Fig 1 - uploaded by Gloria Gagliardi
Content may be subject to copyright.
Validation & Extension interface 4. Conclusions: The IMAGACT e-learning platform Fig.2a shows how the IMAGACT infrastructure returns the information regarding the variation allowed by the English verb to turn. This verb records 12 action types, most of which also are appropriate for the Italian verb girare. The English user learning Italian, however, can discover that the type highlighted does not allow this translation since, on the contrary it is not a possible type for the Italian verb (il ragazzo alza / tira su il colletto vs * il ragazzo gira il colletto). This information can be accessed by clicking on the icon, once Italian is chosen as output language. If the user wants to go through an explicit process of language learning and is interested in mastering the difference between to turn and girare, IMAGACT provides him or her with an explicit method. The infrastructure allows one to compare the possible variations of to turn with those of girare. Through the comparison interface the user will discover both the range of variations allowed by each predicate and their differential. Figure 2.b shows that, among the large set of intersection (only partially reproduced) the semantic competence underling the use of the verbs girare and to turn presents differences in very specific types. The representation of action through scenes is crucial. The learning process bootstrapped by pointing to prototypic 3D scenes allows the learner to foresee that the verb can be applied in all instances of the Action type i.e. it gives rise to the mastering of a productive concept. This information can be acquired by a second language learner through natural language acquisition, however this requires a long exposure to language data. Moreover the productivity of types, which is evident to a mother tongue speaker, can never be taken for granted in a second language speaker. On the contrary the method provided in IMAGACT ensures an easy way to figure out how similar verbs of different languages can be used in different manners .
Source publication
Action verbs express important information in a sentence and they are the most frequent elements in speech, but they are also one of the most difficult part of the lexicon to learn for L2 language learners, because languages segment these concepts in very different ways. The two sentences "Mary folds her shirt" and "Mary folds her arms" refer to tw...
Context in source publication
Context 1
... competence based extension to Chinese Mandarin is in progress and consists in identifying a verb in the target language for each type of the source language and verifying the applicability to all instances in the target language. Fig. 1 shows how this task is accomplished in the case of the second and fourth type of Fig. 2. The Chinese verbs zhuàn e fàn respectively work fine in all instances of the Italian verb girare in those types, showing that the application of the verbs is productive. In principle this procedure can allow the implementation of whatever language ...
Similar publications
The aim of our paper is to present a case study carried out with a Translation and Interpreting student with Asperger's Syndrome (AS) in the framework of the e-learning project "TRADICOR: Corpora Management System for the Teaching Innovation in Translation and Interpreting" (PIE13-054). Our hypothesis is the following: when students with AS transla...
Citations
... IMAGACT4ALL, the competence-based extensions have been extended to incorporate any natural language. Research has already been conducted as to how to make use of the IMAGACT data as an e-learning platform for various languages (Moneglia et al., 2013). Moneglia et al., (2014ahave also explained the annotation of Sanskrit, Hindi and Bengali, the very first Indian languages that have been annotated on the platform. ...
Action verbs are one of the frequently occurring linguistic elements in any given natural language as the speakers use them during every linguistic intercourse. However, each language expresses action verbs in its own inherently unique manner by categorization. One verb can refer to several interpretations of actions and one action can be expressed by more than one verb. The inter-language and intra-language variations create ambiguity for the translation of languages from the source language to target language with respect to action verbs. IMAGACT is a corpus-based ontological platform of action verbs translated from prototypic animated images explained in English and Italian as metalanguages. In this paper, we are presenting the issues and challenges in translating action verbs of Indian languages as target and English as source language by observing the animated images. Among the ten Indian languages which have been annotated so far on the platform are Sanskrit, Hindi, Urdu, Odia (Oriya), Bengali, Manipuri, Tamil, As-samese, Magahi and Marathi. Out of them, Manipuri belongs to the Sino-Tibetan, Tamil comes off the Dravidian and the rest owe their genesis to the Indo-Aryan language family. One of the issues is that the one-word morphological English verbs are translated into most of the Indian languages as verbs having more than one-word form; for instance as in the case of conjunct, compound, serial verbs and so on. We are further presenting a cross-lingual comparison of action verbs among Indian languages. In addition, we are also dealing with the issues in disambiguating animated images by the L1 native speakers using competence based judgements and the theoretical and machine translation implications they bear.
... IMAGACT4ALL, the competence-based extensions have been extended to incorporate any natural language. Research has already been conducted as to how to make use of the IMAGACT data as an e-learning platform for various languages (Moneglia et al., 2013). Moneglia et al., (2014ahave also explained the annotation of Sanskrit, Hindi and Bengali, the very first Indian languages that have been annotated on the platform. ...
Action verbs are one of the frequently occurring linguistic elements in any given natural language as the speakers use them during every linguistic intercourse. However, each language expresses action verbs in its own inherently unique manner by categorization. One verb can refer to several interpretations of actions and one action can be expressed by more than one verb. The inter-language and intra-language variations create ambiguity for the translation of languages from the source language to target language with respect to action verbs. IMAGACT is a corpus-based ontological platform of action verbs translated from prototypic animated images explained in English and Italian as meta-languages. In this paper, we are presenting the issues and challenges in translating action verbs of Indian languages as target and English as source language by observing the animated images. Among the ten Indian languages which have been annotated so far on the platform are Sanskrit, Hindi, Urdu, Odia (Oriya), Bengali, Manipuri, Tamil, Assamese, Magahi and Marathi. Out of them, Manipuri belongs to the Sino-Tibetan, Tamil comes off the Dravidian and the rest owe their genesis to the Indo-Aryan language family. One of the issues is that the one-word morphological English verbs are translated into most of the Indian languages as verbs having more than one-word form; for instance as in the case of conjunct, compound, serial verbs and so on. We are further presenting a cross-lingual comparison of action verbs among Indian languages. In addition, we are also dealing with the issues in disambiguating animated images by the L1 native speakers using competence-based judgements and the theoretical and machine translation implications they bear.
... La scarsa conoscenza degli ambiti di variazione dei predicati nelle diverse lingue ha determinato, di fatto, una carenza di strumenti glottodidattici specifici per l'acquisizione del lessico verbale, nodo cruciale nello sviluppo della competenza della lingua seconda. IMAGACT è idoneo per l'acquisizione lessicale fin dalle prime fasi, anche per apprendenti privi di lingua veicolare: le entrate lessicali presentate si riferiscono, infatti, ad azioni comuni nella vita quotidiana e gli eventi possono essere facilmente identificati dall'utente indipendentemente dalla lingua madre, perché rappresentati in scene filmate o animate (Moneglia et al., 2013). ...
Il volume è strutturato in due parti: la prima, di semantica e linguistica computazionale, ha come argomento l’ontologia interlinguistica dell’azione IMAGACT (capitoli 1, 2 e 3); la seconda è invece dedicata alle applicazioni della risorsa (capitoli 4 e 5). Dopo un breve preambolo, nel capitolo 1 viene presentato il progetto IMAGACT, del quale sono descritti i principi metodologici ed il workflow complessivo. Il progetto ha prodotto un’ontologia interlinguistica che rende esplicito lo spettro di variazione pragmatica associata ai predicati azionali a media ed alta frequenza in italiano ed inglese. Le classi di azioni che rappresentano le entità di riferimento dei concetti linguistici, indotte da corpora di parlato da linguisti madrelingua, sono rappresentate in tale risorsa lessicale nella forma di scene prototipiche (Rosch, 1978, 1999). La metodologia sfrutta la capacità dell’utente/apprendente di trovare somiglianze tra immagini diverse indipendentemente dal linguaggio, sostituendo alla tradizionale definizione semantica, spesso sottodeterminata e linguo-specifica, il riconoscimento e l’identificazione dei tipi azionali. Il capitolo 2 è dedicato alla procedura di annotazione, che ha permesso l’estrazione bottom-up delle classi di azioni che costituiscono l’ontologia. Vengono presentati i dati linguistici di partenza (corpus IMAGACT-IT), nonché l’interfaccia informatica che ha reso possibile lo svolgimento dei vari task. Parallelamente, vengono presentate le linee guida per la standardizzazione delle occorrenze, la loro tipizzazione e l’annotazione delle proprietà linguistiche (struttura tematica, alternanze argomentali, aktionsart). Il capitolo 3 affronta invece la procedura di mapping inter-/intra- linguistico: dopo averne descritto i fondamenti teorici e aver individuato le questioni poste dai dati linguistici in input all’architettura della struttura dati, viene illustrata la metodologia che ha portato alla produzione del database 1.0. Nel capitolo 4 viene proposta una procedura di validazione dei dati linguistici, passaggio imprescindibile per consentirne l’applicabilità in ambiti differenti da quello in cui sono stati prodotti. Viene inoltre fornito un ampio quadro bibliografico sul concetto di “inter-rater agreement”, dando conto degli aspetti critici della valutazione dei dati e dei coefficienti statistici più diffusi. Per contestualizzare, vengono inoltre presentati i risultati di un esteso spoglio bibliografico sui principali studi e sulle maggiori campagne di valutazione nel campo della semantica lessicale. Il metodo viene testato su un sottoinsieme di lemmi coeso dal punto di vista semantico. Il capitolo 5 presenta infine la batteria di test SMAAV (Semantic Memory Assessment on Action Verb), creata dall’autrice a partire dai dati IMAGACT validati nel capitolo 4. La classe grammaticale del verbo, che di solito nelle lingue naturali lessicalizza le azioni, è stata tradizionalmente meno studiata ed utilizzata rispetto a quella del nome in ambito psicometrico. I materiali messi a disposizione dall’ontologia rappresentano una assoluta novità nel settore: la conoscenza sistematica del lessico verbale azionale rende possibile l’utilizzo di un campionario di lemmi ampio e al tempo stesso controllato. Inoltre, a differenza dei materiali utilizzati comunemente nella pratica medica, di tipo statico, gli stimoli multimediali messi a disposizione da IMAGACT garantiscono una rappresentazione migliore, più “ecologica”, delle azioni. Sebbene il test possa essere usato come misura generale dell’acceso al lessico e dell’erosione della memoria semantica, ne viene proposto un impiego diagnostico specifico, lo studio del Mild Cognitive Impairment, un settore di ricerca che ha ricevuto attenzione crescente negli ultimi anni. Dopo aver presentato le caratteristiche del disturbo ed avere illustrato i materiali testologici a disposizione di clinici e ricercatori, viene descritto il processo di realizzazione e taratura della batteria. La sezione si conclude con un’analisi linguistica e gestuale delle risposte date dagli informanti nel corso delle sessioni di test. In una breve conclusione vengono infine riassunti i risultati ottenuti e sono indicate alcune possibilità di sviluppo futuro.
The article considers contrastive linguistics as a discipline that interacts closely with its intralinguistic and applied neighbouring disciplines. Within this framework, the online ontology IMAGACT presents an instrument that allows to contrast how languages lexicalize concrete actions (movements, modification of objects, setting relations among objects, etc.) in their verbs.
German and Italian, the language pair considered here, differ typologically in their lexicalization strategies, which leads to difficulties in L2 acquisition, translation and lexicography. The article shows how the corpus-based IMAGACT database, which presents a set of 1010 actions in short films and links them to the appropriate verbs in 15 languages, provides help in these fields, and how it can at the same time empirically support contrastive-typological findings
