Finding More Needles by Building Bigger Haystacks - Dr. Jacob Eisenstein

Explore the intersection of large-scale text analysis and social science research in this 38-minute conference talk by Dr. Jacob Eisenstein. Delve into the methodological challenges and interdisciplinary opportunities presented by using big data to study linguistic and cultural phenomena. Learn about innovative approaches to discovering new linguistic variables, tracking language change, and analyzing hate speech on social media platforms. Gain insights into sociolinguistic research methods, from classic studies to cutting-edge applications of natural language processing and machine learning techniques. Examine case studies on Twitter language evolution and Reddit content analysis, and consider the ethical implications of defining and detecting hate speech in online communities.

Intro
Sociolinguistics: Big questions, small data
Labov's department store study
Why it worked
Outline
Rare linguistic events on Twitter
Discovering new linguistic variables
Discovering social variables
How does language change?
Language change as a networked cascade
Language change as epidemic
The role of tie strength
Which innovations succeed?
Finding (attempted) innovations
Hate speech on Reddit
A day after the paper came out
What excacly qualifies for hate speech?
Results with and without annotation
Some questions