirical riches to develop a formal syntactic theory of honorific marking. It argues that Magahi’s honorific distinctions can be decomposed into two binary features, [+/–HON] and [+/–HIGH]. These features are specified on an Hon head that is part of the extended projection of any nominal phrase in Magahi; from there they can be copied onto verbal functional heads by Agree. It goes on to argue that Hon heads in Magahi bear a first person designated index that specifies the individual with respect to which the social standing of the person referred to by the nominal inside HonP is evaluated. This first person index is used to account for the intricate ways in which honorific marking interacts with indexical shift in Magahi.

Post-Authorship Attribution Using Regularized Deep Neural Network

Post-authorship attribution is a scientific process of using stylometric features to identify the genuine writer of an online text snippet such as an email, blog, forum post, or chat log. It has useful applications in manifold domains, for instance, in a verification process to proactively detect misogynistic, misandrist, xenophobic, and abusive posts on the internet or social networks. The process assumes that texts can be characterized by sequences of words that agglutinate the functional and content lyrics of a writer. However, defining an appropriate characterization of text to capture the unique writing style of an author is a complex endeavor in the discipline of computational linguistics. Moreover, posts are typically short texts with obfuscating vocabularies that might impact the accuracy of authorship attribution. The vocabularies include idioms, onomatopoeias, homophones, phonemes, synonyms, acronyms, anaphora, and polysemy. The method of the regularized deep neural network (RDNN) is introduced in this paper to circumvent the intrinsic challenges of post-authorship attribution. It is based on a convolutional neural network, bidirectional long short-term memory encoder, and distributed highway network. The neural network was used to extract lexical stylometric features that are fed into the bidirectional encoder to extract a syntactic feature-vector representation. The feature vector was then supplied as input to the distributed high networks for regularization to minimize the network-generalization error. The regularized feature vector was ultimately passed to the bidirectional decoder to learn the writing style of an author. The feature-classification layer consists of a fully connected network and a SoftMax function to make the prediction. The RDNN method was tested against thirteen state-of-the-art methods using four benchmark experimental datasets to validate its performance. Experimental results have demonstrated the effectiveness of the method when compared to the existing state-of-the-art methods on three datasets while producing comparable results on one dataset.

عمل الجملة الاسمية بأنّ المصدرية في رواية أبواب المدينة لإلياس خوري: تحليل نحوي

This study aimed to reveal the functions of the jumlah ismiyyah which consist anna al-maṣdariyyah in the novel Abwābu al-Madīnati by Ilyās Khūrī by utilizing syntactic theory related to functions. The data in this study were in the form of jumlah ismiyyah consisting of anna al-maṣdariyyah. Technique of collecting data in this study used the referral method with tapping technique as a basic technique and note-taking technique as an advanced technique. This study used the agih method as the analysis stage with the Direct Element Technique (BUL) as a basic technique with i‘rāb model and substitution technique as an advanced technique. Furthermore, the presentation of the results of data analysis was done by using informal methods.
 Based on the research that has been done, it can be concluded that in the Abwābu al-Madīnati novel by Ilyās Khūrī, there were 101 jumlah ismiyyah consisting of anna al-maṣdariyyah. Those 101 jumlah ismiyyah were classified based on their functions, such as khabar mubtada`, fā‘il, maf‘ūl bih, mustaṡnā, ma‘ṭūf, and ism majrūr. Jumlah ismiyyah consisting of anna al-maṣdariyyah with the form maṣdar mu`awwal can be replaced with ism in the form of maṣdar ṣarīḥ. The substitution of maṣdar mu`awwal with the form maṣdar ṣarīḥ did not cause a change in meaning and grammatical function because both of them get the same position in a sentence. Then the use of maṣdar mu`awwal in that novel gives a specific time marker for an action in a sentence and makes the meaning of the sentence become more cultured.

Variation at the Syntax–Pragmatics Interface: Discourse Particles in Questions

This paper focuses on the microvariation concerning the distribution and functions of certain interrogative discourse particles found in several central and southern Italian dialects. These particles show many similarities in terms of both their morphological shapes (being homophonous to the wh-phrase corresponding to English ‘what’) and their syntactic distribution within the sentence, in that they all occur at the beginning of polar questions. However, a careful analysis of their distribution across a pragmatically defined typology of canonical and non-canonical polar questions shows that these particles are not possible in all question types. In particular, two patterns emerge: in Pattern B, the particle is associated with a conventional implicature that both the speaker and the addressee are competent with respect to the issue addressed by the polar question, while in Pattern A, this implicature is restricted to the addressee competence. This point of microvariation is then analysed by assuming that both the pragmatic assumptions of competence and the discourse participants are encoded in the syntactic representation; the cartographic framework is adopted to characterize the compositional structure.

Syntactic chunking reveals a core syntactic representation of multi-digit numbers, which is generative and automatic

Representing the base-10 structure of numbers is a challenging cognitive ability, unique to humans, but it is yet unknown how precisely this is done. Here, we examined whether and how literate adults represent a number’s full syntactic structure. In 5 experiments, participants repeated number-word sequences and we systematically varied the order of words within each sequence. Repetition on grammatical sequences (e.g., two hundred ninety-seven) was better than on non-grammatical ones (hundred seven two ninety). We conclude that the participants represented the number’s full syntactic structure and used it to merge number words into chunks in short-term memory. Accuracy monotonously improved for sequences with increasingly longer grammatical segments, up to a limit of ~ 4 words per segment, irrespectively of the number of digits, and worsened thereafter. Namely, short chunks improved memorization, whereas oversized chunks disrupted memorization. This chunk size limit suggests that the chunks are not based on predefined structures, whose size limit is not expected to be so low, but are created ad hoc by a generative process, such as the hierarchical syntactic representation hypothesized in Michael McCloskey’s number-processing model. Chunking occurred even when it disrupted performance, as in the oversized chunks, and even when external cues for chunking were controlled for or were removed. We conclude that the above generative process operates automatically rather than voluntarily. To date, this is the most detailed account of the core representation of the syntactic structure of numbers—a critical aspect of numerical literacy and of the ability to read and write numbers.

Machine-Readable Entailments with the Italian Prendere Construction Expressing Hitting and Insulting Events

The Italian language features a little debated transitive construction with prendere ‘to take/to catch’ in which a prepositional phrase (PP) with an adverbial value occurs mandatorily (e.g. Lui prese a pugni Leo ‘He punched Leo’). Semantically, this construction often implies the use of physical force or verbal offence. In the hitting or insulting event, the notional subject generally is a [+ Human] Agent, whilst the notional direct object generally is a [+ Animate] Affectee ([1]: 4). It can be contended that prendere, which carries no literal meaning, is zero-valent and that the predicate assigning semantic roles is the PP. A computational tool will be illustrated, which automatically performs the following NLP / NLU tasks: it provides a reliable syntactic and semantic representation of the clause type, and it produces machinereadable entailments with three clause types, i.e. sentences with ordinary verbs, support verbs, and the causative verb fare ‘to make/to have’.

Emotional Attractors in Subject-Verb Number Agreement.

Considering the crosstalk between brain networks that contain linguistic and emotional information and that no studies have examined the impact of semantic information of affective nature on subject-verb number agreement, the present Event Related Potential (ERP) study investigated the extent to which emotional local nouns whose number mismatched that of subject head nouns might be considered by the parser during comprehension of grammatically correct sentences. To this end, twenty-eight Spanish native speakers were tested on a self-paced reading task while their brain activity was recorded. The experimental materials consisted of 120 sentences where the valence (negative vs. neutral) and number (singular vs. plural) of the local noun of the singular subject noun-phrase (NP) were manipulated; El gorro de aquel/aquellos cazador(es)/mecánico(s) era… [The hat of that/those hunter(s)/mechanic(s) was…]. ERP results measured in the local noun position showed that valence and number interacted in the 300–500 ms (negative component) and 780–880 ms (late positivity) time windows. In the (target) verb position, the two factors only interacted in the late 780–880 ms time window, revealing an “ungrammatical illusion” for plural marked neutral words. Our findings suggest that number agreement is sensitive to affective meaning but that the emotional information of an attractor is considered in different operations and at different stages during grammatical sentence processing; it can affect lexical and syntactic representation retrieval of a subject-NP and impact agreement encoding only at late stages of processing, during verb agreement and feature integration.

Some structural disanalogies between pronouns and tenses

AbstractA pronominal analysis of tense goes back to Partee (1973), motivated by a series of proposed parallels between the interpretation of tenses and that of pronouns. This article revisits Partee's interpretive parallels, as well as two more identified in Kratzer (1998), in light of subsequent developments in work on both temporal relations and on pronouns. The goal of this article is not to argue for or against a pronominal analysis of tense, but instead to make clearer the syntactic and semantic space within which such an analysis is situated, especially given that pronouns have been given increasingly complex syntactic representations even as tense has remained syntactically simplex.

Probing Word Syntactic Representations in the Brain by a Feature Elimination Method

Neuroimaging studies have identified multiple brain regions that are associated with semantic and syntactic processing when comprehending language. However, existing methods cannot explore the neural correlates of fine-grained word syntactic features, such as part-of-speech and dependency relations. This paper proposes an alternative framework to study how different word syntactic features are represented in the brain. To separate each syntactic feature, we propose a feature elimination method, called Mean Vector Null space Projection (MVNP). This method can remove a specific feature from word representations, resulting in one-feature-removed representations. Then we respectively associate one-feature-removed and the original word vectors with brain imaging data to explore how the brain represents the removed feature. This paper for the first time studies the cortical representations of multiple fine-grained syntactic features simultaneously and suggests some possible contributions of several brain regions to the complex division of syntactic processing. These findings indicate that the brain foundations of syntactic information processing might be broader than those suggested by classical studies.

Analysis of one lexicographic case: the Russian noun svist ‘whistle’

The subject of the article is the application of the case study method to the description of lexical semantics, in particular, to the analysis of a dictionary entry. The author considers the Russian noun SVIST ‘whistle’ as an object, namely, its description in the “Dictionary of the Russian language” (1984). Semantic analysis shows that the description of the word polysemy in the dictionary does not correspond to the nature of the semantic representations that are assigned to this lexeme. The author offers his own version of the semantic description of this noun, in particular, using the category of image-schematic structure. Special attention is devoted to the problem of the dictionary presentation of syntactic representations, i.e. combinatory of the word.

Musical Garden Paths: Evidence for Syntactic Revision Beyond the Linguistic Domain.

While theoretical and empirical insights suggest that the capacity to represent and process complex syntax is crucial in language as well as other domains, it is still unclear whether specific parsing mechanisms are also shared across domains. Focusing on the musical domain, we developed a novel behavioral paradigm to investigate whether a phenomenon of syntactic revision occurs in the processing of tonal melodies under analogous conditions as in language. We present the first proof‐of‐existence for syntactic revision in a set of tonally ambiguous melodies, supporting the relevance of syntactic representations and parsing with language‐like characteristics in a nonlinguistic domain. Furthermore, we find no evidence for a modulatory effect of musical training, suggesting that a general cognitive capacity, rather than explicit knowledge and strategies, may underlie the observed phenomenon in music.

The interface between grammar and bodily enactment in ASL and English

Abstract Users of signed and spoken languages regularly engage bodily enactment (commonly referred to as constructed action [CA] for signers and character viewpoint gestures [CVPT] for speakers) for the creation of meaning, but comparatively few studies have addressed how linguistic grammar interfaces with such gestural depictive devices across language modalities. CVPT gestures have been shown to co-occur with spoken language transitive verbs, and when a reference is definite or more accessible in the discourse. In sign, CA often alternates sequentially with fully conventionalized signs. In both CVPT and CA demonstrations, syntactic and pragmatic factors appear to be important. In this work, we consider these patterns by examining short retellings of video-based elicitation stimuli (silent-movie segments) from 10 deaf users of ASL (American Sign Language) and 20 hearing speakers of English. We describe examples of signs and words that co-occur with or precede specific instances of CA and CVPT. We also examine distributions and degrees of enactment across participants in order to consider the question of gesture threshold (Hostetter and Alibali, 2008, 2019). We provide various examples of how gestural material interfaces with linguistic grammar, which has implications for syntactic theory and possible grammatical constraints on such communicative devices.