Search

• People with autoimmune neurological diseases can have language difficulties. • Language difficulties can be in production and comprehension. • Language difficulties can relate to specific language tasks. • Language difficulties may co-occur with specific brain damage. • Only a few studies administered standardized tests. Autoimmune neurological diseases (ANDs) are a specific type of autoimmune disease that affect cells within the central and peripheral nervous system. ANDs trigger various physical/neuropsychiatric symptoms. However, language impairments in people with ANDs are not well characterized. Here we aimed to determine the kinds of language impairment that most commonly emerge in 10 ANDs, the characteristics of the patients (demographic, neurological damage), and the assessment methods used. We followed the PRISMA Extension for Scoping Reviews (PRISMA-ScR). PubMed and Google Scholar were searched. We used a list of search terms containing 10 types of ANDs (e.g., multiple sclerosis, acute disseminated encephalomyelitis) in combination with the terms aphasia, dysphasia, fluency, language, listening, morphology, phonology, pragmatics, reading, semantics, speaking, syntax, writing. The reference lists and citations of the relevant papers were also investigated. The type of AND, patient characteristics, neurological damage and examination technique, language tests administered, and main findings were noted for each study meeting the inclusion criteria. We found 171 studies meeting our inclusion criteria. These comprised group studies and case studies. Language impairments differed largely among types of ANDs. Neurological findings were mentioned in most of the papers, but specific language tests were rarely used. Language symptoms in people with ANDs are commonly reported. These are often not full descriptions or only focus on specific time points in the course of the disease. Future research needs to assess specific language functions in people with ANDs and relate their language impairments to brain damage at different stages of disease evolution. [ABSTRACT FROM AUTHOR]

Unsupervised dependency parsing is the task of learning a dependency grammar from a set of unannotated sentences. This paper describes a neural network (NN) based probabilistic model, Neural Dependency Model with Valence (NDMV), that combines the dependency parsing with the rich nonlinear featurization of NN approaches. NDMV follows the traditional DMV model and computes the probability of a grammar rule via a feed-forward NN. A strength of this proposed approach is the ability to learn the underlying features of input tokens. In our experiments, this capability leads to gains in both NDMV and its extension version, Lexicalized Neural Dependency Model with Valence (L-NDMV): NDMV achieves better performance on WSJ and datasets of eight additional languages in comparison with previous approaches in the basic setting; L-NDMV achieves a result that is competitive with the current state-of-the-art. 1 1 Part of this paper been presented in a conference: Jiang et al. 2016 [1] and Han et al. 2017 [2]. [ABSTRACT FROM AUTHOR]

Abstract This paper introduces topology finding of patterns for shell structures such as beam grids for gridshells or voussoir tessellations for vaults, among others. The authors refer to topology finding, by analogy and in complement to form finding, as the design of the connectivity of these patterns in order to follow architectural, structural and construction requirements. This paper presents a computational approach relying on a specific design space and data structures based on singularity meshes, which encode the information about the singularities in patterns. The designed patterns are structured, i.e. with a low number of singularities, can include high-valency pole points, and respect alignment to surfaces, curves and points. A feature-based exploration approach is introduced with a generation procedure for singularity meshes following the boundaries of a surface as well as point and curve features, using a topological skeleton or medial axis. These features can stem from statics heuristics, whose efficiency is assessed in a case study. A rule-based editing approach for singularity meshes supplements feature-based topology finding, using a grammar of strip rules as parameters to further explore the singularity design space. This conceptual design approach and its algorithms are an aid for topological exploration of patterns for shell-like structures by architects and engineers. Highlights • Topology finding, by analogy and in complement to form finding, is introduced as a conceptual design approach, which is an aid for topological exploration of patterns for shell-like structures by architects and engineers; • Singularities in patterns are handled through coarse quad meshes, which can include high-valency pole points via pseudo-quad faces; • A feature- based exploration approach is introduced with a generation procedure for singularity meshes following the boundaries of a surface as well as point and curve features, using a topological skeleton or medial axis; • These features can stem from statics heuristics, whose efficiency is assessed in a case study; • A rule-based editing approach for singularity meshes supplements feature-based topology finding, using a grammar of strip rules as parameters to further explore the singularity design space. [ABSTRACT FROM AUTHOR]

Abstract This paper describes an OWL ontology that is a Universal Moral Grammar (UMG). UMG has been hypothesized by students of Chomsky to play the same role in human ethics as Universal Grammar (UG) does in Linguistics. I.e., the UMG describes a moral faculty hypothesized to have a genetic basis. This approach utilizes the modular view of the mind developed by Chomsky and currently utilized by many evolutionary psychology researchers. In this paper I describe the ontology and how it represents ethical choices, rules, scenarios, and systems. This includes representation of choices governed by the golden rule, utilitarianism, and Moral Foundations Theory. The foundation for the model is an Artificial Intelligence model of events, plans, and decisions. This plan model represents what is known as Theory of Mind (TOM). The TOM model is extended to the ethical domain by an analysis of recent research in evolutionary psychology and the representation of 40 different scenarios such as the trolley problem from the philosophical, psychological, and biological literature. The ontology is an example of how semantic technology can be used to provide mathematical rigor to the study of human ethics. This version demonstrates the breadth of the UMG, that it is capable of representing many diverse examples from the literature. However, there are already meaningful results from this version such as a resolution to Hume's Is-Ought problem. [ABSTRACT FROM AUTHOR]

The Tanzanian Bantu language Rangi exhibits a comparatively and typologically unusual word order alternation in the future tense. Whilst declarative main clauses exhibit post-verbal auxiliary placement, the auxiliary appears pre-verbally in wh-questions, sentential negation, relative clauses, cleft constructions and subordinate clauses. This paper examines this alternation from the perspective of Dynamic Syntax ( Cann et al., 2005; Kempson et al., 2001 ). Dynamic Syntax (DS) is a parsing-oriented framework which aims to capture the way in which meaning is established incrementally as a result of lexical input encountered in context. The paper presents a unified analysis of this construction found in Rangi, locating it within the wider workings of the language. It shows that this seemingly idiosyncratic constituent order is in fact predictable on the basis of a general constraint operative in the DS framework which prohibits the co-occurrence of more than one unfixed node, thereby also confirming the claim of Dynamic Syntax to constitute a grammar framework rather than merely a parsing device. [ABSTRACT FROM AUTHOR]

We performed an exhaustive meta-analysis of 73 peer-reviewed journal articles on syntactic priming from the seminal Bock (1986) paper through 2013. Extracting the effect size for each experiment and condition, where the effect size is the log odds ratio of the frequency of the primed structure X to the frequency of the unprimed structure Y , we found a robust effect of syntactic priming with an average weighted odds ratio of 1.67 when there is no lexical overlap and 3.26 when there is. That is, a construction X which occurs 50% of the time in the absence of priming would occur 63% if primed without lexical repetition and 77% of the time if primed with lexical repetition. The syntactic priming effect is robust across several different construction types and languages, and we found strong effects of lexical overlap on the size of the priming effect as well as interactions between lexical repetition and temporal lag and between lexical repetition and whether the priming occurred within or across languages. We also analyzed the distribution of p -values across experiments in order to estimate the average statistical power of experiments in our sample and to assess publication bias. Analyzing a subset of experiments in which the primary result of interest is whether a particular structure showed a priming effect, we did not find evidence of major p -hacking and the studies appear to have acceptable statistical power: 82%. However, analyzing a subset of experiments that focus not just on whether syntactic priming exists but on how syntactic priming is moderated by other variables (such as repetition of words in prime and target, the location of the testing room, and the memory of the speaker), we found that such studies are, on average, underpowered with estimated average power of 53%. Using a subset of 45 papers from our sample for which we received raw data, we estimated subject and item variation and give recommendations for appropriate sample size for future syntactic priming studies. [ABSTRACT FROM AUTHOR]

• A later start of English as a foreign language (EFL) lessons is beneficial to pupils with DLD. • Pupils with more positive attitudes towards EFL lessons obtain higher scores. • In pupils with DLD, more EFL anxiety is associated with lower achievement. • Predictors of EFL achievement change over time. Like children with typical language development, their peers with developmental language disorder (DLD) are expected to learn English as a foreign language (EFL). For pupils without DLD, it is well-established that amount of informal exposure to English outside of the classroom, starting age of EFL instruction and motivation are strong positive predictors of EFL learning rate and/or achievement, whereas anxiety is negatively related to performance. This paper is the first attempt to investigate how these predictors of EFL performance operate in learners with DLD. Participants were nineteen Dutch-speaking 7th graders with DLD learning English as a school subject at a specialist education facility in the Netherlands. English receptive grammar and receptive vocabulary were measured twice, with a four-month interval. Foreign language learning motivation, anxiety and (length and amount of) informal exposure to and instruction in English were measured via questionnaires. The participants did not show any progress on English vocabulary and grammar. At Time 1, vocabulary and grammar scores were positively related to starting age of EFL instruction and negatively related to anxiety. For vocabulary, achievement was also positively predicted by attitudes towards English lessons. Only the relationship between starting age of instruction and vocabulary outcomes was visible at Time 2. Amount and length of informal exposure to English did not predict performance, which is in stark contrast to the patterns observed in EFL learners with typical language development. We conclude that children with DLD benefit from a later onset of foreign language lessons, whereas length and amount of out-of-school exposure to English are less important in the context of DLD, possibly due to difficulty with implicit learning. [ABSTRACT FROM AUTHOR]

[Display omitted] • 4 speeded continuation tasks investigated agreement attraction in Romanian. • Bare nouns cannot typically be preverbal subjects in Romanian. • Bare noun intervenors cause less attraction than modified DP intervenors. • Cue-based retrieval is sensitive to the overtness of the determiner. This paper investigates whether agreement attraction is modulated by distributional properties determining subject-likelihood by examining the degree to which bare nouns and full determiner phrases (DPs) cause agreement attraction effects in Romanian. Romanian represents an ideal testing ground for this, given two distributional constraints making bare nouns less subject-like: Locative Determiner Omission , preventing locative prepositions from taking nouns with definite articles (unless modified by adjectives), and the Naked Noun Constraint , disallowing bare nouns as preverbal subjects. We predicted that bare nouns should be less likely to trigger agreement attraction than overt DPs. We conducted four speeded forced-choice sentence continuation tasks on Romanian native speakers to test this prediction. We observe that overt DPs cause significantly more attraction than bare nouns. We suggest that the results are consistent with a cue-based retrieval mechanism for forming agreement dependencies, where cues that determine subjecthood are used to reactivate elements in working memory upon processing a verb. These cues can be language specific, and in Romanian, this means that agreement attraction is sensitive to the morphophonological overtness of the determiner. [ABSTRACT FROM AUTHOR]

• This work demonstrates the limitations of a transformer-based neural network coupled with conditional random field for the task of persuasive dialogues. • In this work, we consider persuasion strategy recognition as a sequential tagging problem, using transformers and conditional random field to model context information and inter-speaker dependency. Persuasion aims at forming one's opinion and action via a series of persuasive messages containing persuader's strategies. Due to its potential application in persuasive dialogue systems, the task of persuasive strategy recognition has gained much attention lately. Previous methods on user intent recognition in dialogue systems adopt recurrent neural network (RNN) or convolutional neural network (CNN) to model context in conversational history, neglecting the tactic history and intra-speaker relation. In this paper, we demonstrate the limitations of a Transformer-based approach coupled with Conditional Random Field (CRF) for the task of persuasive strategy recognition. In this model, we leverage inter- and intra-speaker contextual semantic features, as well as label dependencies to improve the recognition. Despite extensive hyper-parameter optimizations, this architecture fails to outperform the baseline methods. We observe two negative results. Firstly, CRF cannot capture persuasive label dependencies, possibly as strategies in persuasive dialogues do not follow any strict grammar or rules as the cases in Named Entity Recognition (NER) or part-of-speech (POS) tagging. Secondly, the Transformer encoder trained from scratch is less capable of capturing sequential information in persuasive dialogues than Long Short-Term Memory (LSTM). We attribute this to the reason that the vanilla Transformer encoder does not efficiently consider relative position information of sequence elements. [ABSTRACT FROM AUTHOR]

In this paper, we show that pure insertion grammars of size 2 (i.e. , inserting two symbols in a left and right context, each consisting of two symbols) can characterize all recursively enumerable languages. This is achieved by either applying an inverse morphism and a weak coding, or a left (right) quotient with a regular L O C (2) language, or an intersection with a L O C (2) language and a weak coding. The obtained results improve the descriptional complexity of insertion grammars and complete the picture of known results on insertion-deletion systems that are motivated from the DNA computing area. [ABSTRACT FROM AUTHOR]

In this paper we present a simple linear-time algorithm constructing a context-free grammar of size O ( g log ⁡ ( N / g ) ) for the input string, where N is the size of the input string and g the size of the optimal grammar generating this string. The algorithm works for arbitrary size alphabets, but the running time is linear assuming that the alphabet Σ of the input string can be identified with numbers from { 1 , … , N c } for some constant c . Otherwise, additional cost of O ( N log ⁡ | Σ | ) is needed. Algorithms with such an approximation guarantee and running time are known, the novelty of this paper is a particular simplicity of the algorithm as well as the analysis of the algorithm, which uses a general technique of recompression recently introduced by the author. Furthermore, contrary to the previous results, this work does not use the LZ representation of the input string in the construction, nor in the analysis. [ABSTRACT FROM AUTHOR]

In a recent paper (M. Barash, A. Okhotin, “An extension of context-free grammars with one-sided context specifications”, Inform. and Comput. , 2014), the authors introduced an extension of the context-free grammars equipped with an operator for referring to the left context of the substring being defined. This paper proposes a more general model, in which context specifications may be two-sided, that is, both the left and the right contexts can be specified by the corresponding operators. The paper gives the definitions, presents several examples of grammars and establishes a basic normal form theorem. This normal form, in particular, leads to a simple parsing algorithm working in time O ( n 4 ) , where n is the length of the input string. [ABSTRACT FROM AUTHOR]

Online network platforms provide great convenience for users to obtain information. However, it's challenging to select the required information from enormous texts. Automatic text headline generation methods not only guide users to select the information they are interested in, but also solve the problem of information overload. Nevertheless, the existing works mainly utilize the grammar rules to obtain the key information of the source text, while ignoring the dwell time of user's attention on different text contents. To address this issue, this paper proposes an abstractive text headline generation model based on the eye-tracking attention mechanism. Specifically, this model first relies on the eye-tracking data to establish the mapping relationship between text words and the words' reading time. Then, an eye-tracking attention mechanism is constructed to judge the importance of different words. Finally, this attention mechanism is integrated into the encoder-decoder framework to generate a high-quality headline. Experimental results obtained from different datasets demonstrate that the headline generated by our model is more concise. Moreover, our proposed model outperforms significantly the classical headline generation models on ROUGE-1, ROUGE-2 and ROUGE-L. [ABSTRACT FROM AUTHOR]

Abstract: It follows from the ordering principles that are applied in Functional Discourse Grammar that the positional possibilities of markers of agreement and those of crossreference are different. Markers of crossreference are predicted to occur closer to the verb stem, while markers of agreement would occupy peripheral positions. This paper tests these predictions using data from a variety of languages, and shows that for these languages the predictions indeed hold true. In demonstrating this, the paper furthermore proposes a new treatment for markers of agreement/crossreference in languages in which these optionally co-occur with a corresponding noun phrase. These markers are on a language-specific basis classified as either Contextual Agreement Markers or as Appositional Referential Markers. [Copyright &y& Elsevier]

Abstract: In this paper we offer a panoramic overview of the development of Andean Spanish and Andean Linguistics, from a theoretical point of view of language contact and universal grammar. We discuss how the notion of Andean Spanish came into existence, the issues under debate, and the consequences of different theoretical positions. We also introduce the contents of the papers included in this special issue, and highlight their key points, framing them in the context of the research on universal grammar. The following five areas are covered: (1) the convergence of Quechua and Aymara, (2) the impact of bilingualism on intonation, (3) the sharing of suffixes in contact situations, (4) the modification of the present perfect tense in Spanish in contact with Quechua, and (5) the properties of the Quechua case system from the point of view of contact varieties. Finally, we provide some remarks on bilingual education, and some suggestions for further directions for future research. [Copyright &y& Elsevier]

• Pseudo Relative (PR) availability modulates Relative Clause (RC) attachment in French. • Tense mismatch forces RC-parse, lowering acceptability of locally ambiguous strings. • RC-disambiguation leads to longer fixation durations at critical region. • No effects of tense manipulation in globally unambiguous RCs. • No main effect/interaction in English, a language which disallows PRs. We present the results of three offline questionnaires (one attachment preference study and two acceptability judgments) and two eye-tracking studies in French and English, investigating the resolution of the ambiguity between pseudo relative and relative clause interpretations. This structural and interpretive ambiguity has recently been shown to play a central role in the explanation of apparent cross-linguistic asymmetries in relative clause attachment (Grillo and Costa, 2014; Grillo et al., 2015). This literature has argued that pseudo relatives are preferred to relative clauses because of their structural and interpretive simplicity. This paper adds to this growing body of literature in two ways. First we show that, in contrast to previous findings, French speakers prefer to attach relative clauses to the most local antecedent once pseudo relative availability is controlled for. We then provide direct support for the pseudo relative preference: grammatically forced disambiguation to a relative clause interpretation leads to degraded acceptability and greater processing cost in a pseudo relative environment than maintaining compatibility with a pseudo relative. [ABSTRACT FROM AUTHOR]

So-called psychological verbs such as Italian temere 'fear', preoccupare 'worry', and piacere 'like' denote a particular state that involves an Experiencer and a second role taker that causes, initiates or is related to the psychological state. They present an extremely varied argument structure across languages that arranges these two roles in apparently inverted hierarchies and assigns them different grammatical functions (subject, direct, indirect and prepositional objects). This paper aims to provide a descriptively adequate taxonomy of psych-verbs in Latin in a comparative perspective with Italian. We individuate seven classes of psych-verbs and show that they distribute across the transitive, unergative, unaccusative pattern with the possibility of externalising either argument, therefore creating three "direct" and three "inverted" classes. The seventh class is impersonal, with no external argument. We show that the diachronic variation and apparent idiosyncrasies displayed by some verbs can be explained by the proposal that the seven classes are potentially available to all psych-roots. For this reason, psych-verbs present a high degree of vulnerability in language contact and change which results in intra-language optionality and diachronic variation. [ABSTRACT FROM AUTHOR]

Highlights • The domain of emphasis spread in Urban Jordanian Arabic is the morpheme. • Leftward spreading is almost unbounded within the morpheme, but rightward spreading is relatively more constrained. • Emphasis spread is equally applicable to both monosyllabic and polysyllabic monomorphemic words. • Morpheme boundaries are opaque to emphasis spread. Abstract This paper investigated the domain of emphasis spreading in Urban Jordanian Arabic. Emphasis spreading was examined in two types of words: polysyllabic monomorphemic and polymorphemic. F1 raising, F2 lowering, and F3 raising in the vowels preceding and following the emphatic sound were used as the acoustic correlates of emphasis spreading in the participants' speech who were ten native speakers of Urban Jordanian Arabic. The findings showed that emphasis is a morphophonemic process in that the domain of emphasis spread in Urban Jordanian Arabic is the morpheme rather than the syllable or the word. Second, the study showed that emphasis spreading applies leftward and rightward in all environments but with some significant differences. Morpheme boundaries were found opaque to emphasis spreading. [ABSTRACT FROM AUTHOR]

Highlights • I present a reply to Kim (2018), which is a critique of Lin (2017a). • I show that Kim has misunderstood the main point of Lin (2017a). • I dissolve Kim's objections and dispel the methodological confusions in UG embedded in his arguments. Abstract Kim (2018) argues that the method adopted in universal grammar (UG) is the same as that employed in the empirical sciences such as physics, chemistry and biology, and that my rejection of UG in Lin (2017a) amounts to a rejection of these sciences. In this reply I show that Kim has misunderstood the main point of my original paper, which is that UG is crucially different from typical scientific theories in that it does not consider causal factors. I further show that once this misunderstanding is cleared his objections can be dissolved smoothly, that the fatal problem with UG will become more evident, and that the methodological confusions embedded in Kim's arguments will be dispelled. [ABSTRACT FROM AUTHOR]

Abstract In this paper, we report on a qualitative analysis of co-occurring discourse markers, that is, sequences of adjacent discourse markers that belong to the same unit but may express different functions. We examine several formal and functional features of these co-occurring strings on the basis of corpus examples extracted from conversational data in English. In particular, we focus on scope, meaning-in-context (or functions), syntactic category and position. Our analysis reveals several degrees of integration: differences in scope allow us to differentiate juxtaposition and combination of markers. In the case of combination, difference in meaning integration allows us to distinguish addition from composition of markers. We pay particular attention to in-between and ambiguous cases such as and so or and then , which instantiate different degrees in our cline of co-occurrence depending on the meaning that can be interpreted from the cluster. We finally discuss the implications of such fine-grained distinctions from the perspective of systematic corpus annotation. Highlights • Confrontation of a Romance model to English data • Operational definition of three degrees of co-occurrence • Identification of ambiguous cases, fuzzy borders and non-co-occurring clusters • Comprehensive model covering the syntax, pragmatics and prosody of discourse markers [ABSTRACT FROM AUTHOR]

Highlights • Grammatical conditioning constraints govern (ING) realization in words like talking. • Listeners are sensitive to this probabilistic information in speech perception. • Such expectations are weighted more heavily with less reliable bottom-up info. • Listeners track probabilistic info even when not crucial for word disambiguation. Abstract This paper investigates the extent to which listeners are cued into the systematicity of variability in speech, particularly the grammatical conditioning constraints of the English sociolinguistic variable (ING) (e.g., talking vs. talkin). Listeners’ sensitivity to the realization of (ING) words embedded in sentences was tested under various conditions. Comprehenders demonstrated expectations about the grammatical category constraints conditioning the realization of (ING) even though such knowledge may not be very informative about word recognition (Experiment 1). As more reliable phonetic information was available, listeners weighted grammatical expectations less, favoring other cues in the signal (Experiments 2 and 3). Bridging a gap between psycholinguistic and sociolinguistic accounts of probabilistic conditioning, these results suggest that factors beyond utility for word recognition may contribute to the probabilistic monitoring of a variable, and underscore the opportunistic nature of the speech comprehension system, where listeners make use of whatever information they have to process the signal. [ABSTRACT FROM AUTHOR]

Abstract A neutrosophic uncertain linguistic variable is composed of an uncertain linguistic variable and a neutrosophic number, which has more advantages than both of uncertain linguistic variable and neutrosophic number in expression. Therefore, the decision-making theory and its related methods based on neutrosophic uncertain linguistic variables have received wide attention of scholars. This paper mainly studies the new expression and operations of single value neutrosophic uncertain linguistic variables and its application in multiple attribute group decision-making (MAGDM). First, a new form of single value neutrosophic uncertain linguistic set (NFSVNULS) and its operational rules are defined. Then, a new form of single value neutrosophic uncertain linguistic variable weighted arithmetic average (NFSVNULVWAA) operator and a new form of single value neutrosophic uncertain linguistic variable weighted geometric average (NFSVNULVWGA) operator are proposed. Furthermore, a MAGDM method based on the proposed aggregation operators is put forward. Finally, an example of investment is used to demonstrate the feasibility and effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]

Abstract: Halliday''s Systemic Functional Linguistics (SFL) often pays tribute to one of the great 20th century figures in European Linguistics, Louis Hjelmslev, and his theory of Glossematics. This is surprising because Glossematics is a formally oriented, strictly immanent approach to language while SFL is a functional theory focusing on language as a social semiotic. This paper offers an examination of both theories and attempts to trace similarities and differences on a number of issues with which they share a theoretical concern: the role of paradigmatic relations, the notions of function, text, stratification and immanence, and the overall descriptive goals in linguistics. The paper concludes that except for the basic recognition of the primacy of paradigmatic relations and the abstract idea of the stratification of the content plane, there is little or no similarity between Glossematics and SFL. In fact, many of the references to Glossematics in SFL writings may give the reader an inaccurate impression of Hjelmslev. More generally, while accepting that finding inspiration in a theory does not necessarily involve an obligation to remain faithful to that theory, the paper argues that we all still owe it to earlier generations of scholars to recognize their work for what it really was. [Copyright &y& Elsevier]

Abstract: The paper argues that the notion of emphasis has to be part of the grammar of German. It is demonstrated that in German, Ā-movement to the left periphery of a declarative clause is associated with a conventional implicature which encodes the emphatic interpretation of the moved item. The appropriate notion of emphasis is characterized and it is shown how it differs from a coarse concept of contrast. The paper also studies some IP-internal constructions whose elements’ interpretation has been assumed to be related to emphasis/contrast. It is argued that for the description of these phenomena, the standard notion of focus is adequate. The paper concludes that the syntactic marking of emphasis is a characteristic feature of Ā-movement. [Copyright &y& Elsevier]

Abstract: In this paper we argue that variation in subject-verb agreement in Belfast English and Appalachian English bears on a number of issues that are relevant to current syntactic theory, including clause structure and subject positions, case, the syntax of negative polarity items and negative concord and the structure of the DP. We show that while the nature of the subject plays a role in determining subject-verb agreement in the two varieties (in different ways), subject position is the overt structural manifestation of the differing properties of the subject, independent of which properties are relevant in the variety in question. Our proposal on subject positions, which goes beyond what proposed for Belfast English, is able to capture clusters of properties not discussed by Henry, and thus more fully accounts for the range of micro-parametric variation we find. This paper thus provides support for a configurational approach to subject-verb agreement. [Copyright &y& Elsevier]

Abstract: In the literature on gender assignment, some scholars claim that semantic assignment rules always win, others that there is no type-wise ranking (i.e., ranking according to whether the rules are semantic, morphological or phonological) at all. This discussion necessitates a further look at semantic assignment rules that have nothing to do with the semantic core for gender. Such non-core rules are referred to as “crazy” here. Evidence for non-core rules is presented, and the paper presents criteria for distinguishing between plausible and implausible such rules. Some rules that have been proposed in the literature clearly fall short of the criteria. Other ‘crazy’ rules fare better. The paper also presents a scenario for the origin of crazy rules and discusses what they tell us about gender and about the nature of linguistic generalisations, which can be quite “local”. [Copyright &y& Elsevier]

Abstract: This paper brings to light the contribution the early German semasiologists – most prominently Christian Karl Reisig, Friedrich Haase, and, later, Ferdinand Heerdegen – made to the establishment of semantic theorising as a genuine discipline in linguistics in general and, specifically, in grammaticography. Prior to Bréal, who is often credited with having inaugurated the discipline, these authors developed a programme for a theory of meaning as a part of grammar in its own right. The paper begins with the initial framework proposed by Reisig, who introduced an architecture of grammar that comprises, in addition to the traditional elements “etymology” and “syntax”, an explicit semasiological component. It then investigates the various responses to Reisig’s trichotomy of grammar in 19th century linguistics, in particular in the work of Reisig’s followers Haase and Heerdegen, but also in the work of authors outside of the semasiological movement. As part of the analysis, the origin of the term and concept “semasiology” is traced, from its first mention in Reisig’s lectures and its subsequent spread to its gradual substitution, as a cover term for the theory of meaning, by Bréal’s term “semantique”, and finally to its established present-day use. The second and more specific focus in the paper is on the accounts the early semasiologists give of word formation, and on the place these accounts have in their overall conception of grammar. This particular subject has been virtually unaddressed in historiographic studies so far, and Schmitter’s paper is a substantial step towards the full appreciation of the historical contribution these early theories of word formation made to the development of the field. [Copyright &y& Elsevier]

Abstract: This paper provides a typological interpretation of differences between Chinese and English in grammatical metaphor (GM), a phenomenon arising from the interaction of semantics and lexicogrammar and extending the meaning potential in a language. This paper first describes typological features in Chinese and English in terms of the three variables relevant most to the lexicogrammatical realization of a semantic meaning, namely the order of grammatical constituents, the degree of grammatical specificity and the location of grammatical realization. On the basis of the description of typological features, this paper explores the GM differences in Chinese and English and their inherent reasons from three perspectives: (1) the formal distance between congruent and metaphorical expressions, (2) the subdivision of certain GM categories and (3) the extent of using GM. The study reveals that all the GM differences observed are engendered by one or more than one typological variation between Chinese and English. [Copyright &y& Elsevier]

Abstract: A basic issue in the neurosciences of language is whether an L2 can be processed through the same neural mechanism underlying L1 acquisition and processing. In the present paper I review data from functional neuroimaging studies focusing on grammatical and lexico-semantic processing in bilinguals. The available evidence indicates that the L2 seems to be acquired through the same neural structures responsible for L1 acquisition. This fact is also observed for grammar acquisition in late L2 learners contrary to what one may expect from critical period accounts. However, neural differences for an L2 may be observed, in terms of more extended activity of the neural system mediating L1 processing. These differences may disappear once a more ‘native-like’ proficiency is established, reflecting a change in language processing mechanisms: from controlled processing for a weak L2 system (i.e., a less proficient L2) to more automatic processing. The neuroimaging data reviewed in this paper also support the notion that language control is a crucial aspect specific to the bilingual language system. The activity of brain areas related to cognitive control during the processing of a ‘weak’ L2 may reflect competition and conflict between languages which may be resolved with the intervention of these areas. [Copyright &y& Elsevier]

Abstract: This paper deals with the quantity system of Dinka, a Nilo-Saharan language. Andersen [(1987). The phonemic system of Agar Dinka. Journal of African Languages and Linguistics, 9, 1–27] has postulated that Dinka has a three-level vowel length distinction, i.e., /V/ vs. /VV/ vs. /VVV/. However, it has been argued that three-level vowel length distinctions may not exist in human language, because they may not be reliably produced and perceived [Kohler, K.J. (2001). Überlänge im Niederdeutschen? In: R. Peters, H. P. Pütz, & U. Weber (Eds.), Vulpis Adolatio—Festschrift für Hubertus Menke zum 60. Geburtstag (pp. 385–402). Heidelberg: C. Winter]. Moreover, Andersen''s analysis challenges the widely held view that phonological distinctions are maximally binary (see, e.g. Odden, D. [(1997). Some theoretical issues in Estonian prosody. In: I. Lehiste & J. Ross (Eds.), Estonian prosody: Papers from a symposium (pp. 165–195). Tallinn: Institute of Estonian Language]). In this paper we examine the Dinka quantity system by means of an acoustic analysis. This acoustic analysis reveals that the contrast in question can be accounted for best by the three-level vowel length hypothesis. Also, the realisation of the distinction provides insight as to why three-level vowel length distinctions are uncommon. Our results suggest that the phonetic space for quantity distinctions is limited, so that any increase in categories reduces the phonetic distance between them. [Copyright &y& Elsevier]

This paper discusses the role played by cognition in three linguistic theories which may be labelled as ‘structural-functional’: Functional (Discourse) Grammar, Role and Reference Grammar and Systemic Functional Grammar. It argues that if we are to achieve true cognitive adequacy, we must go well beyond the grammar itself to include the processes of conceptualisation, categorisation and construal, and that the model should be dynamic, in the sense that it should describe the processes of language production and comprehension rather than only the patterns found in languages. The discussion provides evidence that none of the three theories can currently be considered to achieve high standards of cognitive adequacy, despite the efforts made in this direction in Functional Discourse Grammar. Recent work which links Role and Reference Grammar to studies of production and comprehension is, however, promising. The paper ends with a plea for a more highly developed synthesis of work in linguistics, psycholinguistics, cognitive psychology, neurology and neurolinguistics, integrated with perspectives from the study of the sociocultural contexts in which language is used. [Copyright &y& Elsevier]

Abstract: Considerable research has now been undertaken into the development of different approaches to exploiting language corpora for pedagogic purposes in the context of ESP. The question of how language corpora might be utilized by students beyond the immediate language-teaching context is, however, one as yet seldom addressed in the literature. This study attempts to explore the relationship between student use of online corpus tools and academic and professional discourse practices in the context of a professional legal training course at The City University of Hong Kong. Students enrolled in this course were given instruction in how to consult an online concordancer as language support when completing their legal writing assignments. Drawing on narratives of student experience, and other informant data including detailed logs of searches and the outcomes of assessments of English language proficiency, the paper discusses the ways in which students make strategic use of the corpus tools provided to develop competence in writing for legal purposes. The paper concludes by appraising the potential of corpus-based methods as an affordance for studying the practice of Law, in particular as a means of enhancing the acquisition of professional expertise by novice lawyers. [Copyright &y& Elsevier]

Abstract: This paper extends a previous claim that auxiliary selection in the present perfect of monadic verbs reflects a pattern grounded in lexico-aspectual properties of individual verbs though mediated through their syntax. A novel argument in support of this conclusion is made, based on the cross-contextual typology of auxiliaries, showing that the cross-linguistic distribution of auxiliaries in the present perfect cannot be understood in isolation of other constructions requiring an auxiliary. In particular, it is argued that two sets of constraints restrict the mapping from lexico-aspectual features to a particular auxiliary (have or be). One set restricts the mapping from lexical-aspectual features to argument structure at the lexicon-syntax interface, the other restricts the mapping from argument structure to the marked member of the auxiliary pair (have). The analysis succeeds in predicting exactly 15 possible languages or patterns of auxiliary selection with monadic verbs, a variety of which are discussed throughout the paper. While the fundamental syntactic distinction made by the Unaccusative Hypothesis is confirmed by the formal analysis, it is argued that the present perfect imposes restrictions of its own on the argumental status of any verb entering the construction, on a par with constructions such as passives and resultatives. [Copyright &y& Elsevier]

This paper argues that Nepali is neither a fully ergative language nor a simple split-ergative language conditioned by tense/aspect, because the language shows a more complex case-marking pattern than previously described. Furthermore, the paper demonstrates that Nepali is a split-ergative language conditioned by the semantic nature of NPs. In the domain of inanimate NPs, the language is ergative; elsewhere, it resists classification as ergative or accusative. This finding is of typological significance in that it provides a good example for a rare split between inanimate NPs and animate NPs. Moreover, the paper argues that Nepali shows split intransitivity in terms of case marking, and it offers a semantic account of the split in terms of the interaction of two semantic factors, agentivity and telicity. [Copyright &y& Elsevier]

Abstract: This paper takes the view that loanword adaptation results from attempts to match the non-native percept of the L2 input, within the confines of the L1 grammar. Neither a purely perceptual nor a purely grammatical account can explain the facts. The L1 grammar includes constraints specific to mimicking foreign inputs. Where not all aspects can be replicated, the grammar prioritizes. In the Cantonese vowel data examined here, tone (the reflex of English stress) and consonants are more carefully replicated than vowel quality, and vowel length is even less important than vowel quality. The paper examines acoustic and psycholinguistic data on L2 vowel perception, and looks closely at the adaptation of English [æ] and [▪] in English loans into Cantonese, which lacks both these vowels. [Copyright &y& Elsevier]

Summary: Purpose: Over recent years, there has been a growing interest in extracting entities and relations from biomedical literature. There are a vast number of systems and approaches being proposed to extract biological relations, but none of them achieves satisfactory results. These methodologies are either parsing-based or pattern-based, which are not competent to handle the grammatical complexities of biomedical texts, or too complicated to be adapted. It is well known that appositive, coordinative propositions and such grammatical structures are extremely common in biomedical texts, particularly in full texts. However, these problems are still untouched for most of researchers. Methods: In this paper, we have proposed a new approach, which is hybrid with both shallow parsing and pattern matching, to extract relations between proteins from scientific papers of biomedical themes. In the method, appositive and coordinative structures are interpreted based on the shallow parsing analysis, with both syntactic and semantic constraints. Then long sentences are splitted into sub-ones, from which relations are extracted by a greedy pattern matching algorithm, along with automatically generated patterns. Results: Our approach is experimented to extract protein–protein interactions from full biomedical texts, and has achieved an average F-score of 80% on individual verbs, and 66% on all verbs. With the help of shallow parsing analysis, pattern matching is improved remarkably. Compared with the traditional pattern matching algorithm, our approach achieves about 7% improvement of both precision and F-score. In contrast to other systems, our approach achieves performance comparable to the best. A demo system has been available at http://spies.cs.tsinghua.edu.cn. [Copyright &y& Elsevier]

This paper argues against analyzing grammatical category as a morphological epiphenomenon irrelevant for syntax. Using evidence from English and Greek it refutes the account in Barner and Bale [Lingua 112 (2002), 771], whereby coinages are derivations of the free syntactic insertion of roots in nominal and verbal environments, and presents evidence suggesting they are products of recta-linguistic processes. Rehearsing the discussion in Chomsky's Remarks on Nominalization [Readings in English Transformational Grammar (1970) 184], the paper goes on to show that seeking to syntactically derive nouns and verbs from roots stumbles upon the idiosyncratic differences many noun-verb pairs display between them. Towards this, it foregrounds the role of non-zero derivational morphology in Greek and Hungarian. [ABSTRACT FROM AUTHOR]

Japanese has a postpositional particle wa which has been categorized as the expression of ‘literal contrast’ (e.g. John ate a pizza, but Mary didn't). This paper reexamines the functions of contrastive wa based on the collection of newspaper articles from the viewpoint of dialogical voices, and claims that contrastive wa manifests itself as the expression of the author's multivoicedness, which often derives from our consciousness experiences (association, evaluation, conflict). The paper argues that the dialogicality of contrast does not exist in its own right but co-exists with pragmatic acts by which author and reader cooperate in a given context to arrive at agreement on the intended meanings. The import of the discovery of the author's implicit voices is tentatively stated to reflect pedagogical implications. [ABSTRACT FROM AUTHOR]

French questions come into two brands: a movement version and an in-situ alternative. In this context, the present paper argues for a flexible one-to-one mapping between form and interpretation: fronting of the nominal typically correlates with one reading, while the in-situ nominal may, but need not, correlate with another. The contention is nevertheless that there is no such thing as optional WH-movement in French: the EPP/D feature associated with [+WH] in C is always strong; what varies is the placement of the noun. In the dialect/register reviewed in this study, the default interpretation associated with the in-situ position is non-specific while the raising of the noun favours a specific interpretation. Moreover, it is argued that pragmatically, in this variant of French, in-situ fronted-questions are adequate in a prominent context whereas their WH-in-situ counterparts are only felicitous in a non-prominent context. One assumption on which the proposal is based is that single WH-constructions in French are instances of split-DP constructions: a bare operator is separated from its associated semantic nominal restriction in the course of the derivation. Finally, it will be shown that the semantic and discourse properties of in-situ/stranded nominals in the variety of the language examined in this paper follow from the fact they are semantically incorporated. They denote, not an argument, but a property, crucially without scope of their own. Coupled with the assumption that French WH in-situ constructions are split-constructions, the fact that in-situ WH-nominals are semantically incorporated explains the intervention effects exhibited by such constructions. The existential quantifier is provided by the verb. Thus, the existential quantifier will necessarily take scope under all other scopal elements present in the sentence, including negation. [ABSTRACT FROM AUTHOR]

This paper investigates the processing of long-distance filler-gap dependencies in Japanese, a strongly head-final language. Two self-paced reading experiments and one sentence completion study show that Japanese readers associate a fronted why-phrase with the most deeply embedded clause of a multi-clause sentence. Experiment 1 demonstrates this using evidence that readers expect to encounter a scope-marking affix on the verb of an embedded clause in why-fronting constructions. Experiment 2 shows that the we phrase is already associated with the embedded clause before the embedded verb is processed, based on a Japanese counterpart of the Filled Gap Effect (Stowe, 1986). Experiment 3 corroborates these findings in a sentence completion study. These findings clarify the factors responsible for 'active filler' effects in processing long-distance dependencies (Crain & Fodor, 1985; Fodor, 1978; Frazier & Clifton, 1989; Stowe, 1986) in ways not possible in head-initial languages. The results provide evidence that the processing of filler-gap dependencies is driven by the need to satisfy thematic role requirements of the fronted phrase, rather than by the need to create a gap as soon as possible. The paper also discusses implications of these findings for theories of reanalysis. [ABSTRACT FROM AUTHOR]

This paper highlights and accounts for some interesting properties of the superlative construction in Modern Standard Arabic that have not been addressed by previous research on Arabic. One property that is focused on in this paper is the interaction between number and definiteness in this construction. It is shown that superlative adjectives have properties that are different from those of the abstract adjectives, and similar to those of quantifiers and nouns. Superlatives are similar to nouns and quantifiers in that they can host clitics, modify or quantify over noun phrases, and that these noun phrases should always be assigned the genitive case. Given these similarities, it is argued that the superlative construction behaves like a Construct State (CS), which is a nominal construction in Arabic that roughly corresponds to the English genitive construction. On a par with other elements that can head the CS (i.e., numerals, quantifiers, and nouns), the superlative adjective has nominal head properties and is the head of the superlative construction. However, it is shown that there is number ambiguity in the superlative construction that is not present in regular CS constructions, with a noun or a quantifier head. Number ambiguity shows in this construction only when the noun phrase modified by the superlative adjective is definite and plural. In this case, the superlative construction can be either singular or plural. This paper attributes this ambiguity to the quantificational nature of superlative adjectives and its interaction with definiteness within the superlative construction. Superlatives in Arabic, thus, are established as quantifiers, and the superlative construction as a Quantifier Phrase. [Copyright &y& Elsevier]

The efficacy of teacher error/grammar correction in second language writing classes has been the subject of much controversy, including a published debate in an earlier volume of this journal [J. Second Language Writing 8 (1999) 1; J. Second Language Writing 8 (1999) 111]. In this paper, the state-of-the-art in error correction research in L2 writing is described (“Where are we?”), directions for future research are outlined (“Where do we go from here?”) and implications for current L2 composition pedagogy are suggested (“What do we do in the meantime?”). The primary thesis of the paper is that, despite the published debate and several decades of research activity in this area, we are virtually at Square One, as the existing research base is incomplete and inconsistent, and it would certainly be premature to formulate any conclusions about this topic. Thus, findings from previous research on this controversial yet ubiquitous pedagogical issue are recast as “predictions” about what future research might discover, rather than “conclusions” about what the previous research shows us. [Copyright &y& Elsevier]

The use of first and second person pronouns is an important indicator of how audiences are conceptualized by speakers and writers in academic discourse. Several grammarians and linguists have studied more or less specifically this part of the English language. Two of the most widely discussed aspects of pronouns are their reference and their discourse function. In this paper I present research on the use of the pronoun ‘we’ in academic speech. I selected this pronoun since a previous study had signalled it as the most frequent in this type of language. That study was conducted by Rounds in 1985–1987 using a corpus from the University of Michigan. The present research was carried out using the MICASE (Michigan Corpus of Spoken Academic English), also from the University of Michigan. This paper reveals important results about the use of ‘we’ in English academic oral discourse, especially related to reference and discourse function. Some of these results seem to contradict those obtained by Rounds, whereas others support her research and complement it. [Copyright &y& Elsevier]

The aim of this paper is to present a contrastive analysis of reformulation markers in English, Spanish and Catalan. The study is based on a corpus of expository prose (mainly academic writing). The analysis shows coincidences in the sources of the markers and differences in the variety of forms expressing reformulation as well as in their frequency of use. Assuming that grammaticalization processes such as those leading to the creation of connectives have their roots in discourse, parallel differences are expected to be found in text construction. The present paper argues that the differences identified in the grammar of English vs. those of Spanish and Catalan can be associated with two distinct styles for building academic texts. [Copyright &y& Elsevier]

Adjuncts may occur (by ‘adjunct preposing’) before a wh-interrogative clause which is a main clause, but not before one which is subordinate; for example: (i) Tomorrow what shall we do? (ii) I told you (*tomorrow) what we shall do. Why should the possibility of adjunct preposing vary between main and subordinate clauses? The pre-theoretical answer is obvious: the wh-word has the extra function in a subordinate clause of signalling the start of a subordinate clause, so like any other subordinator it must be the first element in its clause. Less obvious is how to capture this insight in a formal grammar, and the paper will show that this challenge favours flexible word-based grammars over the more familiar kind which assign a uniform clause structure. The paper considers and rejects a number of examples of the latter approach, especially that of Rizzi [Rizzi, L. 1997. The fine structure of the left periphery. In: Haegeman, L. (Ed.), Elements of Syntax. A Handbook in Generative Grammer. Klewer, Dordrecht, pp. 281–337]. The proposed solution is based on enriched dependency structure (Word Grammar) which makes head-hood ambiguous in certain constructions. In particular, the head of a wh-interrogative may be its finite verb when it is a main clause but must be the wh-element when it is subordinate. [Copyright &y& Elsevier]

This paper aims to illustrate how we might proceed to attain repeatability and falsifiability in generative grammar, on the basis of discussion of concrete empirical materials in Japanese, such as bound variable anaphora, quantifier scope, and reconstruction effects. After critically examining and improving on the major empirical arguments for the standard view of the phrase structure of Japanese, I propose that a certain type of bound variable anaphora is based on a formal relation, Formal Dependency (FD), between two argument positions. The structural condition on FD and its interpretive consequence proposed here not only enable us to account for the fluctuation and instability of speakers'' judgments but also lead us to predict a number of hitherto unnoticed correlations with regard to the (un)availability of a bound variable construal. The paper attains a significantly higher level of repeatability than before, states explicitly how its proposal can be falsified, and stresses the critical importance of making a rigorous attempt to articulate how every theoretical concept is related to the rest of the theory as well as to the native speaker''s linguistic intuitions so as to be able to make definite predictions and ensure the falsifiability of our hypotheses. [Copyright &y& Elsevier]

Semi-supervised learning has greatly advanced medical image segmentation since it effectively alleviates the need of acquiring abundant annotations from experts, wherein the mean-teacher model, known as a milestone of perturbed consistency learning, commonly serves as a standard and simple baseline. Inherently, learning from consistency can be regarded as learning from stability under perturbations. Recent improvement leans toward more complex consistency learning frameworks, yet, little attention is paid to the consistency target selection. Considering that the ambiguous regions from unlabeled data contain more informative complementary clues, in this paper, we improve the mean-teacher model to a novel ambiguity-consensus mean-teacher (AC-MT) model. Particularly, we comprehensively introduce and benchmark a family of plug-and-play strategies for ambiguous target selection from the perspectives of entropy, model uncertainty and label noise self-identification, respectively. Then, the estimated ambiguity map is incorporated into the consistency loss to encourage consensus between the two models' predictions in these informative regions. In essence, our AC-MT aims to find out the most worthwhile voxel-wise targets from the unlabeled data, and the model especially learns from the perturbed stability of these informative regions. The proposed methods are extensively evaluated on left atrium segmentation and brain tumor segmentation. Encouragingly, our strategies bring substantial improvement over recent state-of-the-art methods. The ablation study further demonstrates our hypothesis and shows impressive results under various extreme annotation conditions. • Experimentally identify a neglected-yet-critical problem, i.e., consistency target selection, that limits the performance of the classical mean-teacher semi-supervised segmentation model. • Propose a new perspective that encouraging perturbed stability in ambiguous-yet-informative regions can drive the model to learn more useful representations from unlabeled data. • Improve the classical mean-teacher model to a novel ambiguity-consensus mean-teacher (AC-MT) model, equipped with four plug-and-play strategies for ambiguous target selection. [ABSTRACT FROM AUTHOR]

One of the principal roles of a PhD examiner is to judge ‘both the potential of the researcher and the quality of the research’ ( Holbrook, Bourke, Fairbairn, & Lovat, 2014 , p. 986). While examiners may be guided by criteria supplied by universities, the descriptors they are provided with can often be open to interpretation. Interpreting an examiner's report can present a challenge to students and their supervisors, exacerbated by the often ambiguous use of language in the reports. This article examines the discourses of evaluation and instruction in 142 PhD examiners’ reports on theses submitted at an Australasian university. The paper draws on systemic functional linguistics, in particular transitivity ( Halliday & Matthiessen, 2014 ), in order to examine the reports. The study revealed that examiners can adopt up to 10 “roles” in their reports, each of which can be co-present in a single report. The inability to differentiate between these roles, we argue, is potentially frustrating for the audience of the reports (candidates, supervisors, departmental heads, etc.), particularly when interpreting whether a comment in the text represents an evaluation, an instruction, or an aside. By revealing these multiple, yet co-present, roles in examiners’ reports and their associated linguistic realisations, we hope to raise examiners’ awareness of the implications of the language they use when writing their reports as well as draw thesis supervisor and institutional attention to the ambiguities inherent in this underexplored genre. [ABSTRACT FROM AUTHOR]

Contra traditional Reichenbach (1947) analysis of tense and aspect, a mere temporal precedence relation does not license a simple past-pluperfect sequence in discourse, and the use of the pluperfect is constrained by further discourse-semantic rules. This paper delineates those discourse-semantic constraints based on a quantitative analysis of naturally occurring narrative corpus examples, and formalizes them in Discourse Representation Theory [DRT] ( Kamp and Reyle, 1993 ). We argue based on the corpus study that, contrary to the common assumption, the pluperfect does not invariably refer to a result or consequent state or ambiguous between aspectual and preterit meaning. Instead, the result state implication derives from the aspectual class of the main predicate. We further argue that the relationship between the anaphoric pluperfect sentence, which forms background and builds a side-structure of ancillary time line ( Hopper, 1979; Reinhart, 1984 ), and its antecedent simple past event sentence, which forms foreground and propels the main narrative time forward, can be best explained by the normative-modal background of the discourse. [ABSTRACT FROM AUTHOR]

Syntactic competence, especially the ability to use a wide range of sophisticated grammatical expressions, represents an important aspect of communicative acumen. This paper explores the question of how to best evaluate the syntactic competence of non-native speakers in an automated way. Using spoken responses of test takers participating in an English practice assessment, three classes of grammatical features – features based on n-grams of part-of-speech tags (POS), features based on various clause types, and features based on various phrases – are compared in an end-to-end assessment system. Feature correlations with human proficiency scores show that POS features and phrase features exhibit the highest correlations with human scores. Including these three classes of grammar features in a baseline scoring model that measures various aspects of spoken proficiency excluding aspects of grammar, we find substantial increases in agreement between machine and human scores. Finally, we discuss the broader implications of our results on the design of automatic scoring systems for spoken language. [ABSTRACT FROM AUTHOR]