The Language of Congress

• Speech counts include only remarks attributable to a specific member by Bioguide ID; procedural speech, submitted Extensions of Remarks, and material printed into the Record are excluded. Named-member coverage begins January 25, 1994; the separate long-run party charts use the Stanford Hein corpus back to 1873.
• Profanity uses a conservative, hand-curated list rather than an exhaustive dictionary. Passages marked as quotations are excluded from a member's rate and retained as a separate audit count.
• Slurs use an exact list of ethnic, sexual-orientation, gender-identity, and disability slurs used in the United States, drawn from Wikipedia's slur lists and cross-checked against a Kaggle profanity collection. Forms that are also ordinary words, names, places, or clinical terms are left out, as are verb uses such as "retard the growth." Quoted slurs are excluded, but a member who repeats a slur to condemn or describe it is still counted, so a count is not evidence of endorsement.
• A member's most-used profanity term is the most frequent accepted, unquoted surface form in the selected period; alphabetical order breaks ties. It does not imply personal preference.
• The Congressional Record is lightly edited rather than verbatim, and source metadata can be corrected after publication.

Speech coverage 2017-01-04 to 2026-10-09. Newest Congressional Record date: 2026-10-09. Site data snapshot: 2026-10-09T00:00:00+00:00. Open member activity and bill tables.