AAVE vocabulary diffusion Twitter/X's #Genz and #BlackTwitter

Overview

This project analyzes how African American Vernacular English (AAVE) terms spread from #BlackTwitter to broader communities like #GenZ over time (2020–2025).

Research Question

How do different types of AAVE words (e.g., nouns vs. verbs, structurally tied vs. not) survive as they spread beyond their originating community?

Methodology

• Words were chosen by a group of AAVE speakers that I consulted, and will be assumed to belong to AAVE, following the intuitions of my consultants
• Tweets were collected using the official Twitter/X API
• Queries included:
  • target word
  • hashtag community (#BlackTwitter or #GenZ)
  • year (2020, 2023, 2025)
• Retweets were excluded
• Tweets were manually filtered for correct semantic meaning
• Duplicate tweets were removed

Key Findings

• Words tied to AAVE grammatical structures show different survival patterns than standalone slang
• Part of speech appears to influence retention across communities
  • Of the observed words, adjectives had the lowest use retention in #GenZ after five years
  • Of the observed words, nouns had the highest use retention in #GenZ after five years
  • Of the observed words, verbs had the lowest use retention in #BlackTwitter after five years
  • Of the observed words, adjectives(only slighter more than nouns) had the highest use retention in #BlackTwitter after five years
  • Across all parts of speech, #BlackTwitter usage frequencies declined
  • #GenZ frequencies of nouns and verbs grew during the five-year interval, while adjective use declined after 5 years
• Frequenices of word usages can be seen in Frequencies.xlsx

Scripts

• unique_tweets.py: python script created by ChatGBT to create corpuses of tweet texts from the .json files, removing duplicates and counting unique tweets
• unique_ids.py: - unique_tweets.py: python script created by ChatGBT to create corpuses of tweet ids to enable publication of tweet data extracted from the .json files while respecting twitter privacy rules

Sources used to define AAVE and linguistic patterns

• Bailey, G. (2022, August 15). The History of African-American Vernacular English. Oxford Research Encyclopedia of Linguistics. Retrieved 17 Mar. 2026, from https://oxfordre.com/linguistics/view/10.1093/acrefore/9780199384655.001.0001/acrefore-9780199384655-e-926
• John R. Rickford. (2003). African American Vernacular English (AAVE). In International Encyclopedia of Linguistics (2nd ed.). Oxford University Press.

Sources that could be considered for future research

• Ilbury, C. (2020). “Sassy Queens”: Stylistic orthographic variation in Twitter and the enregisterment of AAVE. Journal of Sociolinguistics, 24(2), 245–264. https://doi.org/10.1111/josl.12366

Corpuses

• In total, 32,022 tweets were collected throughout the process (corpus of tweet ids can be found in idCorpuses folder), but not all data was used in the statistical analysis described above
• Three corpuses of data were created from #BlackTwitter tweets, corresponding to each year that was examined (#BlackTwitter2025, #BlackTwitter2023, #BlackTwitter2025)
• Three corpuses of data were created from #GenZ, corresponding to each year that was examined (#GenZ2025, #GenZ, #GenZ)
• An additional corpus containint the tweeet IDs for data scraped for the project, that helped confirm the vocabulary term's connection to AAVE and the decision to use #GenZ over other hashtags that were considered like #StanTwt, #StanTwitter, and #tiktok.

Limitations

• Hashtags are imperfect indicators of community identity not all AAVE speakers can be represented through data extracted from #BlackTwitter
• Terms were
• Semantic filtering introduces subjectivity

Ethical Considerations

• This project used the offical Twitter/X API to ensure that the privacy and ethical standards set by Twitter to protect its users were respected and followed.
• Tweets examined in this study were posted publicly
• To further protect the creators of the tweet's used in this study, tweet id's will be published instead of the tweet texts
• AAVE has had a longer history of stigmatization, and it is essential that viewers are careful about the conclusion they draw from this small set, whose scope is affected by the limitations outlined above

Further research

• The main goal of this was to see if there were relationships between word usage and the outlined categories
  • The relationships between part of speech and usage need more words to lead to stronger conclusions beyond the suggestion of a relationship
  • Examining determiners, discourse markers, and other parts of speech could be helpful
• The results and patterns shown are likely to be effected by the sample population, sample size, and medium of data
• Research on broader audiences (of both AAVE and non-AAVE) with more words and more temporal points would lead to clearer, more hollistic results
  • Different age groups, identity groups, and regions of non-AAVE speakers may have differences in rates and distributions of diffused words
  • Historically, the Southern-American has had close contact with AAVE, which could affect diffusion
  • Research has found a relationship between LGBT+ online community vernacular and AAVE (see Ilbury, C)

Author

Alexandra Wingo

Acknowledgments

• I would like to thank Nancy Apollo, Maya Hay, Carys Kirlew, Carter Hoskins, Dominique Quinonez, Erika Sowah, and Israel Allen for their help during the creation of the target word list.