Search

Resumptive pronouns are often regarded as a last-resort strategy for rescuing illicit long-distance dependencies. Previous work has demonstrated a facilitative role for resumptive pronouns in production as well as in comprehension, though not a grammatical option in the languages. This study examined whether the same pattern is found in Cantonese, a language that allows resumption as a legitimate relativization strategy. Young Cantonese monolingual children were tested for both comprehension and production of object relative clauses using different relativization strategies (the gap strategy vs. the resumptive pronoun strategy). Although children did not perform better on either strategy in comprehension, and never employed resumptive pronouns in their production, resumption did prevent misanalysis of prenominal relative clauses as matrix clauses, suggesting that these elements help reveal the full structure of the relative clause to the processor. The study also showed better performance on subject than on object relative clauses, suggesting that the widely observed subject--object asymmetry also applies to Cantonese.

Purpose: Life expectancy is strongly related to national income, whether there is an additional contribution of income inequality is unclear., Methods: We used negative binomial regression to examine the association of neighborhood-level Gini, adjusted for age, sex, and income, with mortality rates in Hong Kong from 1976 to 2006., Results: The association of neighborhood Gini with all-cause mortality varied over time (p-value for interaction < .01). Neighborhood Gini was positively associated with nonmedical mortality in 1976 to 1986; incident rate ratio (IRR) 1.09, 95% confidence interval (95% CI) 1.02-1.16 per 0.1 change and in 1991 to 2006, IRR 1.24, 95% CI 1.13-1.36, adjusted for age, sex and absolute income. Similarly adjusted, Gini was not associated with all-cause mortality in 1976 to 1986 (IRR 0.96, 95% CI 0.93-1.00) but was in 1991 to 2006 (IRR 1.25, 95% CI 1.20-1.29), when Gini was also positively associated with death from cardiovascular diseases, respiratory diseases and some cancers., Conclusions: Independent of income, income inequality was positively associated with nonmedical mortality rates at a low level of spatial aggregation, indicating the consistent harms of social disharmony. However, the impact on medical mortality was less consistent, suggesting the relevance of contextual factors., (Copyright © 2012 Elsevier Inc. All rights reserved.)

Social patterning of disease is pervasive and persistent. Disease patterns change with economic development and the attendant epidemiological transition. It is becoming evident that social patterns of disease are epidemiologically stage specific. In a population with a recent history of rapid economic development we examined social patterns of all-cause and cause-specific mortality over time to elucidate how economic development impacts disparities in health. We used concentration indices to provide a summary measure of disparities by income in potential years of life lost (PYLL) for the Hong Kong population from 1976 to 2006. For all-cause mortality and for each of the specific causes considered the concentration curve in 2006 dominated the 1976 concentration curve. The concentration index for all-cause PYLL was negligible in 1976, but increased over the period. PYLL attributable to injury and poisoning was fairly consistently associated with lower income, but PYLL attributable to cardiovascular diseases and cancer reversed from an association with higher income in 1976 to an association with lower income in 2006. Social disparities in health are not universal or homogeneous in origin. Attention should be focused on disease-specific causes of disparities, so that contextually specific prevention strategies can be implemented. This is of particular relevance to China and other emerging economies where there may be a window of opportunity to prevent disparities in cancer and cardiovascular diseases occurring., (Copyright (c) 2010 Elsevier Ltd. All rights reserved.)