Linguistic Analysis of Fake News

Study of language patterns, stylistic features, and lexical characteristics that distinguish deceptive or unreliable text from trustworthy content.

Key Papers

• Oshikawa, Qian, & Wang (2020) — A Survey on Natural Language Processing for Fake News Detection: Demonstrates that hand-crafted linguistic features (LIWC, POS) improve classical ML baselines but fail to combine with neural networks; emphasizes shift toward learned representations (word embeddings, LSTM, CNN) over explicit linguistic features
• Sharma et al. (2018) — Combating Fake News: A Survey on Identification and Mitigation Techniques: comprehensive survey of linguistic and stylometric methods including Part-of-Speech (POS) tagging, Probabilistic Context Free Grammar (PCFG) features; discusses limitations of shallow linguistic methods compared to deep learning approaches
• Truth of Varying Shades: Analyzing Language in Fake News and Political Fact-Checking