Sentence Selection Task Research Articles

Input factors in the acquisition of evidentiality by Turkish heritage language children and adults in the United States

ABSTRACTAspectual and mood morphology are vulnerable domains in adult heritage speakers. This paper investigates the root of such vulnerability within the domain of Turkish evidentiality system by comparing 20 second-generation adult and 20 school-age child Turkish heritage speakers to 20 first-generation immigrants (main input providers for child heritage speakers) residing in the United States. A monolingual group from Turkey including 20 monolingual adults, 20 7-14-year-old and 20 3-6-year-old Turkish-speaking children were also tested. Results of a story-telling and a sentence selection task revealed weak mappings between the indirect evidential marker –mIş and its pragmatic use in story-telling as opposed to intact knowledge of the Turkish evidentiality system in the first-generation immigrant group. The school-age child and adult heritage speakers, on the other hand, were significantly the least accurate groups in both tasks, performing more variably than the 3-6-year-old monolingual children in Turkey. The individual analysis of the child heritage speaker data as well as a comparison of the child-parent dyads further showed that early heritage language experience and use rather than changes in the parental input are likely to be the source of the variability observed in child and eventually in adult Turkish heritage speakers in this domain.

Capturing Sentence Relations for Answer Sentence Selection with Multi-Perspective Graph Encoding

This paper focuses on the answer sentence selection task. Unlike previous work, which only models the relation between the question and each candidate sentence, we propose Multi-Perspective Graph Encoder (MPGE) to take the relations among the candidate sentences into account and capture the relations from multiple perspectives. By utilizing MPGE as a module, we construct two answer sentence selection models which are based on traditional representation and pre-trained representation, respectively. We conduct extensive experiments on two datasets, WikiQA and SQuAD. The results show that the proposed MPGE is effective for both types of representation. Moreover, the overall performance of our proposed model surpasses the state-of-the-art on both datasets. Additionally, we further validate the robustness of our method by the adversarial examples of AddSent and AddOneSent.

Predicting and Integrating Expected Answer Types into a Simple Recurrent Neural Network Model for Answer Sentence Selection

Since end-to-end deep learning models have started to replace traditional pipeline architectures of question answering systems, features such as expected answer types which are based on the question semantics are seldom used explicitly in the models. In this paper, we propose a convolution neural network model to predict these answer types based on question words and a recurrent neural network model to find sentence similarity scores between question and answer sentences. The proposed model outperforms the current state of the art results on an answer sentence selection task in open domain question answering by 1.88% on MAP and 2.96% on MRR scores.

The Overt Pronoun Constraint through the Spanish (Heritage) Lens

The Overt Pronoun Constraint (OPC) has been studied on numerous occasions, addressing, in particular, the capabilities and limitations of the second language learner (Gürel, 2003; Kanno, 1997; Pérez-Leroux & Glass, 1997; Rothman & Iverson, 2007). However, studies are increasingly showing that the OPC may not be an unyielding restriction on grammar, as previously proposed by Montalbetti (1984) (Alonso-Ovalle, Fernández Solera, Frazier & Clifton, 2002; Gelormini Lezama, Huepe, Herrera, Melloni, Manes, García & Ibáñez, 2016; Keating, Jegerski, & VanPatten, 2016; Lipski, 1996). This study looks at the interpretation restrictions of the OPC by 20 Spanish heritage and 20 monolingual speakers in two experimental tasks: a sentence selection task and a picture-matching task. Results reveal that while participants show a preference towards the interpretation restrictions of the OPC, they do not respond at ceiling, which allows for variability. Additionally, the heritage speaker group varies from the monolinguals, indicating further divergence. This work provides a descriptive analysis of the findings and contributes to the dialogue on the flexibility of the OPC on language.

The Proper Formulation of the Overt Pronoun Constraint: Evidence from Persian and Azeri

Generativists claim that there are some structural properties of the grammar that cannot be acquired solely on the basis of input. It is argued that such properties, known as principles, are part of UG common to all languages. The Overt Pronoun Constraint (OPC) principle, the focus of the present study, asserts that if in any language an alternation between overt/null pronouns be possible, the overt pronoun in an embedded clause cannot take a quantifier such as someone, everybody, and nobody as its antecedent in the main clause. To study OPC principle in four groups of speakers, namely Persian heritage, Persian monolingual, Azeri heritage, and Azeri monolingual, the present study gathered data from 20 participants in each group. Participants completed two comprehension tasks, i.e. a Picture-Matching Task (PMT), and a Sentence-Selection Task (SST). Both tasks tested the interpretation of the implicit knowledge of the OPC with quantified antecedents. All quantifiers were assumed to be treated equally in both tasks. Results of the SST revealed that there was no statistically significant difference between monolingual speakers of Persian and Azeri monolingual languages in understanding the interpretative contrast between overt and null pronouns within OPC contexts. Regarding the heritage speakers, it was found that they seemed to have problems with the OPC in the SST. As for the PMT, it was found that heritage speakers and monolinguals of both languages understood the interpretative contrast present with overt and null pronouns within OPC contexts with no significant difference between them. DOI: 10.5901/mjss.2016.v7n4S2p176

Eventive and stative passives in Spanish L2 acquisition: A matter of aspect

This paper reports on an empirical study that examined knowledge of eventive and stative passives in the L2 Spanish grammar of L1 speakers of English. Although the two types of passive exist in English, the difference between them is not signaled in any specific way. In Spanish, in contrast, the distinction is marked by the choice of copula:seris used to form eventive passives,estarfor statives. Researchers agree that the two copulas, both of which translate as English “to be”, differ in relation to aspect:estaris perfective while ser is not marked for aspect (Schmitt, 1992). The question was whether L2 learners would be able to acquire the aspectual difference of the copulas and apply it to the formation of the passives. Two main tests were used, a Grammaticality Judgment Task and a Sentence Selection Task. The Grammaticality Judgment Task examined properties of the passives related, among other things, to aspect and agentivity. The Sentence Selection Task focused on the interpretation of the subject: only the subject of ser can be interpreted as generic. Although the learners in general distinguished between grammatical and ungrammatical sentences, they had not acquired the restriction on subject interpretation. These results are explained in terms of interfaces.

Second language learnability and the acquisition of the argument structure of English locative verbs by Korean speakers

The present research focuses on Korean English as a foreign language (EFL) learners’ knowledge of the locative alternation (e.g., John loaded hay onto the wagon/John loaded the wagon with hay) and its relationship to theories of language-particular and language-universal properties. Korean, the native language of the participants, has a locative alternation resembling that of English. However, although Korean and English are similar in terms of broad-range constraints, they are dissimilar in terms of narrow-range constraints for locative alternations. This study investigates whether the acquisition of such constraints in English locatives by Korean speakers, and whether the first language (L1) influences the second language (L2) acquisition of locative alternations. Two instruments are used in the experiment: a forced-choice picture-description task and a forced-choice sentence selection task. The study investigates an experimental group of Korean learners of English and a control group of native speakers of English. The results are discussed with reference to universality of linking, to the transfer of argument structure and to Pinker’s learnability theory. The primary results are: • The Korean learners of English had acquired the constructional meaning of the locative construction (which is related to Pinker’s (1989) concept of broad-range rules and broad conflation classes), a property claimed to be universal. • They had not achieved native-speaker knowledge of language-particular properties - which narrow conflation class verbs belong to - so that they did not reject ungrammatical sentences; and • Significant L1 transfer effects were not found.