FOCUS ON: GLOBAL BRANDS
eyes open, worldwide
K Translation, transliteration and phonetic rendering in trademarks
Rachel Gillard-Jones, of CPA’s Trade Marks Directory Service, puts forward a palette of language issues encountered when reviewing official trademark journals from all over the globe
T
he significant growth in the number of trademark applications filed worldwide, and the growing sophistication of applicants, consumers and would-be plagiarists, make it timely to look at this subject in some depth, in particular concerning various language issues. According to statistics provided by the World Intellectual Property Organisation, over the last 20 years there has been a marked increase in trademark applications all over the world. The reasons for this phenomenon lie in both economic and political changes. As a national economy develops into a consumer
In summary
K
Language and script mirror the historical, cultural, political and economic environment of a country, something that global trademark creation has not always taken into account. Trademark owners need to recognise the symbiosis between the development of language and the creation of a brand that will work
K
The use of dialects, jargons, terminology and local expressions, together with different writing systems and languages, makes the world of developing new trademarks and spotting infringements a very challenging and intriguing arena. Careful watching of key marks on a global basis is necessary, as is a thorough understanding of the languages, scripts and contexts of each market
K
As well as linguistic and script differences, political changes will impact the likelihood of infringement famous marks might encounter in different jurisdictions. The disintegration of the Soviet Union, for example, has led the publication of 15 trademark journals in a variety of languages and scripts, where previously just one journal was published by Soviet authorities. Elsewhere, the rise of the Chinese marketplace has fuelled an explosion in trademark registrations, complicated by the use of Chinese characters, rather than an alphabet, to render the marks
32 Trademark World #175 | March 2005
www.ipworldonline.com
FOCUS ON: GLOBAL BRANDS
entity, it becomes an attractive field for companies to apply for trademarks in that domestic market. The free market economy also promotes an awareness of the importance of the registration of trademarks overseas as the best way to protect them. The changes on the political map of the world have also caused a substantial increase in trademark applications. To get a clear idea about how significant the increase is, we can take a look at the number of trademark applications in China over the last 20 years. In 1983 there were 18,565 (17,000 by resident and 1565 by non-resident applicants) new trademark applications. Three years later this number more than doubled to 49,243 (43,445 by resident and 5,798 by non-resident applicants) and in 1993 it was as high as 123,323 (107,758 by resident and 24,565 by non-resident applicants) applications. According to the last available WIPO statistics, in 2001 there were 259,924 (229,775 by resident and 301,49 by non-resident applicants) new trademark applications. The sharpness of the increase is shown in figure 1. This trend is also evident in other countries around the world. In Greece in 1982 there were 5,425 trademark applications, a figure which more than doubled to 12,119 in 2001. A similar pattern emerged in the Benelux; in 1982 there were 9,289 applications soaring to
36,722 in 2001. Another factor, potentially accountable for the sharp increase in the number of trademark applications worldwide, lies in the political changes which affected Eastern Europe. With the disintegration of the Soviet Union, 15 new countries were formed and as a result 15 new trademark journals are now published in various languages and scripts. The Russian Federation, Ukraine and Republic of Belarus publish in Cyrillic, but the situation with other ex-Soviet Union countries is slightly more complex. Since the beginning of last century, official languages and scripts of the Republic of Kazakhstan, Kyrgyz Republic, Republic of
www.ipworldonline.com
Uzbekistan, Turkmenistan and Republic of Azerbaijan, all belonging to the group of Turkic languages, went through several script changes. These changes in script are reflected in the trademark applications today. Turkmenistan initially used Arabic script, in the 1920s Roman script, then Cyrillic. It is now reverting to Roman script and its official trademark journal is trilingual in Turkmen in Roman writing, Russian and English. The official journal of Tajikistan is published in Cyrillic Tajik, Russian and English. Kazakhstan, Kyrghyzstan and Uzbekistan publish their official trademark journal in Russian and their own respective languages. Both Georgia and Armenia abandoned the Cyrillic writing system and started publishing trademark journals only in their own languages, both countries have different writing systems to other known scripts. The Republics of Estonia, Latvia, Lithuania and Moldova publish their journals in Roman script. Trademark owners were obliged to refile their trademarks in the new countries. Trademarks owners were registering their marks in Russian (not realising that they needed to register their marks in the variant forms of Cyrillic such as Tajik Cyrillic in Tajikistan, or in case of Georgian and Armenian, in a completely different script). In Georgia trademark owners registered first their marks in Russian but then realised that they should register their marks in the Georgian script as well. This automatically increases costs and time spent on filing, translation and registration. At present, Georgian trademark office accepts bi-lingual applications and marks are often written both in Cyrillic and Georgian script or applications can be filed in any of these two languages. The Socialist Federal Republic of Yugoslavia published only one trademark journal but after the country’s breakdown, the independent Republics publish their own journals. The Former Yugoslav Republic of Macedonia, and Serbia & Montenegro, publish in Cyrillic whilst the other three countries, namely Republics of Slovenia and Croatia and Bosnia and Herzegovina publish in Roman script. In the early eighties, new trademark applications in Yugoslavia reached approximately 700 per annum, while now, Slovenia alone, the most prosperous country from ex-Yugoslavia, has more than 8,000 new trademark applications per year. The expanded use of different writing systems in the official publications of trademark applications is a barrier to
GLOSSARY
Translation is an act of expressing meaning in another language. Transliteration is the transcription into corresponding letters of another alphabet. Example: GREEK ALPHABET: TRANSLITERATION: portokáli TRANSLATION: orange
Phonetic rendering is the indication of pronunciation of words written in foreign scripts. The sounds in some languages are far from obvious just from the letters and only a language expert can recognise the words in question. There is no simple relationship between the letters or characters and the sounds indicated, and there is sometimes more then one possibility.
Example: WORD IN ENGLISH: “Apple” PHONETIC RENDERING: (katakana) (hiragana) TRANSLATION INTO JAPANESE: “ringo” (kanji) (hiragana) (katakana)
Syllabary is a phonetic writing system consisting of symbols representing syllables.
Trademark World # 1 7 5 | March 2005 33
FOCUS ON: GLOBAL BRANDS
trademark owners finding potential infringement by third parties. There is a greater reliance on a professional watching service to detect these marks in foreign languages. Watching is an essential part of the armoury of trademark owners. Watching services are now faced with more and more trademarks rendered in the different scripts. Unscrupulous third parties are using transliterations, phonetic renderings and translations to disguise the infringement. In order to combat this trend, watching companies need to take on board more language specialists who are able to recognise numerous scripts and various shades of meanings. It’s all Greek The Greek language and alphabet have a long and complex relationship with the languages and script of Western Europe. Trademarks filed in the Greek or Cypriot journals may contain: Greek words in Greek script, Greek words in Roman script (often in inconsistent transliterations) and foreign words in Greek script. In addition, since many upper-case letters have the same shape but different meanings in the Greek and Roman alphabets, it is not always apparent which script a mark is in, and both have to be reviewed. For example, the letter H (= H or E,I); B (= B or V); P (= P or R); Y (= Y or I,F,V – according to circumstances). On some occasions it is the meaning of Greek words that is at issue. , which is transliterated as ‘Ekdosis Asti’, and means ‘City Editions’, where ‘ASTY’ is one of the words for ‘city’ in Greek. Expressing a trademark is more than translating a mark into a foreign language. A trademark owner will consider the richness of the language of the country of registration, using its etymological heritage to select the best rendering of a trademark closest to its meaning and brand value. Where transliterations are concerned, there are several complications. Western European languages contain many words based on Ancient Greek roots, such as biography, gynaecology, telemetry. The spelling used in such words represents Ancient Greek pronunciation; the modern pronunciation of the Greek forms of these words, , , would be viografia, yinekoloyia, and tilemetria respectively. There are particularly clear issues with the letters and (upper case/lower case). In Ancient Greek, _/_ is b and _/_ is e; in the modern pronunciation, they sound like v and i respectively. Andrew Woode, TMDS’ Greek language
34 Trademark World #175 | March 2005
specialist, explains that the choice of which set of equivalents is most appropriate for watching marks or registering new marks in Greek script depends on the context. A technical word using Greek roots will be most appropriately transliterated – in either direction – using the Ancient Greek equivalents. However a modern word, such as English big or Greek (dolmas) would be rendered according to the modern pronunciation. An obvious example is the mark transliterates to ‘Bravo Kalos Ilthate’ in the Modern Greek. The first word can be phonetically rendered as ‘Bravo’ while the combination of the meanings of the second and third words make ‘welcome’ : kalos = well and Ilthate = you came, making ‘welcome’.
The mark is written in a mix of alphabets. There is an obvious use of English ‘easy’, but the whole word is a play on Greek which transliterates to asfalisi and means ‘insurance’, giving a meaning issue which could be important for a trademark owner. Cyrillic imperialism The modern versions of the Cyrillic alphabet, originally used primarily for the Old Church Slavonic, are the standard orthographies for a considerable number of languages. The largest community of speakers is provided by Russians, but other Slav languages written in it include Ukrainian, Belarussian, Bulgarian, Macedonian and Serbian. Under the influence of the above nations (especially Russia) non-
What the brand owners say
Bob Boad, BP: “We do have some registrations of our marks in local scripts eg in Cyrillic, Thai, Katakana and Chinese and in these cases we are guided by our local marketing and other business colleagues in the selection of the exact characters. However, I am not sure that we have any cases where we have registered both Russian Cyrillic and a local Cyrillic but if there was a business need we would certainly do so. Obviously, the cost would be increased but protecting the signs that serve to identify a product or service in the local market place is the critical thing. Both English (Latin script) and local equivalents are commonly registered by us for a particular mark. We also register labels/logos in both versions. The local version is the one the local population recognises and identifies with. We have had infringements where the local version has been copied as well as the English/Latin version. I can think of cases in Kuwait many years ago where we may be able to get permission for details even photos to be used given the passage of time and the fact that it involves a now redundant mark. Infringements would be picked up in the market by our local staff or distributors etc and conflicting TM applications for opposition would be picked up by local TM agents as well as possibly by Trade Marks Directory Service.” Julie Allen, Brown-Forman Corp: “Translations and transliterations are extremely important as trademarks are considered equivalent if meaning is the same, no matter the language. I can’t recall having seen any phonetic renderings, but they would be useful as well. As our company is a global one, we are interested in all languages and scripts in the world, in particular when it comes to worldwide recognised marks such as JACK DANIEL’S and BOLLA.” www.ipworldonline.com
FOCUS ON: GLOBAL BRANDS
Slav languages have also been written in the Cyrillic alphabet. Romanian used it until the 19th century, and it was standard in Moldova for the local version of Romanian in Soviet times. The Soviet Union under Stalin used the Cyrillic alphabet for most of the new orthographies it introduced for minority languages, often replacing earlier uses of the Arabic and/or Roman alphabets. The Central Asian and Caucasian Turkic languages Azeri, Turkmen, Uzbek, Kyrghyz and Kazakh were affected by this, as was Tajik (which is related to Farsi and Dari). Outside the Soviet Union, but heavily influenced by it, Cyrillic was introduced to write Mongolian in place of its own indigenous vertical script which was still retained in Chinese Inner Mongolia. The old script is now being reintroduced in Mongolia itself, but the trademark journal is still in Cyrillic. Like the Roman alphabet, the Cyrillic alphabet has been used to write languages with divergent sound systems. Unlike the Roman alphabet, which often tends to keep the same 26 letters supplemented by various diacritics as appropriate, Cyrillic orthographies add and subtract letters to do the same task. Thus Kazakh and Mongolian use and to represent the sounds of German ö and ü (or French eu and u) which are absent in Slav languages. Serbian has the consonants and representing sounds absent in Russian, while it uses (lj nj) for sounds which Russian possesses but are spelt using different conventions. In addition, the same letter-shape can be used to represent different sounds, which means that the transliterations have to vary according to the language in question. In Russian, is pronounced g, but in Ukrainian (where it generally occurs in the same words), the pronunciation is h; the relatively rare g is written . Transliterations are also subject to traditions. For Russian, an ISO standard exists but is in very restricted use, and most Western European languages transcribe it according to the spelling conventions of their own languages. Thus is rendered shch in popular English transcriptions, schtsch in German, chtch in French and in Czech. By contrast, Serbian is usually written in Cyrillic but also has a standard spelling in Roman script (using the same conventions used routinely by Croatian and Bosnian). So while Russian and Serbian sound identical and mean the same thing (‘our’), one would transcribe the first as nash and the second as . In the opposite direction, transliterations
www.ipworldonline.com
into Cyrillic from Western languages are usually done on the basis of the sound. This means that for instance, ‘eu’ in English, French, German and Italian (such as in a word beginning feud-) might go into Russian as , and (yu, yo, oy and eu) respectively according to the different pronunciations. However if the word was recognised as of classical Greek origin, such as a word beginning euro-, Russian might use the historical equivalent (yev). This means that applicants seeking to register their marks in the Cyrillic world cannot rely on using one standard-issue transliteration, but need to use different versions in different markets. A hypothetical pharmaceutical brand cytocardiahemia would probably be transcribed in Serbian but might be in Russian. In a real-world example, the Russian trademark application , phonetically transcribed as ‘master sheff ’, is an obvious clash with the words ‘Master Chef ’, which would be of great interest to the owners of that mark.
This means that applicants seeking to register their marks in the Cyrillic world cannot rely on using one standard-issue transliteration, but need to use different versions in different markets
Arabic challenges Richard Aitken, TMDS’ specialist for the Arabic language, points out some problems that he encounters when watching trademark journals in Arabic. In Arabic script the ‘short’ vowels are completely left out but ‘long’ vowel sounds count as consonants. The fact that ‘short’ vowels are completely left out can affect the sound of the word without this being detectable on the written page. In addition to this there is no letter ‘p’ in Arabic and it is normally written as ‘b’ – so that in effect ‘p’ and ‘b’ are interchangeable and this can cause problems when trying to detect conflicting marks with a stress on the aforementioned letter. An example from the Saudi Arabian journal (‘BRW’/’PRW’) could be pronounced ‘BREW’, ‘PRU’, ‘BROW’, ‘PROW’, ‘PERRAULT’, ‘BARRAULT’, ‘BERRO’,
‘PERRO’, ‘PRO’ etc. In such a case the trademark watcher when identifying possible infringements has to rely on context as well as the linguistic issues. In the trademark (‘Formula One’) the letter occurs three times, each time with a different function; the first time it is the ‘o’ of ‘Formula’, the second time it is the ‘u’ of ‘Formula’, the third time it is the ‘w’ of ‘Wan’. (The ‘One’ in ‘Formula One’ is written out as ‘Wan’.) In the example of the trademark application from the Sudanese journal (‘Warda’) the meaning of the word is ‘Rose’ and it is the main point of the trademark watcher’s concerns. Furthermore, in the example from the Libyan journal (‘Al-Waaha’) the meaning of the words is ‘The Oasis’ and this mark is of obvious interest to the owner of an identical trademark. In the 28-letter Arabic alphabet there are, from a western viewpoint, two letters ‘t’, two letters ‘d’, two letters ‘s’, two letters ‘th’ and two letters ‘h’; the letter can sometimes be a hard ‘g’ and sometimes ‘j’. Thus the trademark application from the Kuwaiti journal would sound something like ‘Gooda’ or ‘Gouda’ in the Egyptian pronunciation of Arabic but something like ‘Juda’ in the Saudi Arabian pronunciation. The letters and have no exact equivalents in English. In some regional pronunciations the letters and sound identical to speakers of western languages, but in other regional pronunciations, particularly in Saudi Arabia and the Gulf States, and the completely different letter sound identical. Chinese characters According to Michael Rank, TMDS Chinese language specialist, trademarks written in Chinese present special problems and challenges. This is because Chinese does not use an alphabet but many thousands of characters, each of which has a meaning and a pronunciation. Sometimes characters used in trademarks are used only for their pronunciation, and the meaning is irrelevant. This occurs when the character is used to imitate English or other foreign sounds. Some characters are used almost only for their sound rather than meaning, and use of these in a trademark signals that the word is of foreign origin. For example , pronounced ‘ri xing’, means ‘sun star’, while ‘hong pingguo’ means ‘red apple’, however ‘ha li bo te’ makes no sense as a string of characters but is used to imitate the sound of the English word ‘halibut’. In addition, the characters used on the mainland (People’s Republic) and Taiwan are
Trademark World # 1 7 5 | March 2005 35
FOCUS ON: GLOBAL BRANDS
not the same. In mainland China simplified characters, first introduced in the late 1950s, are generally used, while in Taiwan only traditional characters are used. Simplified characters are quicker to write, although they represent only a minority of the total. It is impossible to transcribe exactly the sound of foreign word because of the syllable sound structure of standard Chinese, or Mandarin. All syllables end in a vowel, n, ng or occasionally r and never end in sound like p, t, k or g. This makes rendering a word like ‘brilliant’ in Chinese rather difficult. One could find characters which imitate the sound and write it something like ‘bu li lian te’, but it would be difficult or impossible to find characters to reflect the meaning. The alternative would be to use characters which mean ‘brilliant’ even if they don’t sound anything like the English word. Thus ‘brilliant’ could be written which is pronounced ‘guangming’ and is a typical Chinese trademark. It is vital to consider these language intricacies when applying for, or when watching a trademark. Japanese scripts There are similar concerns with trademark applications in Japanese as Charles Rowe, TMDS’ Japanese language expert, explains. “The Japanese language uses three scripts – kanji or Chinese characters, which it shares with Chinese and Korean, and the two syllabaries hiragana and katakana. Thus, the name ‘Takarajima’ (‘Treasure Island’) could be written in three different ways: , (hiragana),or (katakana). In principle the hiragana syllabary is used for writing native Japanese words (including loanwords from Chinese), while katakana is used to write words from languages other than Chinese, Japanese & Korean languages such as English or French. The English ‘Treasure Island’ would therefore normally be written in katakana as , but might, for special effect, be written in hiragana as .” – explains Charles Rowe and adds – “Writing a word in hiragana
gives it a softer, more ‘elegant’ look than katakana, suitable, say, for a brand of cosmetics, while katakana has a bolder effect, more appropriate for a computer action game, perhaps”. Since Japanese shares Chinese characters with the Chinese language, a Chinese mark (‘Baodao’ in Chinese) would in its written form be identical with the Japanese ‘Takarajima’, although totally different in pronunciation. Trademarks in Chinese characters are found throughout the world, particularly in South-East Asian countries, and it is important to be able to recognise both the Chinese and Japanese readings of such marks, as the same characters are pronounced differently. The English word ‘glen’, for example, is rendered in Chinese by the characters , which to a Japanese speaker would be read as ‘kakuran’ and would be unrecognisable as the word ‘glen’. Thus, when Chinese characters are used for phonetic rendering rather than for their meaning, a mark might be an infringement in one language but not in the other. When they are used for their meaning, however, the written characters are frequently identical, as in the ‘Treasure Island’ example above. An application in Indonesia for the mark was published with a note giving the Chinese reading ‘ju che liang phing’, whereas in fact this is a Japanese name with the pronunciation ‘yu no ryohin’, similar to the name of a famous Japanese store chain. The use of different scripts in Japanese, especially the availability of Chinese characters which have meanings as well as sounds, gives ample scope for wordplay, which happens often in the applications for new trademarks. For example, the Japanese (rakuda – ‘it’s easy’) has the same sound as the word for ‘camel’ . Scripts and languages are freely mixed within the same word, as in the example (kirakuzeshon), which consists of the Japanese word ‘kiraku’ (meaning ‘free and easy’) in kanji and the second half of the English word ‘relaxation’ in katakana, and the practised eye would spot this as a pun on the English word.
Other scripts and languages Other languages such as Armenian, Thai and Korean also use different and unique scripts and a thorough knowledge is needed in order to transliterate and translate them. Even if published in English, the Indian Trademark journal also brings with it a few problems for trademark watchers. Sometimes the mark is written in one of the many scripts of Indian subcontinent (Hindi, Malayalam, Urdu, Gujarati, Bengali, Tamil, Sinhalese, Telegu, Gurmukhi etc) and in order to detect a potentially conflicting mark it is important to have a specialist who can transliterate from all these scripts. Trademark owners should review their portfolios to ensure that their marks are registered appropriately in the countries of registration. For example, a registration in Georgian script for Georgia. It is also important to advise the watching company on how the mark has been registered as the characters used may be different to what would be assumed. A trademark owner may want to review what characters have been used to render their marks in these scripts as some characters can have some bad connotations and lead to the damage to the brand in that country. More then just a language Every country in the world has its own language idiosyncrasies that trademark watchers and translators are encountering every day. A computer system - no matter how advanced it is - cannot replace human knowledge regarding various aspects, shades and idiosyncrasies of the complexity of a language and a script. Language and script mirror the historical, cultural, political and economic environment of a country. Perhaps global trademark creation has not always taken this sufficiently into account. Trademark owners need to recognise the symbiosis that there is between the development of language and the creation of a brand that will work. The use of dialects, jargons, terminology and local expressions, together with different writing systems and languages, makes the world of developing new trademarks and spotting infringements a very challenging and intriguing arena.
K
About the author Rachel Gillard-Jones is the manager of Trade Marks Directory Service, a leading company in trademark watching since 1942. With its in-house expertise, TMDS has the capability to translate, transliterate and establish phonetic rendering for the languages used in all IP jurisdictions. Ms Gillard-Jones joined TMDS in 1990 as a linguist and translator of German, Spanish, French and Swedish, she was made assistant manager of the business in 1994 and now manages TMDS. TMDS’ language specialists, Milka Sculac Sennett, Richard Aitken, Michael Rank and Charles Rowe and Andrew Woode, also contributed to the development of this article.
36 Trademark World #175 | March 2005
www.ipworldonline.com