The study synthesises corpus assisted discourse studies as a methodology, systemic functional linguistics as a theory of language, and sociological theories about social change as a theoretical framework. The data under investigation—all sentences published in the NYT between 1987 and 2014—was chosen for sociological and methodological reasons: sociologically, the publication is a respected and influential part of mainstream media; more practically, it has been digitised, is metadata-rich and provides a sample large enough to observe quantitatively reliable trends. Using a combination of existing and purpose-built tools, we transform NYT articles into a large, grammatically annotated corpus that can be queried and visualised using the Python programming languages. Using topic and title metadata, we also create a subset of articles in the health domain.
[1]
J. Zinn.
Risk as Discourse: Interdisciplinary Perspectives
,
2010
.
[2]
J. Chapman.
Comparative Media History: An Introduction: 1789 to the Present
,
2005
.
[3]
R. Yin.
Case Study Research: Design and Methods
,
1984
.
[4]
Hinrich Schütze,et al.
Book Reviews: Foundations of Statistical Natural Language Processing
,
1999,
CL.
[5]
John A. Bateman,et al.
Systemic-functional linguistics and computation : New directions, new challenges
,
2019
.
[6]
Joakim Nivre,et al.
Towards a Universal Grammar for Natural Language Processing
,
2015,
CICLing.