Patchwriting

Rebecca Moore Howard introduced the term in the early 1990s to describe what she kept seeing in student writing from sources: text produced by deleting some words from the original, swapping in synonyms, and lightly rearranging clauses while the source's sentence skeleton stays intact. Her argument was pedagogical — patchwriting is a transitional stage in learning to write from sources, not theft — but detection software draws no such distinction.

Source: The intervention produced significant gains in reading fluency across all age groups. Patchwritten: The intervention created significant improvements in reading fluency across every age group.

Second-language writers default to patchwriting for understandable reasons. The source's English is often better than what they can produce from scratch, so staying close feels safe; technical meaning feels fragile, so changing the structure feels risky; and genuine restructuring demands exactly the syntactic resources an L2 writer has least of. Patchwriting is what careful, honest effort looks like one step before real paraphrase.

It still triggers flags. Similarity algorithms match strings and word sequences, and a patchwritten paragraph preserves enough of the original skeleton to light up a similarity score report even when the source is cited. The way out is to paraphrase from understanding rather than from the page: read the passage, close it, restate the point in your own structure, then reopen the source only to check accuracy. Push a paraphrasing tool to do the closeness-hiding for you and you trade patchwriting for a worse artifact — see tortured phrases.