Diffusion of Lexical Change in Social Media (1210.5268v4)

Abstract: Computer-mediated communication is driving fundamental changes in the nature of written language. We investigate these changes by statistical analysis of a dataset comprising 107 million Twitter messages (authored by 2.7 million unique user accounts). Using a latent vector autoregressive model to aggregate across thousands of words, we identify high-level patterns in diffusion of linguistic change over the United States. Our model is robust to unpredictable changes in Twitter's sampling rate, and provides a probabilistic characterization of the relationship of macro-scale linguistic influence to a set of demographic and geographic predictors. The results of this analysis offer support for prior arguments that focus on geographical proximity and population size. However, demographic similarity -- especially with regard to race -- plays an even more central role, as cities with similar racial demographics are far more likely to share linguistic influence. Rather than moving towards a single unified "netspeak" dialect, language evolution in computer-mediated communication reproduces existing fault lines in spoken American English.

• The paper demonstrates that geographic proximity and demographic similarity, such as race, significantly influence the spread of new lexical entries.
• The authors employ a latent vector autoregressive model with Bayesian inference to analyze 107 million tweets and capture irregular sampling rates.
• Empirical results challenge traditional gravity models by showing that language change on social media reinforces existing social divisions.

Diffusion of Lexical Change in Social Media

The paper entitled "Diffusion of Lexical Change in Social Media" by Eisenstein et al. investigates the dynamics of language evolution facilitated by social media, specifically focusing on Twitter data. The paper employs extensive computational and statistical methodologies to analyze 107 million tweets, authored by approximately 2.7 million users, to understand the diffusion of new lexical entries across different geographical and demographic contexts.

The authors utilize a latent vector autoregressive model to parse the complex interactions inherent in this vast dataset. This model is particularly effective in accommodating the irregular sampling rates of Twitter and capturing patterns in linguistic change diffusion across the United States. The analysis delineates the influence of geographic proximity, population size, and demographic similarities—particularly race—in dictating how language evolves online, contradicting the broader notion of converging into a homogeneous "netspeak."

Key findings affirm that geographical and demographic lines significantly direct the spread of language changes. For instance, racial demographic similarities between cities, such as high proportions of African Americans or Hispanics, strongly correlate with shared linguistic influences. This is evidenced by the enhanced spread of terms such as "ion" and "ctfu" within demographically similar urban settings, even when geographically distant. As opposed to a sweeping linguistic uniformity, the paper suggests a scenario wherein language change accentuates existing social divisions.

Theoretical implications of this paper challenge models like the cascade or gravity models by emphasizing race and other demographic factors over pure geographic or population metrics. Practically, these findings could influence how marketers or policymakers understand and utilize social media language trends in demographically targeted strategies.

By capturing individual word usage dynamics with a Gaussian noise model and probabilistic frameworks, the research offers a robust approach to understanding language change in a digital age. Bayesian inference and logistic regression further elucidate demographic and geographic ties in the induced network of linguistic diffusion.

Future work could explore sub-regional or community-specific linguistic dynamics within metropolitan areas to gain deeper insights. Additionally, interdisciplinary methodologies could integrate insights from sociolinguistics and computational sciences to model other layers of language change beyond lexical items alone, such as syntactic or semantic shifts.

In summary, Eisenstein et al.'s research elucidates the localized and demographically influenced nature of lexical change in social media, underpinning the persistent influence of social structures on linguistic evolution. This work underscores the importance of considering diversity and demographic homophily in understanding the landscape of language change in the era of digital communication.