Regex to find repeated words ("the the")

Answer

\b(\w+)\s+\1\b with the gi flags finds a word followed by whitespace and the same word again. \1 is a backreference: it matches exactly the text that group 1 captured. Replace with $1 to remove the duplicate.

JavaScript ✓PCRE ✓

No catastrophic-backtracking patterns found. (A heuristic check: nested and adjacent repeats, overlapping alternatives.)

\bA word boundary: the edge between a word character and a non-word character (or the string's start or end)anchor
(…)Capturing group #1: remembers what it matchedgroup
\wA word character: a letter, digit or underscore (A-Z, a-z, 0-9, _)character type
+Repeat it one or more times, as many as possible (greedy)quantifier
\sA whitespace character (space, tab, line break and other Unicode spaces)character type
+Repeat it one or more times, as many as possible (greedy)quantifier
\1Backreference: the exact text that group #1 matchedbackreference
\bA word boundary: the edge between a word character and a non-word character (or the string's start or end)anchor

1 capturing group.

Examples

Checked at build time in JavaScript and in Python's re.

Input	Matches?
`the the cat`	✓ match
`The the cat`	✓ match
`this is is fine`	✓ match
`the cat`	✗ no match
`the theory`	✗ no match
`end end`	✓ match
`that that's it`	✓ match
`bye-bye`	✗ no match

Try your own text

One string per line: each line is tested on its own (the m flag is added here).

How it works

Group 1 captures a whole word, since \b forces the match to start at a word boundary. \s+ allows any whitespace between the two copies, including a line break, which is where doubled words usually hide in edited text. The final \b stops the theory from matching: after the comes o, a word character, so there is no boundary.

With the i flag, the backreference compares case-insensitively in both JavaScript and Python, so The the matches. s.replace(/\b(\w+)\s+\1\b/gi, "$1") keeps the first copy.

Variants

/(?<![\p{L}\p{N}_])([\p{L}\p{N}_]+)\s+\1(?![\p{L}\p{N}_])/giu
Unicode-aware version for non-English text. Lookarounds replace \b, which is ASCII-only in JavaScript.

Gotchas

that that's matches because the apostrophe is not a word character, so \b holds after the second that. Expect some false positives in English text.
Some doubled words are correct: "I knew that that was wrong" or "had had". Use this to flag for review, not to auto-fix.
\w and \b are ASCII-only in JavaScript, so café café is not found at all. Python 3 finds it, because its \w is Unicode-aware. Use the variant below for non-English text.
Repeated punctuation-separated words like bye-bye are not caught, by design.

Related explainers

← Explain any regular expression