Corpus-based Measures Research Articles

The probability of external causation: An empirical account of crosslinguistic variation in lexical causatives

AbstractThis corpus-driven computational study addresses the question of why some verbs in some languages participate in the causative alternation while their counterparts in other languages do not. The results of this study suggest that the lexical property that underlies this variation is the probability of external causation. Alternating verbs are distributed on a scale of increasing probability for an external causer to occur. The probability of external causation can be empirically assessed in two ways, among others: first, by observing the typological distribution of causative and anticausative morphological markings across a wide range of languages; second, through the frequency distribution of causative and anticausative uses of the alternating verbs in a corpus of a single language. Our study reveals that these two measures are correlated. Moreover, we demonstrate that the corpus-based measure is applicable to a wide range of verbs. Extending the corpus-based investigation comparatively across two languages, English and German, we find that the frequencies of crosslinguistic realizations of lexical causatives are modulated by the probability of external causation, an underlying parameter assigned to verb types. Finally, we propose a probabilistic graphical model that clusters verbs based on the relation between the crosslinguistic distribution of their causative and anticausative realizations and the probability of external causation.

SodhanaRef: a reference management software built using hybrid semantic measure

Reference management softwares are widely used by the researchers to maintain their collection of scholarly literature that exist in PDF format. Though widely used most of the reference management softwares have no sophisticated Information retrieval except few which offer advanced search that includes search for title, author etc., These softwares in the present day market do not give importance to the semantic similarity or relatedness concept, query expansion and finding the context within the query to find the concept behind the user mentioned query.With SodhanaRef, a solution is offered to deal with the above-mentioned issues by building reference management software using a mix up of corpus-based and knowledge-based semantic measures. Based on the evolution done on about 200 various scholarly literatures in the PDF form, SodhanaRef shows a good performance over Mendeley when compared between these two reference management softwares for title search. The other evaluations for finding the semantic similarity between the user mentioned query and the existing titles in the title search and for identifying the concept behind the query along with identifying the concept of a research publication have shown good results with an average precision between 0.8 to 1 for each query.

Semiautomatic Construction of Cross-Period Thesaurus

A cross-period (diachronic) thesaurus enables users to search for information using modern terminology and obtain semantically related terms from earlier historical periods. The complex task of supporting the construction of a diachronic thesaurus by a domain expert lexicographer has hardly been addressed computationally until now. In this article, we introduce a semiautomatic iterative Query Expansion (QE) scheme for supporting diachronic thesaurus construction, which identifies candidate related terms based on statistical corpus-based measures. We use ancient-modern period classification to increase the performance of the statistical cooccurrence measures and extend our methods to deal with Multi-Word Expressions (MWEs). We demonstrate the empirical benefit of our scheme for a Jewish cross-period thesaurus and evaluate its impact on recall and on the effectiveness of the lexicographer’s manual efforts.

Opposing effects of semantic diversity in lexical and semantic relatedness decisions.

Semantic ambiguity has often been divided into 2 forms: homonymy, referring to words with 2 unrelated interpretations (e.g., bark), and polysemy, referring to words associated with a number of varying but semantically linked uses (e.g., twist). Typically, polysemous words are thought of as having a fixed number of discrete definitions, or “senses,” with each use of the word corresponding to one of its senses. In this study, we investigated an alternative conception of polysemy, based on the idea that polysemous variation in meaning is a continuous, graded phenomenon that occurs as a function of contextual variation in word usage. We quantified this contextual variation using semantic diversity (SemD), a corpus-based measure of the degree to which a particular word is used in a diverse set of linguistic contexts. In line with other approaches to polysemy, we found a reaction time (RT) advantage for high SemD words in lexical decision, which occurred for words of both high and low imageability. When participants made semantic relatedness decisions to word pairs, however, responses were slower to high SemD pairs, irrespective of whether these were related or unrelated. Again, this result emerged irrespective of the imageability of the word. The latter result diverges from previous findings using homonyms, in which ambiguity effects have only been found for related word pairs. We argue that participants were slower to respond to high SemD words because their high contextual variability resulted in noisy, underspecified semantic representations that were more difficult to compare with one another. We demonstrated this principle in a connectionist computational model that was trained to activate distributed semantic representations from orthographic inputs. Greater variability in the orthography-to-semantic mappings of high SemD words resulted in a lower degree of similarity for related pairs of this type. At the same time, the representations of high SemD unrelated pairs were less distinct from one another. In addition, the model demonstrated more rapid semantic activation for high SemD words, thought to underpin the processing advantage in lexical decision. These results support the view that polysemous variation in word meaning can be conceptualized in terms of graded variation in distributed semantic representations.

Empréstimo lexical, conceptualização e variação: Para a abordagem sociocognitiva e socioletométrica dos estrangeirismos no português

This paper first advocates an onomasiological, concept-based and socio-cognitive approach to lexical borrowing, expanding the current loanword research from lexical items towards concepts. Second, it presents a corpusbased and concept-based sociolectometrical study on differences in the use of loanwords in European Portuguese and Brazilian Portuguese and their impact on diachronic lexical variation between the two national varieties. In the first part, the main topics and contributions of the Cognitive Sociolinguistic perspective on borrowability, and concept-based sociolectometrical methods of measuring variation in the success of loanwords are highlighted. In the second part, English and French loanwords in the field of football and clothing terminologies are analyzed through possible receptor Portuguese equivalents and advanced corpus-based sociolectometrical measures, such as featural measures (calculating the proportion of terms possessing a special feature) and uniformity measures (calculating onomasiological homogeneity and convergence/divergence between language varieties). These measures are based on onomasiological profiles, i.e. sets of alternative synonymous terms, together with their frequencies. As a development of our previous research on lexical convergence and divergence between European and Brazilian Portuguese (Soares da Silva 2010), the data include thousands of observations of the usage of alternative terms to refer to 43 football and clothing concepts. Corpus material was extracted from sports newspapers and fashion magazines from the 1950s, 1970s and 1990s/2000s, Internet chats related to football, and labels and price tags pictured from clothes shop windows. Football and clothing concepts confirm the hypothesis that the influence of foreign languages is stronger in the Brazilian variety than in the European variety. The use of loanwords has contributed towards onomasiological heterogeneity within and across the two national varieties in the last 60 years.

Words that go together

Although formulaic language has been studied extensively from both a linguistic and psycholinguistic perspective, little is known about the relationship between individual speakers’ knowledge of collocations and their linguistic experience, or between collocational knowledge and other aspects of linguistic knowledge. This is partly because work in these areas has been hampered by lack of an adequate instrument measuring speakers’ knowledge of collocations. This paper describes the development of such an instrument, the “Words that go together” (WGT) test, and some preliminary research using it. The instrument is a multiple choice test consisting of 40 items of varying frequency and collocation strength. The test was validated with a sample of 80 adult native speakers of English. Test-retest reliability was 0.80 and split-half reliability was 0.88. Convergent validity was established by comparing participants’ scores with measures expected to correlate with language experience (print exposure, education, and age) and other linguistic abilities (vocabulary size, grammatical comprehension); divergent validity was established by comparing test scores with nonverbal IQ. The results of the validation study are then used to compare speakers’ performance on the WGT with corpus-based measures of collocation strength (mutual information, z-score, t-score and simple frequency); however, no statistically reliable relationships were found.

Examining the N400 semantic context effect item-by-item: Relationship to corpus-based measures of word co-occurrence

With increasing availability of digital text, there has been an explosion of computational methods designed to turn patterns of word co-occurrence in large text corpora into numerical scores expressing the “semantic distance” between any two words. The success of such methods is typically evaluated by how well they predict human judgments of similarity. Here, I examine how well corpus-based methods predict amplitude of the N400 component of the event-related potential (ERP), an online measure of lexical processing in brain electrical activity. ERPs elicited by the second words of 303 word pairs were analyzed at the level of individual items. Three corpus-based measures (mutual information, distributional similarity, and latent semantic analysis) were compared to a traditional measure of free association strength. In a regression analysis, corpus-based and free association measures each explained some of the variance in N400 amplitude, suggesting that these may tap distinct aspects of word relationships. Lexical factors of concreteness of word meaning, word frequency, number of semantic associates, and orthographic similarity also explained variance in N400 amplitude at the single-item level.

Arabic Short Answer Scoring with Effective Feedback for Students

In this paper, we explore text similarity techniques for the task of automatic short answer scoring in Arabic language. We compare a number of string-based and corpus-based similarity measures, evaluate the effect of combining these measures, handle student’s answers holistically and partially, provide immediate useful feedback to student and also introduce a new benchmark Arabic data set that contains 50 questions and 600 student answers. Overall, the obtained correlation and error rate results prove that the presented system performs well enough for deployment in a real scoring environment. General Terms Natural Language Processing, Text Mining

In support of multiword unit classifications: Corpus and human rating data validate phraseological classifications of three different multiword unit types

AbstractMultiword units (MWUs) are word combinations which sit within the continuum of formulaic language. Many experimental studies have focused on the online processing of MWUs by native and non-native speakers, and the processing of idioms in particular. However, some studies use a mix of various MWU subtypes, while other studies have varying definitions for the same subtypes. For results from MWU studies to be useful to theories of language processing, storage and access, clearer classifications are needed for MWU subtypes. This study aims to empirically validate MWU categories as described by certain phraseologists in the European tradition. This will be done using MWUs from the British National Corpus, from across the continuum of frequent to infrequent occurrence and co-occurrence. Hence, in this paper I will describe the empirical findings that may validate the classifications for MWU categories of restricted collocations, idioms, and lexical bundles, using corpus-based measures and human ratings.

Short Answer Grading Using String Similarity And Corpus-Based Similarity

Most automatic scoring systems use pattern based that requires a lot of hard and tedious work. These systems work in a supervised manner where predefined patterns and scoring rules are generated. This paper presents a different unsupervised approach which deals with students’ answers holistically using text to text similarity. Different String-based and Corpus-based similarity measures were tested separately and then combined to achieve a maximum correlation value of 0.504. The achieved correlation is the best value achieved for unsupervised approach Bag of Words (BOW) when compared to previous work.

The Meaning of Lexical Causatives in Cross-Linguistic Variation

The causative alternation has been recognised in the linguistic literature as one of the most widely spread linguistic phenomena, attested in almost all languages, although differently realised and involving partially different sets of verbs. In this paper, we identify the degree of spontaneity of the event described by a verb as a general component of meaning of alternating verbs which underlies the within-language and cross-linguistic variation in their realisations. We first establish that a corpus-based measure of this property, the ratio of the frequency of usage of the causative and anticausative form, is strongly correlated to an independent typological measure (Haspelmath, 1993). Then we examine the influence of this property on the cross-linguistic realisations of verbs. We find that the degree of variation and parallelism in forms across languages is strongly related to the degree of spontaneity of the verb.

Comparable Evaluation of Contemporary Corpus-Based and Knowledge-Based Semantic Similarity Measures of Short Texts

This paper presents methods for measuring the semantic similarity of texts, where we evaluated different approaches based on existing similarity measures. On one side word similarity was calculated by processing large text corpuses and on the other, commonsense knowledgebase was used. Given that a large fraction of the information available today, on the Web and elsewhere, consists of short text snippets (e.g. abstracts of scientific documents, image captions or product descriptions), where commonsense knowledge has an important role, in this paper we focus on computing the similarity between two sentences or two short paragraphs by extending existing measures with information from the ConceptNet knowledgebase. On the other hand, an extensive research has been done in the field of corpus-based semantic similarity, so we also evaluated existing solutions by imposing some modifications. Through experiments performed on a paraphrase data set, we demonstrate that some of proposed approaches can improve the semantic similarity measurement of short text.

Method and System for Assessing Pronunciation Difficulties of Non-Native Speakers by Entropy Calculation

The present disclosure presents a useful metric for assessing the relative difficulty which non-native speakers face in pronouncing a given utterance and a method and systems for using such a metric in the evaluation and assessment of the utterances of non-native speakers. In an embodiment, the metric may be based on both known sources of difficulty for language learners and a corpus-based measure of cross-language sound differences. The method may be applied to speakers who primarily speak a first language speaking utterances in any non-native second language.

Corpus-Based Frequency Profiling: Migration To A Word List Based On The British National Corpus

The selection and assessment of ELT materials involve multiple criteria. The use of frequency word lists to profile the vocabulary makeup of a text is one such criterion. It provides a quantifiable characterization and classification of lexical material in terms of corpus-based frequency measures. The process of vocabulary profiling is not without challenges, first among which is the identification of a word list adequate for ELT. The choice will determine the amount of information, if any, that can be derived from a text. This paper provides an appraisal of a frequency word list based on the British National Corpus (BNC) and shows the benefits that can be gained by profiling with this list rather than with the long-established General Service List (West, 1953).

Semantic text similarity using corpus-based word similarity and string similarity

We present a method for measuring the semantic similarity of texts using a corpus-based measure of semantic word similarity and a normalized and modified version of the Longest Common Subsequence (LCS) string matching algorithm. Existing methods for computing text similarity have focused mainly on either large documents or individual words. We focus on computing the similarity between two sentences or two short paragraphs. The proposed method can be exploited in a variety of applications involving textual knowledge representation and knowledge discovery. Evaluation results on two different data sets show that our method outperforms several competing methods.

Automatically learning semantic knowledge about multiword predicates

Highly frequent and highly polysemous verbs, such as give, take, and make, pose a challenge to automatic lexical acquisition methods. These verbs widely participate in multiword predicates (such as light verb constructions, or LVCs), in which they contribute a broad range of figurative meanings that must be recognized. Here we focus on two properties that are key to the computational treatment of LVCs. First, we consider the degree of figurativeness of the semantic contribution of such a verb to the various LVCs it participates in. Second, we explore the patterns of acceptability of LVCs, and their productivity over semantically related combinations. To assess these properties, we develop statistical measures of figurativeness and acceptability that draw on linguistic properties of LVCs. We demonstrate that these corpus-based measures correlate well with human judgments of the relevant property. We also use the acceptability measure to estimate the degree to which a semantic class of nouns can productively form LVCs with a given verb. The linguistically-motivated measures outperform a standard measure for capturing the strength of collocation of these multiword expressions.

Contextual Diversity, Not Word Frequency, Determines Word-Naming and Lexical Decision Times

Word frequency is an important predictor of word-naming and lexical decision times. It is, however, confounded with contextual diversity, the number of contexts in which a word has been seen. In a study using a normative, corpus-based measure of contextual diversity, word-frequency effects were eliminated when effects of contextual diversity were taken into account (but not vice versa) across three naming and three lexical decision data sets; the same pattern of results was obtained regardless of which of three corpora was used to derive the frequency and contextual-diversity values. The results are incompatible with existing models of visual word recognition, which attribute frequency effects directly to frequency, and are particularly problematic for accounts in which frequency effects reflect learning. We argue that the results reflect the importance of likely need in memory processes, and that the continuity between reading and memory suggests using principles from memory research to inform theories of reading.

Models of high-dimensional semantic space predict language-mediated eye movements in the visual world

In the visual world paradigm, participants are more likely to fixate a visual referent that has some semantic relationship with a heard word, than they are to fixate an unrelated referent [Cooper, R. M. (1974). The control of eye fixation by the meaning of spoken language. A new methodology for the real-time investigation of speech perception, memory, and language processing. Cognitive Psychology, 6, 813–839]. Here, this method is used to examine the psychological validity of models of high-dimensional semantic space. The data strongly suggest that these corpus-based measures of word semantics predict fixation behavior in the visual world and provide further evidence that language-mediated eye movements to objects in the concurrent visual environment are driven by semantic similarity rather than all-or-none categorical knowledge. The data suggest that the visual world paradigm can, together with other methodologies, converge on the evidence that may help adjudicate between different theoretical accounts of the psychological semantics.

A corpus-based analysis of Mandarin nominal root compound

Mandarin has many compound-like expressions which seem to be derived directly from (non-word) roots. For example, yi, the root of the word ma yi ‘ant’, occurs in many derived words, including gong yi ‘worker ant’. It seems reasonable to analyze these constructions as compounds, yet traditional theories of compounding as well as recent work on Mandarin morphology (Dai (1992)) have assumed that true compounding is word-based. Dai further implies that only true compounding is productive and that what we shall term ‘root compounding’ is unproductive, thus apparently strengthening the distinction between the two types. We show, using a corpus-based measure of morphological productivity due to Baayen (1989), that many nominal roots in fact productively form root compounds; thus, at least on the basis of productivity, there is no compelling reason to consider word and root compounding as distinct morphological processes. The question arises of why root compounds (as opposed to word compounds) are formed. As a partial answer we provide a preliminary analysis of the semantics of nominal roots, arguing that they represent kinds (Carlson (1977)); root compounds are formed when the ‘kind’ interpretation is appropriate, though stylistic issues are argued also to be involved. We also consider a prosodic constraint on (root) compounding proposed by Lu and Duanmu (1991) and show that this constraint can be overridden when semantic or stylistic considerations demand it.