Every writer leaves a fingerprint. Not a physical one, a linguistic one. The words we choose, the sentence structures we prefer, the punctuation habits we develop, the errors we consistently make, these patterns are surprisingly individual, surprisingly stable over time, and surprisingly difficult to suppress even when someone is actively trying to disguise their writing.
Forensic authorship analysis, the science of identifying or eliminating suspects based on linguistic evidence in written texts, has been used in criminal cases, fraud investigations, intellectual property disputes, and national security contexts around the world. In Chinese, it presents a fascinating set of challenges and opportunities that are almost entirely unexplored in English-language forensic literature.
What exactly constitutes a linguistic fingerprint? Forensic linguists look at multiple levels of language simultaneously:
Chinese presents several features that make authorship analysis both more complex and potentially more powerful than in alphabetic languages.
Character choice variation. Chinese has thousands of characters, many with overlapping meanings. A writer's character preferences, particularly for less common synonyms, are highly individual. Someone who consistently chooses 迅速 over 快速 for "fast," or 凝视 over 盯着 for "stare," is leaving a consistent stylistic trail.
Simplified vs. Traditional characters. A text written in Simplified characters by someone whose native writing system is Traditional, or vice versa, often contains revealing inconsistencies. Writers who learned Traditional Chinese and switched to Simplified may revert to Traditional forms for certain high-frequency characters under time pressure or inattention. These reversions are diagnostic.
Dialect interference. A writer whose native dialect is Cantonese, Shanghainese, or another variety will often allow dialect-specific vocabulary, grammar patterns, or phonetically-influenced character substitutions to appear in their written Mandarin. These interference patterns are highly geographically specific and can narrow a suspect pool dramatically.
Input method artifacts. Most Chinese digital text is produced using phonetic input methods (Pinyin or Zhuyin) that suggest characters based on sound. A writer who consistently selects the wrong homophone, using 在 where 再 was intended, for example, may be doing so habitually, creating a traceable error pattern tied to their specific input habits and dialect background.
Chinese authorship analysis has applications across a range of legal and investigative contexts:
The field is young, particularly as applied to Chinese. But the underlying principle is well established: language is behavior, behavior is individual, and individuality leaves traces. In Chinese, those traces are rich, varied, and, for someone with the right expertise, highly readable.