Examinations
Sound-alike trademarks: how marks are compared by sound and what a linguist checks
Section 2(d) of the Lanham Act bars registration of a mark that so resembles a registered mark that it is likely to cause confusion, and sound is one of the ways two marks are compared. The Federal Circuit has said there is no correct pronunciation of a trademark, so the owner's intended reading does not settle the question. A linguist describes how speakers are likely to say each mark and where the sounds differ.

When do two trademarks sound alike in the eyes of the law?
When the sound brings the marks close enough that the later mark is likely to cause confusion. Section 2(d) of the Lanham Act bars registration of such a mark, and the first DuPont factor covers sound.
The rule is in 15 U.S.C. 1052(d). Registration may be refused to a mark that "consists of or comprises a mark which so resembles a mark registered in the Patent and Trademark Office, or a mark or trade name previously used in the United States by another and not abandoned, as to be likely, when used on or in connection with the goods of the applicant, to cause confusion, or to cause mistake, or to deceive." The statute does not mention sound. The method of comparison comes from case law.
In In re E. I. du Pont de Nemours & Co., 476 F.2d 1357, decided May 3, 1973, Chief Judge Markey of the Court of Customs and Patent Appeals listed the factors that, "when of record, must be considered" under Section 2(d). The first is "the similarity or dissimilarity of the marks in their entireties as to appearance, sound, connotation and commercial impression." The second is the similarity and nature of the goods or services. The Federal Circuit still calls them the DuPont factors. In 2012 it repeated that the Board "should simply use the DuPont factors" when it compares depictions of a standard character mark with another mark.
The USPTO page on likelihood of confusion puts the starting point in one line: "Trademarks don't have to be identical to be confusingly similar." Its example of marks that sound alike while spelled differently is T. Markey and Tee Marquee.
In In re Viterra Inc., 671 F.3d 1358 (Fed. Cir. 2012), the court called the Board's determination of likelihood of confusion a legal determination and added: "The ultimate question of likelihood of confusion, however, is based upon factual underpinnings." How the marks sound is one of those underpinnings. The court reviews the Board's legal conclusions de novo and its factual findings for substantial evidence.
Is similarity in sound enough on its own?
Sound can carry a case, but it is one factor. The USPTO summary of the May 2026 TMEP edition repeats that similarity in sound is one factor in determining whether two marks are confusingly similar.
The USPTO summary of changes in the May 2026 TMEP lists section 1207.01(b)(iv), on similarity in sound and phonetic equivalents. Its first point reads: "Similarity in sound is one factor in determining whether the marks are confusingly similar." The authorities given are the DuPont opinion at page 1361 and In re White Swan, Ltd., a Board decision of 1988.
The summary of the same section lists Board decisions in which marks were found close in sound. In In re Cresco Mfg. Co. (1963) the Board wrote: "The terms 'CRESCO' and 'KRESSCO' are phonetically indistinguishable." ENTELEC and INTELECT were found "similar in appearance and substantially identical in pronunciation" in 1983. AFCO and CAFCO, which differ by the letter C, were "substantially similar in appearance and sound" in 1977.
In Viterra the Federal Circuit held that substantial evidence supported the Board's determination "that any minor differences in the sound of these marks may go undetected by consumers and, therefore, would not be sufficient to distinguish the marks." The goods in that case were the same: agricultural seed in the application and agricultural seeds in the cited registration.
The other factors still count. In the GASPER ROOFING appeal described below, the services were identical and the Board reversed the refusal anyway, because it found the words different in sound and the buyers of roofing services careful.
Is there a correct way to pronounce a trademark?
No. In In re Viterra (2012) the Federal Circuit repeated that there is no correct pronunciation of a trademark and that consumers may pronounce a mark differently than its owner intends.
Viterra Inc. filed an intent-to-use application, Serial No. 77/608,885, on November 6, 2008, to register XCEED in standard characters for "agricultural seed." The examining attorney cited Registration No. 3,339,424, owned by X-Seed, Inc., for "agricultural seeds." That mark consists of X-SEED with a stylized letter X and surrounding dots, and the registrant disclaimed the word "seed." Viterra argued that the marks are pronounced differently and offered dictionary evidence.
The Board made the factual finding that "it is hard to imagine that the two marks will not sound alike when spoken" and was not persuaded that a difference in pronunciation would even be noticed by prospective purchasers. On March 6, 2012, the Federal Circuit affirmed in an opinion by Judge O'Malley. It agreed with the Board that "there is no correct pronunciation of a trademark" and that consumers may pronounce a mark differently than intended by the brand owner, and it cited the Board's LEGO and MEGO decision of 2002. The May 2026 TMEP summary describes that decision: the two marks differ in one letter and could be pronounced similarly as "leg" and "meg" followed by a long "o" sound.
The USPTO summary of the May 2026 TMEP ties the rule to the kind of word: the principle that there is no correct pronunciation applies "with greater force" when one of the marks is a coined term without a specific meaning. Its examples are coined words. IKEA and AKEA "are coined terms that look alike and sound alike" with "no apparent meaning in any language" (2014). In 1960 the Board found DIABINESE and DIABLAISE "substantially similar in sound" when spoken, since both were "coined marks having no specific meaning and no established pronunciation."
What if a word already has a settled pronunciation?
Then the Board can rely on it. In In re Jimenez (2025), as reported by TTABlog, the Board reversed a refusal of GASPER ROOFING over JASPER CONTRACTORS: Gasper takes a hard g, Jasper a soft j.
The USPTO cites In re Jimenez, Serial No. 97551823, decided in 2025, for the sentence on coined terms quoted above. The facts of the appeal below are taken from the TTABlog report of November 7, 2025, which quotes the precedential opinion written by Judge Angela Lykos.
The applicant sought to register GASPER ROOFING for roofing services. The examining attorney refused it over the registered mark JASPER CONTRACTORS and argued that GASPER and JASPER are phonetic equivalents. The Board noted that "Gasper" is a surname, that it begins with the letters of the word "gasp," which is pronounced with a hard "g," and that the dictionary entry for "jasper" gives the pronunciation "ˈja-spər." It concluded that "Gasper" would be pronounced with a hard "g" sound and "Jasper" with a soft "j" sound, so that the two words "sound different when verbalized."
The services were identical, and the trade channels and classes of buyers were the same. The Board still found that the marks differ in sound, appearance, meaning and commercial impression, took into account the care that buyers of roofing services exercise, and reversed the refusal.
After this decision a linguist who compares two marks by sound records, for each mark, whether it is a dictionary word, a surname or a coined term, and names the dictionary or the surname source behind that answer.
What evidence shows how buyers will say a mark?
In the decisions discussed here: dictionary entries, evidence that a word is a surname, a survey of how people read the marks aloud, and a linguist's affidavit on English sound and spelling rules.
Dictionary evidence did not win Viterra's appeal: the Board was not persuaded that buyers would notice the difference in pronunciation it pointed to. In Jimenez the dictionary entry for "jasper" and the evidence that Gasper is a surname supported the applicant.
A survey can test pronunciation directly. In Aveda Corp. v. Evita Marketing, Inc., 706 F. Supp. 1419 (D. Minn. 1989), the defendants added horizontal lines over letters of their mark Avita. The court noted that "Even with the horizontal lines added, twenty-five out of the forty people surveyed" still pronounced Avita very much like Aveda. The defendants pointed out that only two of the women sampled were recorded as pronouncing the two marks exactly the same. The court answered: "The likelihood of confusion standard, however, does not require an infringing mark to be a phonetic duplicate, only phonetically similar."
The same case had a third kind of evidence, an affidavit of a linguist, described in the next section.
Has a court relied on a linguist's analysis of how marks sound?
Yes. In Aveda Corp. v. Evita Marketing (1989) a federal court in Minnesota relied on the affidavit of a University of Minnesota linguist, together with a survey, to find AVEDA and AVITA confusingly similar in pronunciation.
Aveda sold hair and skin care products under AVEDA through authorized salons and beauty schools. The defendants sold shampoos and conditioners under Avita, mainly through salons, and both lines were expensive. Aveda filed an affidavit of Genevieve Escure, an associate professor of English and linguistics at the University of Minnesota, on how the two names would be pronounced.
Her two main points were that T and D tend to sound alike in casual speech between vowels when the second vowel is unstressed, and that the letters E and I correspond to several pronunciations, at least two of which are shared. She added that the lines over the letters suggest a stress pattern that English rules exclude, because those rules "do not allow two consecutive syllables both bearing primary stress," and that such lines "are not generally recognized as pronunciation guides by lay people." She concluded that Aveda and Avita "are likely to be confused."
In the opinion of March 2, 1989, Judge MacLaughlin noted that the defendants "do not directly contest that conclusion" and found that the evidence "clearly does establish a confusing similarity between the pronunciation of Aveda and Avita." The court granted a preliminary injunction against the use of Avita for hair care products and granted Aveda summary judgment on the counterclaim to cancel its registrations.
What does a linguist compare in two marks?
The sounds and syllables of each mark, their order and structure, where the differences fall, how the spelling is likely to be read, and what each word means, checked against dictionaries and records of word use.
Daniel A. Dinnsen proposed measures for this comparison in "Phonetic Similarity and Trademark Law," reissued in 2017 in a special volume of the IULC Working Papers (vol. 17, no. 6) that reprints volume 5 of Innovations in Linguistic Education. The abstract states that "the law sets out no explicit or objective criteria for assessing phonetic similarity and likelihood of confusion," and that "The Courts do nonetheless allow expert testimony from linguists on this point."
According to the abstract, the paper presents "objective measures from linguistics and cognitive psychology" as criteria for assessing phonetic similarity and likelihood of confusion, and "These criteria have been successfully applied in several recent court cases."
The Board decisions listed in the USPTO summary of the May 2026 TMEP show the units such a comparison works with. AFCO and CAFCO differ only by the letter C. LEGO and MEGO differ in one letter. CRESCO and KRESSCO are spelled differently, and the Board found them phonetically indistinguishable.
Escure's affidavit in Aveda drew on the same kinds of evidence: rules of casual speech, the sounds that English spelling assigns to vowel letters and the stress rules of English. Spelling and meaning are compared in the same report, because DuPont looks at the marks as to appearance, sound, connotation and commercial impression.
What happens when one of the marks is a foreign word?
Under the doctrine of foreign equivalents, words from common languages are translated into English, when the ordinary American purchaser would stop and translate them. A linguist can state the language and the translation.
In In re Spirits International, N.V., decided April 29, 2009, Judge Dyk quoted the Federal Circuit's earlier decision in Palm Bay Imports v. Veuve Clicquot (2005): "Under the doctrine of foreign equivalents, foreign words from common languages are translated into English." The ordinary American purchaser "includes all American purchasers, including those proficient in a non-English language who would ordinarily be expected to translate words into English."
The mark in Spirits International was MOSKOVSKAYA for vodka with no connection to Moscow. The Board had taken notice that Russian is spoken by 706,000 people in the United States, which the court put at about 0.25% of the population. The case arose under Section 2(e)(3). The court held that the Board had applied the wrong test, because that section requires that a substantial portion of the relevant consumers be likely to be deceived, and it vacated and remanded.
In Spirits International the court also quoted Palm Bay on the limits of the doctrine: it "is not an absolute rule" and "should be viewed merely as a guideline," and it applies when the ordinary American purchaser "would stop and translate the word into its English equivalent." As the court later described Palm Bay in Vetements, it was a likelihood of confusion case: the court found it improbable that the average American purchaser would stop and translate "VEUVE" into "widow," and the refusal of VEUVE ROYALE still stood on the marks VEUVE CLICQUOT PONSARDIN and VEUVE CLICQUOT.
In In re Vetements Group AG, decided May 21, 2025, the court applied the doctrine to VETEMENTS, the French word for clothing, filed for clothing and for online retail store services for clothing. It placed the burden on the party opposing translation and held that it is enough that an "appreciable" number of Americans are capable of translating the term. The marks were found generic, and the refusal was affirmed. In disputes of this kind the linguist states what the word means, in which language, and which dictionaries of that language record it.
Can a linguist testify that the marks are likely to be confused?
Rule 704(a) does not bar an opinion on an ultimate issue, but its advisory note targets opinions phrased in legal criteria. Likelihood of confusion is the tribunal's legal determination, made on findings about sound, appearance and meaning.
Rule 704(a) of the Federal Rules of Evidence says: "An opinion is not objectionable just because it embraces an ultimate issue." The advisory committee note adds that the rules "stand ready to exclude opinions phrased in terms of inadequately explored legal criteria." Its example: the question "Did T have capacity to make a will?" would be excluded, while a question that lists the elements of capacity to make a will would be allowed.
In Viterra the Federal Circuit treated the sound of the marks as a factual finding of the Board and likelihood of confusion as a legal determination. In Aveda the linguist concluded that Aveda and Avita "are likely to be confused." The defendants did not directly contest that conclusion, and the court also had a survey in front of it.
A report that answers language questions gives the tribunal statements it can check against the record: how English speakers are likely to pronounce each mark, which sounds and syllables coincide and in what positions, whether a word has a meaning in a common foreign language, and which dictionaries and usage records support each answer. Whether those facts add up to a likelihood of confusion is decided by the Board or the court.
What does IAE prepare in a dispute over similar-sounding marks?
An IAE linguist compares two marks in spelling, pronunciation and meaning, with dictionary and corpus evidence of word use, and prepares a report on the client's questions or a review of the opposing linguist's report.
Trademark similarity is one of the subjects on the IAE forensic linguistics page, together with disputed statements, contract wording and the authorship of anonymous texts. IAE puts questions to the linguist about the language of the text and leaves legal conclusions to the court. The questions are confirmed with the client before work starts.
The secretariat confirms receipt of the materials the same day. Within one business day IAE confirms the format, cost, deadline and the specialist from the registry, and the client approves the candidate and the CV before work starts. The document is issued in up to 5 business days after the format and cost are confirmed. For cases with tight deadlines there is a 48-hour rush.
Every specialist is a member of the public IAE registry and can be checked by Member ID or QR code. The specialist confirms in the document that he or she has no interest in the outcome of the case. A report is evidence that the tribunal weighs with the other evidence, and neither IAE nor the specialist guarantees the outcome. IAE is not a law firm and does not advise on case strategy. Materials are sent through the contact form and are not passed to third parties without the client's consent.
Sources
- 15 U.S.C. 1052, Trademarks registrable on principal register (Cornell LII)
- In re E. I. du Pont de Nemours & Co., 476 F.2d 1357 (C.C.P.A. 1973) (vLex)
- USPTO, Likelihood of confusion
- USPTO, summary of changes in the Trademark Manual of Examining Procedure, May 2026 (PDF)
- In re Viterra Inc., 671 F.3d 1358 (Fed. Cir. 2012) (law.co)
- TTABlog, Precedential No. 10: In re Jason Jimenez, GASPER ROOFING and JASPER CONTRACTORS, November 7, 2025
- Aveda Corp. v. Evita Marketing, Inc., 706 F. Supp. 1419 (D. Minn. 1989) (Justia)
- D. A. Dinnsen, Phonetic Similarity and Trademark Law, IULC Working Papers, vol. 17, no. 6 (2017), special volume reissuing Innovations in Linguistic Education, vol. 5
- In re Spirits International, N.V., No. 2008-1369 (Fed. Cir. Apr. 29, 2009) (Justia)
- In re Vetements Group AG, Nos. 2023-2050, 2023-2051 (Fed. Cir. May 21, 2025) (FindLaw)
- Federal Rules of Evidence, Rule 704 and advisory committee note (Cornell LII)
Read also
All articles
Puffery or false advertising: where is the line, and what does a linguist examine in the claim?
"Better Ingredients. Better Pizza." was puffery, and so was "America's Favorite Pasta." "50% Less Mowing" was a statement of fact. The line between
What is a true threat, when is an angry message a crime, and what does a linguist examine in the words?
A true threat is a serious expression of an intent to commit unlawful violence against a particular person or group. The First Amendment does not
Is satire or parody defamation, how do courts read a joke about a real person, and what does a linguist examine?
Satire about a real person is not defamation if a reasonable reader could not take it as stating actual facts. That is the test the D.C. CircuitExpert examinations
Ready to discuss your case?
Send the materials: the secretariat will confirm the format, cost, deadline and the specialist within one business day.

