Trademark Strategy for Vernacular-First Brands: Navigating India's Multilingual Mark Space

Introduction

India’s approach to branding has become different. A decade back, the rule was to start in English and then localize. Now, brands starting D2C or regional first take the opposite route, choosing a Hindi name, a Devanagari-Roman combination logo, or even building the mark to be flexible with scripts from the beginning. While this may be seen as only a marketing decision, it places the brand firmly within an area that remains largely unaddressed in India’s trademark law, the question of how a mark with altered script or sound translates between languages.

The Legal Baseline: Deceptive Similarity

All depends upon Section 2(1)(h) of the Trade Marks Act, 1999, according to which a mark can be said to be “deceptively similar” to another if it bears such close resemblance that there is likelihood of deception or confusion. All other provisions like sections 9, 11, 29 and 34, depend upon different phrases used, but all are construed as synonymous by the Courts. In Amritdhara Pharmacy v. Satya Deo Gupta, Supreme Court held that there is no hard and fast rule for determining deceptive similarity and it entirely depends upon the impression made upon an ordinary customer with imperfect memory. Multi-factor test as laid down in Cadila Health Care Ltd. v. Cadila Pharmaceuticals Ltd. is applied even today - nature of the marks, the phonetic, visual and conceptual similarity, the nature of the goods, mode of trade and, most importantly, class of consumers, their education and language of understanding.

The myth of founders about the uniqueness of the Hindi/indigenous language mark just because it is unfamiliar to an English trained examiner is prevalent. This is not so from a legal perspective, where generic or descriptive words in Hindi are taken as generic English words. The word “Doodh” for the dairy product bears the same risk as “milk” does since the Indian consumer knows its meaning. Section 9 holds equally irrespective of the script, and the examiners now transliterate and translate the applied mark to understand whether it is conceptually descriptive.

Translation Is Not a Loophole

The oldest thread in Indian case law is that translating a mark does not create a new mark for infringement purposes. This traces to J.C. Eno Ltd. v. Vishnu Chemical Co. (AIR 1941 Bom 3), where the defendant sold saline under the Marathi name “Falaxar,” fala and kshar being the ordinary Marathi words for “fruit” and “salt,” against Eno’s established “Fruit Salt” mark. The court restrained the defendant, reasoning that a trader who has acquired an exclusive right to a mark in one Indian language can restrain a competitor from copying it via translation into another, given how many Indians know several languages.

The same principle barred registration in T.G. Balaji Chettiar v. Hindustan Lever Ltd., where a Tamil "Surian" soap mark was refused as deceptively similar to prior "Sun" mark, both built on the idea of the sun, and in Bhatia Plastics v. Peacock Industries Ltd., where "Peacock" was restrained as the semantic equivalent of the prior "Mayur" mark for plastic goods, even though the words share no sound or script. This doctrine protects the idea behind a mark, not merely its surface form, and it remains active: in February 2026, the Delhi High Court in Victoria Foods Private Limited v. Ashad Trading Co., granted an ex parte ad interim injunction after finding a prima facie case that "CAPITAL," used as the English translation of the registered "RAJDHANI" mark for food products, infringed the plaintiff's trademark.

Transliteration Is the Newer, Sharper Problem

Although translation and transliteration may be used interchangeably in common parlance, they remain separate concepts under the law. While translation refers to the use of the same meaning in two languages, transliteration is the representation of the same name by using a different script without altering its pronunciation - such as Devanagari, Tamil, or Roman scripts. This problem arises in the case of brands operating across different scripts, since the Indian statutes have failed to make a separate doctrinal classification regarding transliteration, and rather incorporate it within normal phonetic similarity test under Sections 9(2)(a) and 11(1).

In the case of Hitachi Ltd. v. Ajay Kr. Agarwal, the mark “HITAISHI” in Roman and Devanagari scripts was held to be phonetically similar to the already registered trademark “HITACHI” and restrained. This should be read alongside Rule 28 of the Trade Marks Rules, 2017, which requires that any mark containing words or numbers in a script other than Hindi or English be accompanied by a precise transliteration and translation of each such word or number, along with a statement of the language it belongs to, the message is that registering only one script version of a name does not insulate the other versions; a same-sounding mark in a different script is not a separate asset to defend later.

Phonetic Similarity as the Dominant Axis

Phonetic similarity has now been recognized as the most crucial factor in Indian trademark disputes due to the country's linguistic diversity and varying levels of literacy amongst its consumers. In the case of Consitex S.A. v. Ms. Kamini Jain & Ors., the Delhi High Court relied upon documentary evidence and an investigator's affidavit, filed in support of the interim application to conclude that "ZEGNA" sounds like "ZEN-YAH" in the Indian marketplace, and accordingly held "JENYA" to be deceptively similar, along with the fact that the plaintiffs had already independently registered the Hindi transliteration of the mark.

Conceptual similarity alone has been found to be enough to constitute deceptive similarity even without any phonetic similarity: in Surya Roshni Ltd. v. Electronic Sound Components Co. (1994), "BHASKAR" along with a depiction of a sun was found to be deceptively similar to "SURYA," since both the wordmark and the device were well-known conceptual substitutes for the word "sun." This case relied on the precedent set in Balaji Chettiar, and the very case has been followed in Bhatia Plastics v. Peacock Industries (later that year) when "PEACOCK" and "MAYUR" were considered to be conceptually similar.

Why the Domain Layer Matters Too

Vernacular branding does not remain confined to the online space in India. Indian courts apply the same principle regarding deceptive similarity to domain names as a mark. In the Satyam Infoway Ltd. v. Siffynet Solutions, the rationale for the Supreme Court to regard a domain name as having greater protection than any ordinary trademark is that a domain name, since it can be accessed by everyone using the internet, demands global exclusivity unlike a mark, which can easily coexist in different regions and even products.

Another issue, specific to the Indian languages, is that of internationalized domain names, in which the languages used can create confusion between themselves like Tamil and Malayalam.

Search and Enforcement Are Catching Up, Unevenly

The practical constraint was always going to be search. The transliteration rule in Rule 28 allows the Registry to phonetically index non-English marks, and Registry tools already pick up phonetic matches among different script marks. But Devanagari, Tamil, Kannada, Malayalam, Bengali, and other Brahmi-derived scripts organize their alphabets by place and manner of articulation rather than by convention, so marks which are phonetically equivalent across scripts have structural equivalencies even when the marks look completely different.

An approach to computational phonetics from IIIT Hyderabad takes advantage of this fact by giving each letter a feature vector (place and manner of articulation, voiced/unvoiced, aspirated, length), and using a step-wise distance metric to score word similarities, identifying matches that would be missed by simple search criteria like p/b, ch/j, and r/l substitutions that consumers hear but would not be caught by a mere match. Peer-reviewed research on assessing trademark similarity, which considers visual, semantic, and phonetic components using a fuzzy logic engine, shows statistically significant increases in accuracy compared to any one of those measures in a large database of actual dispute cases.

The examination procedure of India is gradually turning toward this as well, as reports have shown that there are applications of machine learning and NLP to help the examiners make comparisons regarding the phonetics and concepts in the case of Hindi, Tamil, Bengali, and other regional languages, due to the growing backlog of applications. However, there are limitations to these systems; the systems may not be able to take into account the cultural considerations, which the human examiner may pick up immediately, the peacock’s connection with pride being the typical example, and they might bring in data bias or opacity, which does not gel well with a test requiring reasonable decision-making.

Conclusion

First vernacular brand strategy is not a means of coping with the normal trials of trademark law in India; issues like descriptiveness, distinctiveness, and deceptive similarities remain relevant with equal intensity regardless of the script used. What is different is the approach of analysis, Indian courts have always focused on the ear rather than the eye, considering phonetic and conceptual similarities across scripts and languages an essential part of the confusion test.

Author: Annika Tony, in case of any queries please contact/write back to us via email to chhavi@khuranaandkhurana.com or at  Khurana & Khurana, Advocates and IP Attorney.

References

1. Trade Marks Act, 1999, Sections 2(1)(h), 9, 11, 29, 34 (India)

2. Trade Marks Rules, 2017, Rule 28.

3. Amritdhara Pharmacy v. Satyadeo Gupta, AIR 1963 SC 449 (India).

4. Cadila Health Care Ltd. v. Cadila Pharmaceuticals Ltd., AIR 2001 SC 1952 (India).

5. J.C. Eno Ltd. v. Vishnu Chemical Co., AIR 1941 Bom 3 (India).

6. T.G. Balaji Chettiar v. Hindustan Lever Ltd., AIR 1967 Mad 148 (India).

7. Bhatia Plastics v. Peacock Industries Ltd., AIR 1995 Del 144 (India).

8. Victoria Foods Priv. Ltd. v. Ashad Trading Co., CS (COMM) 135/2026 (Del. H.C. Feb. 12, 2026) (India)

9. Hitachi Ltd. v. Ajay Kr. Agarwal, 59 (1995) DLT 213 (Del.) (India).

10. Consitex S.A. v. Kamini Jain, 2011 (47) PTC 337 (Del.) (India).

11. Surya Roshni Ltd. v. Electronic Sound Components Co., AIR 1995 Del 92 (India).

12. Satyam Infoway Ltd. v. Sifynet Solutions Pvt. Ltd., (2004) 6 SCC 145 (India).

13. Bisman Kaur, Protection of Translations and Transliterations in India, IAM Brands in the Boardroom 75, 75 (2009), https://www.remfry.com/wp-content/uploads/2018/02/protection-of-translations-and-transliterations-in-india1.pdf.

14. Nikita Banerjee, Transliteration Conflicts for Indian-Language Trademarks, 7 Ind. J.L. & Legal Rsch. 7702 (2023), https://3fdef50c-add3-4615-a675-a91741bcb5c0.usrfiles.com/ugd/3fdef5_1e536bed5221496cba3b9e1579968ad7.pdf.

15. Sankalp Jain, Deceptive Similarity of Trade Marks in India (Sept. 8, 2020), https://ssrn.com/abstract=3902406.

16. Rossitza Setchi & Fatahiyah Mohd Anuar, Multi-faceted Assessment of Trademark Similarity, 65 Experts Sys. Applications 16 (2016), https://doi.org/10.1016/j.eswa.2016.08.028.

17. Anita H. Pathak & Anjula S. Chowbe, Critical Analysis of AI's Impact on Trademark Examination and Its Role in Preventing the Registration of Deceptively Similar Marks in India, 26 SEEJPH 2177 (2025), https://doi.org/10.70135/seejph.vi.4103.

18. Anil Kumar Singh, A Computational Phonetic Model for Indian Language Scripts, Language Technologies Research Centre, IIIT Hyderabad (2006), https://cdn.iiit.ac.in/cdn/ltrc.iiit.ac.in/anil/papers/cpms-long-iwlc-06.pdf.