Social Sciences

Seth Joshua Goss, Phd

Assistant Professor, ECAS, Japanese Linguistics

Halle Institute for Global Research/URC Award

Testing an information-based model of the perception of nonnative speech prosody

How listener derives meaning from spoken language using prosody the melodic features above individual consonants and vowels such as tone, stress, and intonation is a question at the core of speech perception research. In some languages, such as Mandarin, prosody is vital for word identification; in others, like English, it provides the listener little help at all. How then, for example, does a speaker of English (where prosody is not very informative), who has learned Mandarin as a second language (where prosody is very informative) perceive the relevant prosodic features of Japanese as they begin learning this new third language? Moving beyond the linguistic pattern-by-pattern comparisons found in previous research, the current proposal aims to test an information-based framework of prosodic perception that accounts for how knowledge of, and level of proficiency inmultiple languages influence the perception of new languages.

Two questions frame the goals of this proposal: 1) To what degree do previously learned languages that differ in their usefulness of word-level prosody help or hinder the learning of prosody in new languages? and 2) Given that a learner's native language is not static, do nonnative languages reshape the perception of word-level prosody in one's native language? By answering these questions, I aim to bring nonnative speech perception research in line with recent descriptions of the multilingual mind that construe linguistic knowledge as a system shared across all languages. The proposal contributes to a comprehensive model of prosodic perception that provides a clearer understanding of the complexity and interconnectedness of language knowledge in the multilingual mind.

Hong Li, PhD

Professor, ECAS, Department of Russian and East Asian Languages and Cultures

Halle Institute for Global Research/URC Award

Investigating Parental Motivations Behind the Decision by Chinese Parents to Send Their Children to Study at Emory University

According to the data in Open Doors 2019, released by the Institute of international Education (IIE), China remained thelargestsource ofinternational students in the United States in 2018/19 for the 10th year with over 369,000 students in undergraduate, graduate, and optional practical training(OPT) programs. Previous research has mostly focused on the motivations of Chinese students in choosing to pursue an undergraduate or graduate education at universities in the United States; however, a major limitation ofcurrent research is the nan-ow focus on student motivation as the sole determinant behind the decision of Chinese students to study abroad.

Understood within both a global and Chinese cultural context, a nan-ow academic focus on student motivation obscures the possibly significant role that Chinese parentalmotivation plays in sending their children touniversities in the U.S. Drawing from a modified "push-pull model" as a framework for analyzing trends in parental and student motivations and gathering data through surveys and interviews, this research will identifyand analyzethe factors influencing parental motivation, such as the political and economic status quo of the United States and China, parental socioeconomic status, parent-student aspirations, andthe acquisitionofcultural capital. The resulting research should satisfy a gap in existing academicresearchwithregardsto the motivations of Chinese parents in sendingtheir children to study abroad and may be useful to stake holders and university staff working in offices that manage international students.

Elena Pesavento, PhD

Associate Professor, ECAS Economics

Dynamic Causal Effect; the dos and donts of local projections in linear and nonlinear models

There is a large literature interested in estimating the effect of unexpected shocks such as monetary policy or oil shocks. Traditionally, the dynamic effect of these shocks has been estimated in linear vector autoregressive (VAR) models by computing impulse response functions. Recently, a relatively new approach to estimation of impulse response functions has gained significant popularity: local projections (LP). The main motivation for using LP as alternative to tra- ditional estimators are its simplicity - it amounts to repeatedly estimating a single regression for the variables of interest by OLS - and the belief that LP estimates may be more robust to model misspecifications and nonlinearities. Recent theoretical work proves that VAR and LP estimate the same impulse response functions and, thus, share the same population estimand. In other words, asymptotically, there is no advantage to use LP over traditional estima- tors of impulse response functions. However, the two methods differ in terms or their finite-sample performance and the ranking between the two methods depends on assumptions about the autocovariance function of the data. In addition, while the equivalence is valid if we view VARs or LP regressions as linear approximations of a nonlinear processes, it does not hold if we augment the regressions with nonlinear terms or we are directly interested in the form of nonlinearity. The aim of this project is to closely analyze which estima- tion method (LP versus VAR) is more efficient in small samples in linear and nonlinear models with the aim of giving recommendations for practitioners.

Renard Sexton, PhD

Assistant Professor, ECAS, Political Science

Halle Institute for Global Research/URC Award

South China Sea Data Initiative

The South China Sea region, at the core of a rapidly growing Asia Pacific, has experienced growing levels of conflict in recent years due to overlapping maritime claims between Vietnam, the Philippines, Malaysia, Brunei, Indonesia and China. These disputes have provoked violent encounters between military and commercial vessels among these nations, as well as protracted diplomatic rifts and international court cases. Despite growing interest among scholars and policymakers, there is almost no systematic data on the incidence of conflict in the South China Sea, and accordingly empirical analyses to date have relied on largely anecdotal evidence to support their claims. This project will employ a combination of ‘big data,’ student research assistants in Southeast Asia and fieldwork to compile a comprehensive dataset on conflict incidents in the South China Sea. This dataset will allow us to substantially push the boundaries of empirical research on the region, and provide a public good to other scholars who may use the data. We will also employ modern techniques of causal inference and spatial econometrics in the study, improving on most existing empirical research on conflict in international relations. Finally, we will make the aggregate data available to the public through a partnership with the Emory Center on Digital Scholarship, creating a data portal and visualization tool.