personssuch

Personssuch is a coinage used in discussions of linguistics and natural language processing to describe a textual token that results from merging a two-word phrase such as “such persons” or “persons such” into a single word. This concatenation typically arises from OCR errors, auto-correct or hurried typing, but it can also appear as a deliberate stylistic variant in informal writing. In corpus work, personssuch is treated as a single token, even though it encodes a familiar multi-word unit.

Etymology and scope: The term blends the noun form “person” (often pluralized as “persons”) with the determiner or adjective “such.” The combined form signals the underlying phrase while highlighting how word boundaries can blur in automated text processing. While not a standard linguistic category, personssuch is discussed as an example of data noise and tokenization challenges in large text datasets.

Occurrence and contexts: Personssuch is most commonly observed in scanned or digitized texts where optical character recognition misreads spacing, as well as in noisy user-generated content, search logs, and code comments. It can also appear in historical corpora where spacing conventions differ from contemporary norms. The phenomenon is of interest mainly to researchers studying data cleaning, text normalization, and robust parsing.

Implications for processing: In natural language processing and information retrieval, personssuch can affect tokenization, part-of-speech tagging, and entity recognition. Preprocessing pipelines may include normalization steps to split merged tokens back into their constituent words (e.g., separating “personssuch” into “such persons” or “persons such”). This improves search accuracy and downstream linguistic analyses.

See also: tokenization, text normalization, OCR errors, corpus linguistics.