Multilingual lexical resources to detect cognates in non-aligned texts

The identification of cognates between two distinct languages has recently start- ed to attract the attention of NLP re- search, but there has been little research into using semantic evidence to detect cognates. The approach presented in this paper aims to detect English-French cog- nates within monolingual texts (texts that are not accompanied by aligned translat- ed equivalents), by integrating word shape similarity approaches with word sense disambiguation techniques in order to account for context. Our implementa- tion is based on BabelNet, a semantic network that incorporates a multilingual encyclopedic dictionary. Our approach is evaluated on two manually annotated da- tasets. The first one shows that across different types of natural text, our method can identify the cognates with an overall accuracy of 80%. The second one, con- sisting of control sentences with semi- cognates acting as either true cognates or false friends, shows that our method can identify 80% of semi-cognates acting as cognates but also identifies 75% of the semi-cognates acting as false friends.

Assessing multilingual lexical incorporation hypotheses through a primed picture-naming task

The incorporation of new representations into the mental lexicon has raised numerous questions about the organisational principles that govern the process. A number of studies have argued that similarity between the new L3 items and existing representations in the L1 and L2 is the main incorporating force (Hall & Ecke, 2003; Herwig, 2001). Experimental evidence obtained through a primed picture-naming task with L1 Polish-L2 English learners of L3 Russian supports Hall and Ecke’s Parasitic Model of L3 vocabulary acquisition, displaying a significant main effect for both priming and proficiency. These results complement current models of vocabulary acquisition and lexical access in multilingual speakers.

Cross-lingual ontology matching as a challenge for the Multilingual Semantic Web

Recently, the Semantic Web has experienced significant advancements in standards and techniques, as well as in the amount of semantic information available online. Nevertheless, mechanisms are still needed to automatically reconcile information when it is expressed in different natural languages on the Web of Data, in order to improve the access to semantic information across language barriers. In this context several challenges arise [1], such as: (i) ontology translation/localization, (ii) cross-lingual ontology mappings, (iii) representation of multilingual lexical information, and (iv) cross-lingual access and querying of linked data. In the following we will focus on the second challenge, which is the necessity of establishing, representing and storing cross-lingual links among semantic information on the Web. In fact, in a “truly” multilingual Semantic Web, semantic data with lexical representations in one natural language would be mapped to equivalent or related information in other languages, thus making navigation across multilingual information possible for software agents.

Developing and testing readability measurements for second language learners

This research constructed a readability measurement for French speakers who view English as a second language. It identified the true cognates, which are the similar words from these two languages, as an indicator of the difficulty of an English text for French people. A multilingual lexical resource is used to detect true cognates in text, and Statistical Language Modelling to predict the predict the readability level. The proposed enhanced statistical language model is making a step in the right direction by improving the accuracy of readability predictions for French speakers by up to 10% compared to state of the art approaches. The outcome of this study could accelerate the learning process for French speakers who are studying English. More importantly, this study also benefits the readability estimation research community, presenting an approach and evaluation at sentence level as well as innovating with the use of cognates as a new text feature.

Localização de ontologias - elementos para a construção de uma abordagem

O desenvolvimento de recursos multilingues robustos para fazer face às exigências crescentes na complexidade dos processos intra e inter-organizacionais é um processo complexo que obriga a um aumento da qualidade nos modos de interacção e partilha dos recursos das organizações, através, por exemplo, de um maior envolvimento dos diferentes interlocutores em formas eficazes e inovadoras de colaboração. É um processo em que se identificam vários problemas e dificuldades, como sendo, no caso da criação de bases de dados lexicais multilingues, o desenvolvimento de uma arquitectura capaz de dar resposta a um conjunto vasto de questões linguísticas, como a polissemia, os padrões lexicais ou os equivalentes de tradução. Estas questões colocam-se na construção quer dos recursos terminológicos, quer de ontologias multilingues. No caso da construção de uma ontologia em diferentes línguas, processo no qual focalizaremos a nossa atenção, as questões e a complexidade aumentam, dado o tipo e propósitos do artefacto semântico, os elementos a localizar (conceitos e relações conceptuais) e o contexto em que o processo de localização ocorre. Pretendemos, assim, com este artigo, analisar o conceito e o processo de localização no contexto dos sistemas de gestão do conhecimento baseados em ontologias, tendo em atenção o papel central da terminologia no processo de localização, as diferentes abordagens e modelos propostos, bem como as ferramentas de base linguística que apoiam a implementação do processo. Procuraremos, finalmente, estabelecer alguns paralelismos entre o processo tradicional de localização e o processo de localização de ontologias, para melhor o situar e definir.

Delimitação e alinhamento de conceitos lexicalizados no inglês norte-americano e no português brasileiro

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

A interlíngua e o alinhamento léxico-conceitual da base de dados bilíngue REBECA

In the architecture of a natural language processing system based on linguistic knowledge, two types of component are important: the knowledge databases and the processing modules. One of the knowledge databases is the lexical database, which is responsible for providing the lexical unities and its properties to the processing modules. The systems that process two or more languages require bilingual and/or multilingual lexical databases. These databases can be constructed by aligning distinct monolingual databases. In this paper, we present the interlingua and the strategy of aligning the two monolingual databases in REBECA, which only stores concepts from the “wheeled vehicle” domain.

Production and comprehension of English and Hindi in multilingual transcortical aphasia

This study investigates the two later-acquired but proficient languages, English and Hindi, of two multilingual individuals with transcortical aphasia (right basal ganglia lesion in GN and brain stem lesion in GS). Dissociation between lexical and syntactic profiles in both the languages with a uniform performance across the languages at the lexical level and an uneven performance across the languages at the syntactic level was observed. Their performances are discussed in relation to the implicit/explicit language processes (Paradis, 1994 and Paradis, 2004) and the declarative/procedural model (Ullman, 2001b and Ullman, 2005) of bilingual language processing. Additionally, their syntactic performance is interpreted in relation to the salient grammatical contrasts between English and Hindi.

English compound and non-compound processing in bilingual and multilingual speakers: effects of dominance and sequential multilingualism

This article reports on a study investigating the relative influence of the first and dominant language on L2 and L3 morpho-lexical processing. A lexical decision task compared the responses to English NV-er compounds (e.g., taxi driver) and non-compounds provided by a group of native speakers and three groups of learners at various levels of English proficiency: L1 Spanish-L2 English sequential bilinguals and two groups of early Spanish-Basque bilinguals with English as their L3. Crucially, the two trilingual groups differed in their first and dominant language (i.e., L1 Spanish-L2 Basque vs. L1 Basque-L2 Spanish). Our materials exploit an (a)symmetry between these languages: while Basque and English pattern together in the basic structure of (productive) NV-er compounds, Spanish presents a construction that differs in directionality as well as inflection of the verbal element (V[3SG] + N). Results show between and within group differences in accuracy and response times that may be ascribable to two factors besides proficiency: the number of languages spoken by a given participant and their dominant language. An examination of response bias reveals an influence of the participants' first and dominant language on the processing of NV-er compounds. Our data suggest that morphological information in the nonnative lexicon may extend beyond morphemic structure and that, similarly to bilingualism, there are costs to sequential multilingualism in lexical retrieval.

The Automatic Mapping of Princeton WordNet Lexical-Conceptual Relations onto the Brazilian Portuguese WordNet Database

Princeton WordNet (WN.Pr) lexical database has motivated efficient compilations of bulky relational lexicons since its inception in the 1980's. The EuroWordNet project, the first multilingual initiative built upon WN.Pr, opened up ways of building individual wordnets, and interrelating them by means of the so-called Inter-Lingual-Index, an unstructured list of the WN.Pr synsets. Other important initiative, relying on a slightly different method of building multilingual wordnets, is the MultiWordNet project, where the key strategy is building language specific wordnets keeping as much as possible of the semantic relations available in the WN.Pr. This paper, in particular, stresses that the additional advantage of using WN.Pr lexical database as a resource for building wordnets for other languages is to explore possibilities of implementing an automatic procedure to map the WN.Pr conceptual relations as hyponymy, co-hyponymy, troponymy, meronymy, cause, and entailment onto the lexical database of the wordnet under construction, a viable possibility, for those are language-independent relations that hold between lexicalized concepts, not between lexical units. Accordingly, combining methods from both initiatives, this paper presents the ongoing implementation of the WN.Br lexical database and the aforementioned automation procedure illustrated with a sample of the automatic encoding of the hyponymy and co-hyponymy relations.

Annotating the meaning of discourse connectives in multilingual corpora

Discourse connectives are lexical items indicating coherence relations between discourse segments. Even though many languages possess a whole range of connectives, important divergences exist cross-linguistically in the number of connectives that are used to express a given relation. For this reason, connectives are not easily paired with a univocal translation equivalent across languages. This paper is a first attempt to design a reliable method to annotate the meaning of discourse connectives cross-linguistically using corpus data. We present the methodological choices made to reach this aim and report three annotation experiments using the framework of the Penn Discourse Tree Bank.

Multilingual Variation in the context of Linked Data

In this paper we present a revisited classification of term variation in the light of the Linked Data initiative. Linked Data refers to a set of best practices for publishing and connecting structured data on the Web with the idea of transforming it into a global graph. One of the crucial steps of this initiative is the linking step, in which datasets in one or more languages need to be linked or connected with one another. We claim that the linking process would be facilitated if datasets are enriched with lexical and terminological information. Being that the final aim, we propose a classification of lexical, terminological and semantic variants that will become part of a model of linguistic descriptions that is currently being proposed within the framework of the W3C Ontology-Lexica Community Group to enrich ontologies and Linked Data vocabularies. Examples of modeling solutions of the different types of variants are also provided.

Multilingual variation in the context of Linked Data

In this paper we present a revisited classification of term variation in the light of the Linked Data initiative. Linked Data refers to a set of best practices for publishing and connecting structured data on the Web with the idea of transforming it into a global graph. One of the crucial steps of this initiative is the linking step, in which datasets in one or more languages need to be linked or connected with one another. We claim that the linking process would be facilitated if datasets are enriched with lexical and terminological information. Being that the final aim, we propose a classification of lexical, terminological and semantic variants that will become part of a model of linguistic descriptions that is currently being proposed within the framework of the W3C Ontology- Lexica Community Group to enrich ontologies and Linked Data vocabularies. Examples of modeling solutions of the different types of variants are also provided.

A Statistical Approach for Multilingual Document Clustering and Topic Extraction from Clusters

2000 Mathematics Subject Classification: 62H30

Using parent report to assess early lexical production in children exposed to more than one language

Limited expressive vocabulary skills in young children are considered to be the first warning signs of a potential Specific Language Impairment (SLI) (Ellis & Thal, 2008). In bilingual language learning environments, the expressive vocabulary size in each of the child’s developing languages is usually smaller compared to the number of words produced by monolingual peers (e.g. De Houwer, 2009). Nonetheless, evidence shows children’s total productive lexicon size across both languages to be comparable to monolingual peers’ vocabularies (e.g. Pearson et al., 1993; Pearson & Fernandez, 1994). Since there is limited knowledge as to which level of bilingual vocabulary size should be considered as a risk factor for SLI, the effects of bilingualism and language-learning difficulties on early lexical production are often confounded. The compilation of profiles for early vocabulary production in children exposed to more than one language, and their comparison across language pairs, should enable more accurate identification of vocabulary delays that signal a risk for SLI in bilingual populations. These considerations prompted the design of a methodology for assessing early expressive vocabulary in children exposed to more than one language, which is described in the present chapter. The implementation of this methodological framework is then outlined by presenting the design of a study that measured the productive lexicons of children aged 24-36 months who were exposed to different language pairs, namely Maltese and English, Irish and English, Polish and English, French and Portuguese, Turkish and German as well as English and Hebrew. These studies were designed and coordinated in COST Action IS0804 Working Group 3 (WG3) and will be described in detail in a series of subsequent publications. Expressive vocabulary size was measured through parental report, by employing the vocabulary checklist of the MacArthur-Bates Communicative Development Inventory: Words and Sentences (CDI: WS) (Fenson et al., 1993, 2007) and its adaptations to the participants’ languages. Here we describe the novelty of the study’s methodological design, which lies in its attempt to harmonize the use of vocabulary checklist adaptations, together with parental questionnaires addressing language exposure and developmental history, across participant groups characterized by different language exposure variables. This chapter outlines the various methodological considerations that paved the way for meaningful cross-linguistic comparison of the participants’ expressive lexicon sizes. In so doing, it hopes to provide a template for and encourage further research directed at establishing a threshold for SLI risk in children exposed to more than one language.