Do Court Reporters Include "Uh" and "Um"? What Verbatim Really Requires
What the rules actually say
"Verbatim" sounds absolute, and most reporter statutes and manuals use the word without ever defining how far down it reaches — every word? Every sound? Search the manuals for an answer on hesitation sounds specifically and you find almost nothing. Almost.
The federal policy answers it directly
The Judicial Conference of the United States prescribes the transcript format for all federal district courts in the Guide to Judiciary Policy, Volume 6, Chapter 18. Its "Editing of Speech" rule (§ 18.10.1.b) is the clearest published statement in American court reporting on this exact question:
"The transcript should provide an accurate record of words spoken in the course of proceedings. All grammatical errors, changes of thought, contractions, misstatements, and poorly constructed sentences should be transcribed as spoken. … In the interest of readability, however, false starts, stutters, uhms and ahs, and other verbal tics are not normally included in transcripts; but such verbalizations must be transcribed whenever their exclusion could change a statement's meaning." Guide to Judiciary Policy, Vol. 6, Ch. 18, § 18.10.1.b
Notice the structure. Everything a speaker actually says — bad grammar, misstatements, half-finished thoughts — must go down as spoken. The license to omit covers only nonword verbal tics, it exists "in the interest of readability," and it ends the moment an omission could change meaning. That is the professional definition of verbatim in practice: every word, not every noise.
State manuals are silent — including Michigan's
Michigan's SCAO Manual for Court Reporters and Recorders binds reporters to "make a verbatim record of the proceedings" and forbids striking "that which actually took place" — but every example it gives concerns substance: stricken testimony, off-record discussions, judge-ordered omissions. We searched all 208 pages: it never mentions hesitation sounds. Its only discussion of "uh"-type utterances concerns "uh-huh" and "uh-uh" — the yes/no answer forms, which are a different thing entirely (more below).
Texas is the same story. The Uniform Format Manual for Texas Reporters' Records § 3.8 requires "all English words and other verbal expressions uttered during the course of the proceedings," and then its § 3.9 — the one rule that names an "uh" form — turns out to be about answers: "Uh-huh" is affirmative, "Huh-uh" is negative, transcribe them accordingly. New Jersey's transcript standards parallel the federal all-words language but omit the federal filler exception — again, silence. We found no state authority anywhere that requires keeping "uh" and "um," and none that forbids it.
What certified transcripts actually do
Rules aside, what do certified reporters actually deliver? We pulled certified transcripts from the public record — filed in dockets, posted by courts and agencies — and counted, using word-boundary matching with the uh-huh answer forms excluded and every hit checked in context:
| Certified transcript | Live, unscripted speech? | Standalone "uh"/"um" |
|---|---|---|
| 4 U.S. Supreme Court oral arguments, 2025 terms (Heritage Reporting) — Kennedy v. Braidwood, TikTok v. Garland, Trump v. CASA, Ames v. Ohio Dept. of Youth Services | Yes — rapid-fire argument | 0 across ~380,000 characters |
| Michigan PSC Case U-21990 public comment hearing (Dec. 2025) — dozens of ordinary citizens speaking unrehearsed | Yes — the most filler-prone speech there is | 0 — while 15 false starts stay on the page with the dash ("numbers of -- oh, I just lost it") |
| Michigan PSC transcripts by four more reporters (U-22036, U-21825, U-21904, U-21906 and others) | Colloquy and hearings | 0 in every one |
| All-day Michigan deposition, Doe v. Whitmer (E.D. Mich. No. 2:22-cv-10209, filed 2023) | Yes — full live cross-examination | 0 — while 30 "uh-huh"/"uh-uh" answers are kept |
| U.S. v. Google trial transcript, Day 7 (D.D.C. 2023, official court reporter) | Yes — witness cross-examination | 0 in 120 pages |
| FTC and FERC evidentiary hearings (several reporters, hundreds of pages incl. sworn cross) | Yes | 0 |
| Michigan depositions, Rodgers v. Fenton Area Public Schools (E.D. Mich. No. 2:23-cv-13110) | Yes — live sworn testimony | Kept: "Um, I mean, I don't -- I don't think it was a falling out." |
The public-comment hearing is the tell. Dozens of unrehearsed lay speakers produced a transcript with fifteen dash-marked false starts — the reporter was plainly capturing broken speech faithfully — and not a single "um." That is deliberate editing, not tidy speakers. And it is exactly the federal formula: real words kept, verbal tics dropped.
The Rodgers depositions are the honest counterweight: a certified Michigan reporter who keeps "um," in transcripts filed in federal court without objection. Keeping fillers is a real minority practice, not an error. Legal scholarship confirms the split leans heavily one way — a 2014 study in Legal Communication & Rhetoric ("The Lawyer's Guide to Um") documents that court reporters routinely excise fillers from deposition records. So the accurate statement is: dropping is the dominant convention; keeping is a permissible verbatim practice.
"Uh-huh" is not a filler — it's an answer
Whatever you do with "um," never touch the answer forms. "Uh-huh" (yes) and "uh-uh"/"huh-uh" (no) are testimony — often the entire answer to the question that matters. Every authority that speaks at all speaks to this: Texas UFM § 3.9 defines the affirmative and negative forms; Michigan's SCAO manual discusses "guttural utterances such as 'uh-huh' or 'uh-uh'" and urges getting a clear yes or no on the record; and it's why the deposition admonition ritual exists — the attorney warning the witness that "uh-huh and huh-uh make a confusing record." In our survey, even the strictest filler-dropping transcripts keep every one: the Doe v. Whitmer deposition has zero "um" and thirty "uh-huh"/"uh-uh" answers.
The same goes for the hyphenated spellings generally — uh-huh, huh-uh, mm-hmm — which are the standard dictionary forms. Drop the hesitations if you drop them; the answers stay, spelled the conventional way.
The one time you must keep them
The federal rule's own limit is the rule to remember: verbal tics "must be transcribed whenever their exclusion could change a statement's meaning." A long hesitation before a key admission, an "um" that marks a witness abandoning one answer for another, a stall that opposing counsel comments on — if the sound carries meaning, it is testimony and it goes on the page. New Mexico's reporter board put the same idea as a warning: editing for readability must never reach the point of "misrepresenting the tenor" of the testimony.
That judgment call — noise or meaning? — cannot be delegated to a style rule. It belongs to the certifying reporter, listening to the audio, one instance at a time. It is also, frankly, rare: the overwhelming majority of fillers are pure throat-clearing, which is why the overwhelming majority of certified pages don't carry them.
So which should you do?
If you're deciding your own practice
Either choice is defensible; the dominant convention is to drop. Judges and attorneys read transcripts for a living, and a page of "Um, all right. Well, I will, um, tell all of you that, uh, um…" serves nobody — least of all the speaker. If you drop, do it the way the certified record does it: hesitation sounds only; every real word, false start, and repeat stays; answer forms stay; meaning-bearing hesitations stay.
If a firm or agency sets the style
Match it. "Clean verbatim" (fillers out) and "true verbatim" (everything in) are both recognized products, and some clients — particularly where recorded statements or interpreter proceedings are involved — specifically order true verbatim. The style choice is the ordering party's; the accuracy underneath it is yours.
Whatever you choose, be consistent
A transcript that keeps some speakers' fillers and silently drops others' reads as editorializing — the one thing a reporter must never appear to do. Pick the rule, apply it to every speaker on every page, and let the certificate mean what it says.
Primary sources
- Guide to Judiciary Policy, Vol. 6, Ch. 18 — Transcript Format (§ 18.10.1.b, "Editing of Speech")
- Michigan SCAO Manual for Court Reporters and Recorders — Section 5 verbatim-record rule; guttural-utterances guidance
- Uniform Format Manual for Texas Reporters' Records — §§ 3.8–3.9
- U.S. Supreme Court certified oral-argument transcripts
- MPSC e-dockets — filed Michigan transcripts (e.g., Case Nos. U-21990, U-22036, U-21825)
- Gotthelf, "The Lawyer's Guide to Um," Legal Communication & Rhetoric: JALWD, Vol. 11 (2014)
Common questions
Do court reporters include "uh" and "um" in transcripts?
Most don't. Across 20+ certified transcripts we surveyed — SCOTUS, federal trials, Michigan hearings and depositions, 14 reporters and firms — the overwhelming majority contain zero standalone fillers. A minority of reporters keep them, and no rule prohibits either choice.
Does "verbatim" legally require filler words?
No published authority says so. The federal transcript policy expressly licenses omitting them for readability; Michigan, Texas, and other state manuals are silent on hesitation sounds. What verbatim does require: every actual word — grammar errors, misstatements, and false starts included — transcribed as spoken.
Are "uh-huh" and "uh-uh" fillers?
No. They're answers — affirmative and negative — and always stay in the transcript, in every style. Even the strictest filler-dropping certified transcripts keep every one.
When is dropping an "um" wrong?
When the hesitation itself carries meaning — the federal rule requires keeping verbal tics "whenever their exclusion could change a statement's meaning," and readability editing must never misrepresent the tenor of testimony. That call belongs to the certifying reporter.