Using modern linguistic technologies, the project has compiled automatically thousands of articles on poverty, published in the leading newspapers in Argentina, Brazil, Colombia and Mexico: The Polame-Corpus.
The term "poverty" ("pobreza" in Spanish and Portuguese) is derived from latin, and its numerous meanings can be easily found in any dictionary. However, the concept of poverty may appear in a linguistic corpus as a synonym (indigente), antonym (rico), as part of complex expressions(pobreza absoluta), combined with specific verbs a large number of times, such as "combatir la pobreza" (fight poverty) or in a metaphor (pobreza de ideas).
Additionally, because the Polame-corpus has been gathered in four different countries and two languages, it is also important to take into account the geographically different Spanish variants of the countries studied, a fact that requires the work of specialists in Latin American Spanish and culture, as well as in Brazilian Portuguese and culture to correctly interpret the contextual meaning or meanings of the poverty expressions found in the Polame-corpus of newspaper texts. The POLAME-corpus is available online via a query interface, and project researchers have access to it through a web browser.
A combination of corpus-based (deductive) and corpus-driven (inductive) techniques to query the data is used. In the corpus-based approach, key poverty-related concepts that appear in existing resources are queried by using corpus data, e.g. the UNDP Thesaurus. For the corpus-driven work, an empirical study using linguistic and statistical methods is used to extract additional poverty-related concepts from the POLAME-corpus data, as well as from reference material, investigating consistent phrases (collocates) and frequencies of key terms related to poverty notions. Thus, the combination of both approaches allows for the study of concordances containing poverty-related expressions or allusions and lexical, semantic and hierarchical relations between terms, among others.
An analysis of meaning from various academic perspectives is being undertaken to characterize the semantic features of each context. Contexts will then be compared to suggest how the language used in newspapers can affect the perceptions of poverty of the readership. At this stage, a detailed context analysis will be done from the perspective of the social and political situation of each country interpreting each of the instances found according to the background of the press article containing it, and other relevant features.