Morphosyntactic Information Research Articles

Grammatical number processing and anticipatory eye movements are not tightly coordinated in English spoken language comprehension.

Recent studies of eye movements in world-situated language comprehension have demonstrated that rapid processing of morphosyntactic information – e.g., grammatical gender and number marking – can produce anticipatory eye movements to referents in the visual scene. We investigated how type of morphosyntactic information and the goals of language users in comprehension affected eye movements, focusing on the processing of grammatical number morphology in English-speaking adults. Participants’ eye movements were recorded as they listened to simple English declarative (There are the lions.) and interrogative (Where are the lions?) sentences. In Experiment 1, no differences were observed in speed to fixate target referents when grammatical number information was informative relative to when it was not. The same result was obtained in a speeded task (Experiment 2) and in a task using mixed sentence types (Experiment 3). We conclude that grammatical number processing in English and eye movements to potential referents are not tightly coordinated. These results suggest limits on the role of predictive eye movements in concurrent linguistic and scene processing. We discuss how these results can inform and constrain predictive approaches to language processing.

Do grammatical-gender distinctions learned in the second language influence native-language lexical processing?

How does learning a second language influence native language processing? In the present study, we examined whether knowledge of Spanish - a language that marks grammatical gender on inanimate nouns - influences lexical processing in English - a language that does not mark grammatical gender. We tested three groups of adult English native speakers: monolinguals, emergent bilinguals with high exposure to Spanish, and emergent bilinguals with low exposure to Spanish. Participants engaged in an associative learning task in English where they learned to associate names of inanimate objects with proper names. For half of the pairs, the grammatical gender of the noun's Spanish translation matched the gender of the proper name (e.g., corn-Patrick). For half of the pairs, the grammatical gender of the noun's Spanish translation mismatched the gender of the proper noun (e.g., beach-William). High-Spanish-exposure bilinguals (but not monolinguals or low-Spanish-exposure bilinguals) were less accurate at retrieving proper names for gender-incongruent than for gender-congruent pairs. This indicates that second-language morphosyntactic information is activated during native-language processing, even when the second language is acquired later in life.

Extraction of relations between genes and diseases from text and large-scale data analysis: implications for translational research.

BackgroundCurrent biomedical research needs to leverage and exploit the large amount of information reported in scientific publications. Automated text mining approaches, in particular those aimed at finding relationships between entities, are key for identification of actionable knowledge from free text repositories. We present the BeFree system aimed at identifying relationships between biomedical entities with a special focus on genes and their associated diseases.ResultsBy exploiting morpho-syntactic information of the text, BeFree is able to identify gene-disease, drug-disease and drug-target associations with state-of-the-art performance. The application of BeFree to real-case scenarios shows its effectiveness in extracting information relevant for translational research. We show the value of the gene-disease associations extracted by BeFree through a number of analyses and integration with other data sources. BeFree succeeds in identifying genes associated to a major cause of morbidity worldwide, depression, which are not present in other public resources. Moreover, large-scale extraction and analysis of gene-disease associations, and integration with current biomedical knowledge, provided interesting insights on the kind of information that can be found in the literature, and raised challenges regarding data prioritization and curation. We found that only a small proportion of the gene-disease associations discovered by using BeFree is collected in expert-curated databases. Thus, there is a pressing need to find alternative strategies to manual curation, in order to review, prioritize and curate text-mining data and incorporate it into domain-specific databases. We present our strategy for data prioritization and discuss its implications for supporting biomedical research and applications.ConclusionsBeFree is a novel text mining system that performs competitively for the identification of gene-disease, drug-disease and drug-target associations. Our analyses show that mining only a small fraction of MEDLINE results in a large dataset of gene-disease associations, and only a small proportion of this dataset is actually recorded in curated resources (2%), raising several issues on data prioritization and curation. We propose that joint analysis of text mined data with data curated by experts appears as a suitable approach to both assess data quality and highlight novel and interesting information.Electronic supplementary materialThe online version of this article (doi:10.1186/s12859-015-0472-9) contains supplementary material, which is available to authorized users.

The differential time course for consonant and vowel processing in Arabic: implications for language learning and rehabilitation.

Educators and therapists in the Arab world have not been able to benefit from the recent integration of basic behavioral science with neuroscience. This is due to the paucity of basic research on Arabic. The present study is a step toward establishing the necessary structure for the emergence of neuro-rehabilitory and educational practices. It focuses on the recent claim that consonants and vowels have distinct representations, carry different kinds of information, and engage different processing mechanisms. This proposal has received support from various research fields, however it suprisingly stops short of making any claims about the time course of consonant and vowel processing in speech. This study specifically asks if consonants and vowels are processed differentially over time, and whether these time courses vary depending on the kind of information they are associated with. It does so in the context of a Semitic language, Arabic, where consonants typically convey semantic meaning in the form of tri-consonantal roots, and vowels carry phonological and morpho-syntactic information in the form of word patterns. Two cross-modal priming experiments evaluated priming by fragments of consonants that belong to the root, and fragments of vowels belonging to the word pattern. Consonant fragments were effective primes while vowel fragments were not. This demonstrates the existence of a differential processing time course for consonants and vowels in the auditory domain, reflecting in part the different linguistic functions they are associated with, and argues for the importance of assigning distinct representational and processing properties to these elements. At broader theoretical and practical levels, the present results provide a significant building block for the emergence of neuro-rehabilitory and neuro-educational traditions for Arabic.

The use of case marking for predictive processing in second language Japanese

Research on processing in English has shown that verb information facilitates predictive processing. Because Japanese verbs occur at the ends of clauses, this information cannot be used to predict the roles of preceding nominals. Kamide, Altmann and Haywood (2003) showed that native Japanese speakers use case markers to predict forthcoming linguistic items. In the present study, we investigated whether second language learners of Japanese demonstrate such predictive effects when processing sentences containing either the monotransitive or ditransitive constructions. A visual-world paradigm experiment showed that, although native speakers generated predictions for syntactic outcomes, the learners did not. These findings underscore the usefulness of morphosyntactic information in processing Japanese and indicate that learners fail to make full use of case markers to generate expectations regarding syntactic outcomes during online processing. Learners may rely on nonlinguistic information to compensate for this deficit in syntactic processing.

Semantics and morphosyntax in predictive L2 sentence processing

AbstractIn the context of current approaches to anticipation in native and non-native sentence processing, this paper investigates whether late second-language (L2) learners integrate morphosyntax, i.e. case marking, and verb semantics to generate anticipations in L2 sentence comprehension. In a visual-world eye-tracking experiment with 45 L1 English L2 learners of German and 12 German natives, German natives are found to integrate morphosyntactic and lexical-semantic information in anticipatory processing, while L2 learners only rely on lexical-semantic information for prediction. Moreover, there is no indication that increasing proficiency leads to the involvement of morphosyntax in predictive L2 processing. We discuss reasons for the lower sensitivity to morphosyntax in anticipatory L2 sentence processing.

Assessing document and sentence readability in less resourced languages and across textual genres

In this paper, we tackle three underresearched issues of the automatic readability assessment literature, namely the evaluation of text readability in less resourced languages, with respect to sentences (as opposed to documents) as well as across textual genres. Different solutions to these issues have been tested by using and refining READ‑IT, the first advanced readability assessment tool for Italian, which combines traditional raw text features with lexical, morpho-syntactic and syntactic information. In READ‑IT readability assessment is carried out with respect to both documents and sentences, with the latter constituting an important novelty of the proposed approach: READ‑IT shows a high accuracy in the document classification task and promising results in the sentence classification scenario. By comparing the results of two versions of READ‑IT, adopting a classification‑ versus ranking-based approach, we also show that readability assessment is strongly influenced by textual genre; for this reason a genre-oriented notion of readability is needed. With classification-based approaches, reliable results can only be achieved with genre-specific models: Since this is far from being a workable solution, especially for less resourced languages, a new ranking method for readability assessment is proposed, based on the notion of distance.

MORPHOSYNTAX IN THE BILINGUAL MENTAL LEXICON

Although morphosyntax has been identified as a major source of difficulty for adult (nonnative) language learners, most previous studies have examined a limited set of largely affix-based phenomena. Little is known about word-based morphosyntax in late bilinguals and of how morphosyntax is represented and processed in a nonnative speaker’s lexicon. To address these questions, we report results from two behavioral experiments investigating stem variants of strong verbs in German (which encode features such as tense, person, and number) in groups of advanced adult learners as well as native speakers of German. Although the late bilinguals were highly proficient in German, the results of a lexical priming experiment revealed clear native-nonnative differences. We argue that lexical representation and processing relies less on morphosyntactic information in a nonnative than in a native language.

An Arabic CCG approach for determining constituent types from Arabic Treebank

Converting a treebank into a CCGbank opens the respective language to the sophisticated tools developed for Combinatory Categorial Grammar (CCG) and enriches cross-linguistic development. The conversion is primarily a three-step process: determining constituents’ types, binarization, and category conversion. Usually, this process involves a preprocessing step to the Treebank of choice for correcting brackets and normalizing tags for any changes that were introduced during the manual annotation, as well as extracting morpho-syntactic information that is necessary for determining constituents’ types. In this article, we describe the required preprocessing step on the Arabic Treebank, as well as how to determine Arabic constituents’ types. We conducted an experiment on parts 1 and 2 of the Penn Arabic Treebank (PATB) aimed at converting the PATB into an Arabic CCGbank. The performance of our algorithm when applied to ATB1v2.0 & ATB2v2.0 was 99% identification of head nodes and 100% coverage over the Treebank data.

Morphosyntax can modulate the N400 component: Event related potentials to gender-marked post-nominal adjectives

Event-related potential studies of grammatical gender agreement often report a left anterior negativity (LAN) when agreement violations occur. Some studies have shown that during sentence comprehension gender violations can also interact with semantic processing to modulate a negativity associated with processing meaning — the N400. Given that the LAN and N400 overlap in time, they are identified by their scalp distributions and purported functional roles. Critically, grammatical gender violations also elicit a right posterior positivity that can overlap temporally and potentially affect the scalp distribution of the LAN/N400. We measured the effect of grammatical gender violations in the LAN/N400 window and late positive component (LPC) during comprehension of Spanish sentences. A post-nominal adjective could either make sense or not, and either agree or disagree in gender with the preceding noun. We observed a negativity to gender agreement violations in the LAN/N400 window (300–500ms post stimulus onset) that was smaller than the semantic-congruity N400, but overlapped with it in time and distribution. The early portion of the LPC to gender violations was modulated by sentence constraint, occurring as early as 450ms in highly constraining sentences. A subadditive interaction occurred at the later portion of the LPC with equivalent effects for single and double violations (gender and semantics), reflecting a general stage of reprocessing. Overall, our data support models of language comprehension whereby both semantic and morphosyntactic information can affect processing at similar time points.

On modular approaches to grammar: Evidence from Polish

AbstractModularity of grammar has been explicitly or tacitly assumed in many generative analyses. Modules are separate computational systems that perform specific tasks and make use of domain-specific information. It is argued that the concept is difficult to maintain in the light of evidence from Polish. I look at palatalization effects before vowels and conclude that phonological regularities must have access to morphosyntactic information. In addition, certain regularities in the selection of diminutive allomorphs suggest that morphology must have access to phonetic information. As domain specificity, the core concept of modular approaches, is compromised, modularity does not seem a likely candidate for a universal property of grammar.

A New Approach to Tagging in Indian Languages

In this paper, we present a new approach to automatic tag- ging without requiring any machine learning algorithm or training data. We argue that the critical information required for tagging comes more from word internal structure than from the context and we show how a well designed morphological analyzer can assign correct tags and disam- biguate many cases of tag ambiguities too. The crux of the approach is in the very denition of words. While others simply tokenize a given sen- tence based on spaces and take these tokens to be words, we argue that words need to be motivated from semantic and syntactic considerations, not orthographic conventions. We have worked on Telugu and Kannada languages and in this paper, we take the example of Telugu language and show how high quality tagging can be achieved with a ne grained, hierarchical tag set, carrying not only morpho-syntactic information but also some aspects of lexical and semantic information that is necessary or useful for syntactic parsing. In fact entire corpora can be tagged very fast and with a good degree of guarantee of quality. We give details of our experiments and results obtained. We believe our approach can also be applied to other languages.

Izgradnja modelov za prepoznavanje imenskih entitet za hrvaščino in slovenščino

Prispevek predstavlja razvoj prosto dostopnih modelov za prepoznavanje in klasifikacijo imenskih enot za hrvaški in slovenski jezik. Poskusi se osredotočajo na najbolj informativne jezikovne lastnosti, pri čemer upoštevajo dostopnost jezikovnih orodij za oba jezika. Poleg standardnih jezikovnih lastnosti so upoštevane tudi distribucijske lastnosti, ki so bile izračunane iz velikih neoznačenih enojezičnih korpusov. Uporaba distribucijskih lastnosti izboljša rezultate za 7-8 točk v meri F1, uporaba oblikoslovnih informacij pa dodatno za 3-4 točke, in to pri obeh jezikih. Najboljši naučeni model skupaj s testno množico za primerjavo z obstoječimi in bodočimi sistemi ter model za oblikoslovno označevanje hrvaščine s programom HunPos so dostopni za prenos za uporabo v znanstvene in komercialne namene.

Suffix -mente Adverbs in DAELE, A Spanish Learners' Dictionary

This paper looks at how Spanish –mente adverbs are shown in DAELE, an electronic dictionary for advanced-level students of Spanish, currently being developed at the Universitat Pompeu Fabra (Barcelona). Since a learners’ dictionary is a tool conceived not only for decoding but also for production, information about use is deemed highly relevant in this dictionary. That is why new ways to provide adverb-production information are explored. Our discussion considers aspects of both the macrostructure and the microstructure. In terms of the macrostructure, the main problem arising from the lexicographical representation of adverbs ending in –mente is to decide which adverbs are to be included in the dictionary. In the Spanish lexicographical tradition there are several answers to this question, which range from including only a few adverbs to including all documented adverbs. In terms of the microstructure, it is important to provide all the semantic, morphosyntactic and pragmatic information on how to use an adverb, by means of a set of definition formulae which are both user friendly and informative.

Animacy information outweighs morphological cues in Russian

The present study examined the robustness of morphosyntactic information (inflectional case marking) vs. semantic information (animacy) in non-canonical SOV-and-V structures in Russian, a morphologically rich language with relatively free word order. Results from a self-paced moving window paradigm followed by a comprehension question indicated that case marking was a relatively weak cue during online reading and offline comprehension compared to animacy, which was relied upon heavily throughout. The results suggest that even in languages with rich inflectional morphology, morphosyntax is fragile. Furthermore, the results provide support for memory-based models of sentence processing in which similarity-induced interference affects memory for previously encountered constituents and the integration of incoming material.

Integrating meaning and structure in L1–L2 and L2–L1 translations

This article examined the integration of semantic and morphosyntactic information by Korean learners of English as a second language (L2). In Experiment 1, L2 learners listened to English active or passive sentences that were either plausible or implausible and translated them into Korean. A significant number of Korean translations maintained the original passive/active structure, but switched the thematic roles of the actors in the sentences. In Experiment 2, the direction of translation was reversed and participants made very few translation errors, showing that the errors in Experiment 1 were not due to participants’ lack of control over the English passive morphosyntax. The results are strikingly similar to previous results in the first language (L1) psycholinguistics literature, and support a view of L2 processing (like L1 processing) that is ‘good enough’ in nature: misinterpretations arise from only a ‘good enough’ integration of semantic and morphosyntactic information in the input.

Individual differences in the second language processing of object–subject ambiguities

ABSTRACTThis study investigates whether and how individual differences modulate the adult second language (L2) processing of syntactic ambiguities. In a linear mixed regression analysis, we test how proficiency, working memory, reading speed, automaticity in lexical access, and grammatical integration ability affect the resolution of temporary object–subject ambiguities in L2 English. The results from 75 first language German advanced learners attest that individual differences in syntactic integration ability modulate the reliance on morphosyntactic and plausibility information. Similar to native speakers, L2 learners are found to adopt two different routes in L2 processing. The findings highlight the role of individual differences and qualify previous generalizations about the relative use of morphosyntactic and other types of information in L2 processing.

Syntactic computation in the human brain: the degree of merger as a key factor.

Our goal of this study is to characterize the functions of language areas in most precise terms. Previous neuroimaging studies have reported that more complex sentences elicit larger activations in the left inferior frontal gyrus (L. F3op/F3t), although the most critical factor still remains to be identified. We hypothesize that pseudowords with grammatical particles and morphosyntactic information alone impose a construction of syntactic structures, just like normal sentences, and that “the Degree of Merger” (DoM) in recursively merged sentences parametrically modulates neural activations. Using jabberwocky sentences with distinct constructions, we fitted various parametric models of syntactic, other linguistic, and nonlinguistic factors to activations measured with functional magnetic resonance imaging. We demonstrated that the models of DoM and “DoM+number of Search (searching syntactic features)” were the best to explain activations in the L. F3op/F3t and supramarginal gyrus (L. SMG), respectively. We further introduced letter strings, which had neither lexical associations nor grammatical particles, but retained both matching orders and symbol orders of sentences. By directly contrasting jabberwocky sentences with letter strings, localized activations in L. F3op/F3t and L. SMG were indeed independent of matching orders and symbol orders. Moreover, by using dynamic causal modeling, we found that the model with a inhibitory modulatory effect for the bottom-up connectivity from L. SMG to L. F3op/F3t was the best one. For this best model, the top-down connection from L. F3op/F3t to L. SMG was significantly positive. By using diffusion-tensor imaging, we confirmed that the left dorsal pathway of the superior longitudinal and arcuate fasciculi consistently connected these regions. Lastly, we established that nonlinguistic order-related and error-related factors significantly activated the right (R.) lateral premotor cortex and R. F3op/F3t, respectively. These results indicate that the identified network of L. F3op/F3t and L. SMG subserves the calculation of DoM in recursively merged sentences.

Using the crowd for readability prediction

AbstractWhile human annotation is crucial for many natural language processing tasks, it is often very expensive and time-consuming. Inspired by previous work on crowdsourcing, we investigate the viability of using non-expert labels instead of gold standard annotations from experts for a machine learning approach to automatic readability prediction. In order to do so, we evaluate two different methodologies to assess the readability of a wide variety of text material: A more traditional setup in which expert readers make readability judgments and a crowdsourcing setup for users who are not necessarily experts. To this purpose two assessment tools were implemented: a tool where expert readers can rank a batch of texts based on readability, and a lightweight crowdsourcing tool, which invites users to provide pairwise comparisons. To validate this approach, readability assessments for a corpus of written Dutch generic texts were gathered. By collecting multiple assessments per text, we explicitly wanted to level out readers' background knowledge and attitude. Our findings show that the assessments collected through both methodologies are highly consistent and that crowdsourcing is a viable alternative to expert labeling. This is a good news as crowdsourcing is more lightweight to use and can have access to a much wider audience of potential annotators. By performing a set of basic machine learning experiments using a feature set that mainly encodes basic lexical and morpho-syntactic information, we further illustrate how the collected data can be used to perform text comparisons or to assign an absolute readability score to an individual text. We do not focus on optimising the algorithms to achieve the best possible results for the learning tasks, but carry them out to illustrate the various possibilities of our data sets. The results on different data sets, however, show that our system outperforms the readability formulas and a baseline language modelling approach. We conclude that readability assessment by comparing texts is a polyvalent methodology, which can be adapted to specific domains and target audiences if required.

The Authorship of the Disputed Federalist Papers with an Annotated Corpus

The present paper discusses the potentials of annotated corpora in authorship attribution studies. These corpora allow the retrieval of morpho-syntactic information, which may serve as stylistic features, getting beyond the most common word-based approaches. In itself, annotation is meant to contain the structural mark-up, the lemmatization, the part-of-speech mark-up, and the textual reference. For the purpose, the corpus of the Federalist Papers is tested to explore Burrows's Delta from a twofold perspective, lemma-based Delta and POS-based Delta. The experiments come to demonstrate the validity of these approaches for authorial attribution with the finding of empirical evidence to assign, in the majority of cases, the hand responsible for the disputed Federalist Papers, but for paper 55, traditionally considered less Madisonian in the relevant literature.