1 Within the rich tapestry of Singapore’s culinary culture, cheese holds a significant place not only as a delectable food item that may be enjoyed in and of itself, but also as an essential ingredient of many staple dishes in the Singapore diet. In the context of the Geographical Indications Act 2014 (Act 19 of 2014) (the “GIA”), the protection of geographical indications (“GIs” or “GI”) safeguards the interests of Singapore’s consumers by providing greater assurance that food products truly carry the characteristics that they are known for, and which are attributable to their geographical origin. This is important not least because the Singapore marketplace, nestled as it is at the crossroads of international trade and commerce, offers the average consumer an abundance of choices. Consumers must therefore be able to rely on GIs – being indications used in trade to identify goods as originating from a certain place – to convey accurate information about the geographical origins of the product in question and the given quality or reputation of the product that is essentially attributable to that place. The registration of a GI should not, however, prevent fair and established competition from products of a similar nature which have their origins outside of the registered geographical area.

2 The present case concerns a request for a qualification of rights under s 46(1)(b) read with s 46(2)(b) of the GIA, namely, a request that the protection of the GI “Parmigiano Reggiano” should not extend to the use of the term “Parmesan”. The ground of the request is that “Parmesan” is not a translation of “Parmigiano Reggiano”.

3 One notable feature of the GIA is that translations of a GI can also be protected by default. In the case of GIs identifying wines or spirits, the protection of translations finds its basis in s 4(2)(c) and (2)(d) of the GIA, which set out the “special non-confusion based protection reserved for GIs identifying wines or spirits”: Ng-Loy Wee Loon, Law of Intellectual Property of Singapore (Sweet & Maxwell, 3rd Ed, 2021) (“Law of Intellectual Property of Singapore”) at para 28.3.5. This protection is afforded to all such GIs regardless of whether the GI has been registered under the GIA: Law of Intellectual Property of Singapore at para 28.3.19. In the case of GIs identifying agricultural products (other than wines and spirits), such as in the present case, the protection of translations finds its basis in s 4(6) of the GIA. Importantly, such protection is available only to registered GIs:

4 Therefore, s 4(6) “extends the special non-confusion-based protection previously reserved for GIs identifying wines and spirits to registered GIs identifying the other specified categories of agricultural products” [emphasis added]: Law of Intellectual Property of Singapore at para 28.3.20. The reason for this was explained by the Senior Minister of State for Education and Law at the Second Reading of the Geographical Indications Bill 2014 (Singapore Parl Debates; Vol 91, Sitting No 15; [14 April 2014] (Indranee Rajah, Senior Minister of State for Education and Law)):

5 Notably, applicants for the registration of a GI (also referred to under the GIA as the “registrant”) do not need to specify all the translations which they wish to protect at the outset, ie, at the time of registration. Under the scheme and design of the GIA, the determination of whether a term is a translation of a GI falls to be undertaken on a case-by-case basis, either at the time when infringement proceedings are brought under s 4 of the GIA, or (in the case of a registered GI) upon a request for a qualification of rights under s 46 of the GIA: see, Public Consultation on Changes to be made to the Geographical Indications Act and the Trade Marks Act to Enhance Singapore’s Regime for the Protection of Geographical Indications, prepared by the Intellectual Property Office of Singapore on 1 November 2013 (“IPOS Public Consultation”) at paras 4.4 and 4.15.

6 The appellant is Fonterra Brands (Singapore) Pte Ltd (“Fonterra”), a wholly owned subsidiary of a New Zealand-based co-operative company owned by 10,000 dairy farmers in New Zealand. Fonterra claims to be the world’s largest dairy exporter, and is involved in the collection, manufacture and sale of milk and milk-derived products. These products include cheese sold by Fonterra under the “Perfect Italiano” trade mark in Singapore.

7 The respondent is the Consorzio del Formaggio Parmigiano Reggiano (the “Consorzio”). The Consorzio is a voluntary consortium of Parmigiano Reggiano cheese producers which has taken the mantle of a similar organisation established in Italy in 1934. It is tasked by the Italian Ministry of Agricultural, Food and Forestry Policies with the protection, promotion, enhancement, consumer information and general care of the interests relating to Parmigiano Reggiano cheese. The Consorzio is the holder of “Parmigiano Reggiano” as a protected designation of origin under European Union law.

8 In Singapore, the Consorzio is the registrant of the registered GI No. 50201900057U for “Parmigiano Reggiano” for cheese originating from the specified region of “the part of the province of Bologna to the left of the River Reno, the part of the province of Mantua to the right of the River Po, and the provinces of Modena, Parma and Reggio Emilia” in Italy (the “Specified Region”). The GI was formally registered on 22 June 2019.

9 On 16 September 2019, after the GI was successfully registered, Fonterra filed a request to qualify the rights conferred in respect of the GI “Parmigiano Reggiano” under s 46(1)(b) read with s 46(2)(b) of the GIA on the basis that the term “Parmesan” is not a translation of “Parmigiano Reggiano” (the “Request”). The Request sought to enter the following qualification in the register of geographical indications:

10 In the two-page statement of ground(s) accompanying the Request (“Fonterra’s Statement of Grounds”), Fonterra asserted that “Parmigiano Reggiano” cheese is “from the province of Bologna to the left of the River Reno, the part of the province of Mantua to the right of the River Po, and the provinces of Modena, Parma and Reggio Emilia, whereas ‘Parmesan’ cheese can be made anywhere” [emphasis added]. Furthermore, “[n]ot only does Parmesan cheese need not originate from Italy, ‘Parmesan’ cheese is also not regulated in the same manner as ‘Parmigiano Reggiano’ cheese and they differ in terms of its milk content, regulations, taste, colour and texture”.

11 On 12 November 2019, the Registrar of Geographical Indications (the “Registrar”) proposed to allow the Request and published the proposed qualification of rights in the Geographical Indications Journal No. 13/2019 for opposition purposes in accordance with r 40(4) of the Geographical Indication Rules 2019 (the “GIR”). It is not disputed that no evidence accompanied Fonterra’s Statement of Grounds. It also appears that the Registrar did not, at any time prior to publishing the proposed qualification of rights in the Geographical Indications Journal, require Fonterra to furnish evidence in respect of the Request pursuant to r 40(2) of the GIR.

12 The Consorzio filed an opposition to the Request on 12 March 2020 (the “Opposition”), pursuant to r 41(1) of the GIR. The statement of ground(s) accompanying the notice of opposition stated that the Consorzio “oppose[d] the proposed qualification of rights on the basis that the ground upon which [Fonterra] relies – section 46(2)(b) – is not established”, and denied the Fonterra’s assertions in its own statement of grounds.

13 The learned Principal Assistant Registrar (the “PAR”) heard the Opposition on 15 March 2022. She allowed the Opposition and rejected the Request: see Consorzio del Formaggio Parmigiano-Reggiano v Fonterra Brands (Singapore) Pte Ltd [2022] SGIPOS 11. The PAR held that the legal burden of proof in establishing whether a particular term was a translation of a GI lay on the party opposing the qualification of rights, which was the Consorzio in this case. As to the meaning of “translation” in s 46(1)(b) of the GIA, the PAR took the view that what was required was a translation of the GI as a whole. The question therefore was whether “Parmigiano Reggiano” directly translated to “Parmesan”. The PAR also held that all that was required was a translation of the GI – there was no requirement that the possible translation be the only translation. Further, there was nothing in s 46 of the GIA which required that the translation of the GI be an English translation. Finally, based on a purposive interpretation of s 46(2)(b) of the GIA, “translation” simply referred to the question of whether words had the same meaning in a different language.

14 The PAR concluded, after examining the evidence before her, that the Consorzio had discharged its burden and shown that “Parmesan” was indeed a translation of “Parmigiano Reggiano”. Central to the PAR’s reasoning was that various dictionary entries reflected the understanding of the consumer in the local context and showed that “Parmesan” was indeed a translation of “Parmigiano Reggiano”. Among other things, the PAR considered that an extract from the Collins Dictionary was persuasive because Singapore adopts British English as our official working language.

15 On appeal in HC/TA 8/2022 (“TA 8”), the learned judge of the General Division of the High Court (the “Judge”) held that “Parmesan” was a translation of the GI “Parmigiano Reggiano” and dismissed Fonterra’s appeal against the PAR’s decision: Fonterra Brands (Singapore) Pte Ltd v Consorzio del Formaggio Parmigiano Reggiano [2023] SGHC 77 (“Parmesan (HC)”) at [90].

16 The Judge agreed with the PAR, in relation to the burden of proof where a proposed qualification of rights was opposed, that the burden should lie on the party opposing the qualification of rights, since it was this party (in this case, the Consorzio) which sought to persuade the court that its rights ought to be extended in respect of a specific translation: Parmesan (HC) at [40] and [47].

17 Before the Judge, the parties’ positions diverged on whether a “faithful” or a “strict literal” approach to translation ought to be taken, and whether expert evidence was required: Parmesan (HC) at [55]. The Judge ruled that a faithful translation which captures the meaning of the words in question should be adopted for the purposes of the GIA. The Judge considered that an insistence on word-for-word translations may result in inaccuracies in meaning, which may either allow protection to be extended in an unprincipled manner or unjustifiably restrict protection altogether: Parmesan (HC) at [59] and [61]. The Judge also ruled that the court did not require expert evidence to make a finding on whether “Parmesan” was a translation of “Parmigiano Reggiano”: Parmesan (HC) at [64] and [65]. On the role of consumer perception in the translation inquiry, the Judge ruled that consumer perception was generally irrelevant, save to the extent where such perception was reflected in extracts taken from reputable dictionaries: Parmesan (HC) at [72] and [77]. Furthermore, the Judge observed that the ground pertaining to the potential genericism of the GI was not pleaded by Fonterra: Parmesan (HC) at [51] and [77].

18 The Judge, taking into consideration the probative value of three dictionary extracts as a whole, concluded that the Consorzio had fulfilled its burden of proving that “Parmesan” was indeed a translation of the term “Parmigiano Reggiano”: Parmesan (HC) at [86]. The Judge further considered that, even if consumer perception and marketplace evidence were relevant to the translation inquiry, Fonterra had not provided concrete proof of consumer perception, through means such as consumer surveys or other relevant market data, to show that consumers perceived “Parmesan” and “Parmigiano Reggiano” as referring to different products: Parmesan (HC) at [89].

19 Fonterra now appeals against the whole of the Judge’s decision. At the hearing of the appeal, counsel for Fonterra, Mr Chan Wenqiang (“Mr Chan”), confirmed that Fonterra was not insisting on the position that expert evidence from a translator would always be required. While there is no cross-appeal, the Consorzio submits that the Judge’s finding that it is the party opposing the qualification of rights that bears the burden of proving the ground in s 46(2)(b) of the GIA should be varied. The Consorzio submits that the Judge’s decision should be affirmed based on the grounds in the judgment and on grounds other than those relied upon in the judgment, including certain other dictionary extracts.

20 In the hearing of this appeal, we were greatly assisted by the views of Professor Gordon Ionwy David Llewelyn (“Prof Llewelyn”), who was appointed as Independent Counsel.

21 The issues raised on appeal are:

(a) Which party bears the burden of proof where a proposed qualification of rights is opposed?

(b) What is the proper approach to the meaning of a “translation” under s 46(1)(b) read with s 46(2)(b) of the GIA?

(c) Applying this approach, is “Parmesan” a translation of “Parmigiano Reggiano”?

22 To understand the proper context in which these issues arise, we begin by setting out the broad features of the qualification of rights mechanism under the GIA.

23 Under the GIA, as originally enacted, s 46 provides for a system where the rights in respect of a registered or soon to be registered GI can be qualified. Such requests for a qualification to be entered in the register can be made either before the GI is registered, or after the GI has been registered. Section 46 provides:

24 The consequence of a qualification of rights is that the protection afforded to the registered GI is restricted accordingly: s 46(7) of the GIA. Therefore, separate from opposition proceedings under s 45 of the GIA, the qualification of rights process provides an avenue for interested persons (who are typically third parties) to request that there should be a qualification or carve-out to the rights conferred under the GIA in relation to certain elements of the GI for which registration is sought. Per ss 46(1)(a) or 46(1)(b) of the GIA, these elements would comprise either “any name contained in the geographical indication”, or “any term which may be a possible translation of the geographical indication”.

25 The purpose of the qualification of rights process thus emanates from its statutory text. Providing an avenue for interested persons to request a qualification of the rights conferred under the GIA in respect of a registered GI allows both registrants of the GI and interested third party traders to achieve clarity on whether specific terms will or will not be available for use by third parties: see IPOS Public Consultation at para 4.15. The same point was subsequently repeated by Mr Edwin Tong in the Second Reading of the Geographical Indications (Amendment) Bill (Bill No 4/2020) (“2020 Amendment Bill”) (Singapore Parl Debates; Vol 94, Sitting No 116; [3 February 2020] (Edwin Tong Chun Fai, Senior Minister of State for Law)):

It is clear from a contextual reading of the Second Reading speech as a whole that the above-cited passage was intended to reflect the pre-existing concept of a request for a qualification of rights under s 46 of the GIA.

26 As to the grounds for a request for a qualification of rights, s 46(2) of the GIA sets out the relevant grounds on which a request under s 46(1) may be made.

27 Notably, the translation ground under s 46(2)(b) of the GIA is only available to a request for a qualification of rights in relation to “any term which may be a possible translation of [the GI]” (per s 46(1)(b)).

28 Meanwhile, the grounds under s 46(2)(a) read with Part III of the GIA appear to be available to both a request for a qualification of rights in relation to “any name contained in the [GI]” (per s 46(1)(a)) and a request for a qualification of rights in relation to “any term which may be a possible translation of the [GI]” (per s 46(1)(b)). One such ground is that of genericness, provided for under ss 11(c) and 15(a) of the GIA. In the IPOS Public Consultation, it was clearly envisaged (at para 4.15) that:

This reference to a “disclaimer request process” was eventually enacted as s 46 of the GIA.

29 We observe that, in GI jurisprudence, “genericness”, a somewhat awkward term, is typically used to refer to the situation where a name, despite denoting the place from where a product originates, has become the term customary for such a product. The relevant public therefore understands the term in question to be commonly used for a type of product, such as “Camembert” for cheese. GIs can become generic as a result of the owner’s failure to prevent others from using the term for other goods or services not necessarily originating from the region suggested by the GI. The transformation of a GI into a generic term may occur in different countries and at different times: see, Bernard O’Connor, The Law of Geographical Indications (Cameron May, 2004) at p 95. Nonetheless, as Fonterra’s sole pleaded ground in the present case is not the genericness ground, we do not propose to say more on the genericness inquiry.

30 Rounding off our preliminary discussion, we observe that the present case relates to the GIA in force as of 16 September 2019 (ie, on the date that Fonterra filed its Request). After the Request had been filed, the GIA was amended on 15 August 2020 by the Geographical Indications (Amendment) Act 2020 (No 5 of 2020) (thereafter, the “Post-2020 GIA”). The legislative changes effected by the amendments have no bearing on the present appeal. However, it is worth noting that the post-registration qualification of rights regime has since been replaced by a new limitation of scope (“LoS”) regime. The reason for this was explained by Mr Edwin Tong, who was then Senior Minister of State for Law, in the Second Reading of the 2020 Amendment Bill (Singapore Parl Debates; Vol 94, Sitting No 116; [3 February 2020] (Edwin Tong Chun Fai, Senior Minister of State for Law)):

Nonetheless, the pre-registration qualification of rights regime remains governed by s 46 of the Post-2020 GIA.

31 It is apparent, if one compares s 46 of the GIA with s 48A of the Post-2020 GIA, that the material change wrought by the legislative amendments was as to which body would hear an application to qualify the rights of a GI, after that GI had been registered. There is, however, no change as to the basis on which one can seek to qualify or, to use the terminology under the GIA 2020, limit the rights accorded in respect of a registered GI. An application to qualify or limit the rights in respect of a registered GI can still be made on the basis that a term which may be a possible translation of a registered GI is not actually a translation of that registered GI: see s 46(2)(b) of the GIA and s 48A(2)(b) of the Post-2020 GIA.

32 We deal first with the question of which party bears the burden of proof where a proposed qualification of rights is opposed. Regardless of whether the proposed qualification of rights is opposed or unopposed, we agree that, as submitted by Prof Llewelyn, there are two thresholds that a party who makes a request (a “Requestor”) must cross before the requested qualification of rights under s 46 of the GIA can be allowed:

(a) first, the Registrar must be satisfied pursuant to r 40(4) of the GIR that the request discloses a prima facie case, and if he is so satisfied only then is the Registrar obliged to publish the proposed qualification of rights in the Geographical Indications Journal for opposition purposes; and

(b) second, after the proposed qualification of rights has been published for opposition purposes, the Requestor must satisfy the Registrar on a balance of probabilities that it has made out either or both of the grounds in s 46(2) of the GIA, and only then will the Registrar be obliged to enter the qualification of rights in the register. In this regard, the fact that there has been an opposition to the proposed qualification of rights does not shift the burden of proof to the registrant who has opposed the qualification.

33 We come to this conclusion for three main reasons. This flows from the prescribed procedure set out by r 40 of the GIR, as well as the nature and function of the qualification of rights mechanism.

34 The prescribed procedure for a request for a proposed qualification of rights is found in r 40 of the GIR, which provides as follows:

35 Firstly, r 40(4) of the GIR is not necessarily inconsistent with the two thresholds that a Requestor must cross before the proposed qualification of rights under s 46 of the GIA is allowed. Fonterra takes the view that the Registrar, in proposing to allow the qualification request under r 40(4), must have been satisfied by that time that at least one of the grounds for qualification is made out. However, we agree with Prof Llewelyn that what is required under r 40(4) of the GIR is that the Registrar must be satisfied that the request discloses a prima facie case before the Registrar becomes obliged to publish for opposition the proposed qualification of the registered GI. In this regard, the Requestor must provide such evidence as may be requested by the Registrar pursuant to r 40(2) of the GIR. Understandably, it would not make sense to waste resources entertaining a request which does not even disclose a prima facie case.

36 Secondly, the phrase “and the Registrar is satisfied that either or both of the grounds in section 46(2) of the Act is or are made out” (the “Second Phrase”) [emphasis added] in r 40(5) of the GIR suggests that at the time the Registrar proposes to allow the request under r 40(4), he need not yet be satisfied on a balance of probabilities that at least one of the grounds in s 46(2) of the GIA is or are made out. Therefore, the absence of an opposition does not necessarily mean that the Registrar is obliged, without more, to enter the proposed qualification of rights in the register. It may be that, in a straightforward case, the Registrar is already satisfied with the grounds for the proposed qualification of rights at the time of publishing it for opposition purposes so that he goes on to enter the qualification of rights in the register once the opposition period has passed and no notice of opposition has been filed. But it is not inconceivable that, in an exceptional case, the Registrar, after having published the proposed qualification of rights in the Geographical Indications Journal, of his own motion comes across evidence that casts doubt on the grounds relied upon by the Requestor. Suppose, taking the present GI as an example, that even if there was no opposition to the Request filed within the prescribed period, the Registrar subsequently came across a dictionary extract not relied upon by Fonterra (as Requestor), or came across certain Internet websites claiming that “Parmesan” is a translation of “Parmigiano Reggiano” cheese. Nothing restricts the Registrar to considering only evidence that has been put forward by the Requestor (or by the opponent, if there is one), or from exercising his powers under r 40(2) of the GIR to call for further evidence. It follows that the Requestor must go on to satisfy the Registrar on a balance of probabilities that at least one of the grounds for the proposed qualification of rights is made out.

37 Fonterra submits that r 40(5) of the GIR only applies to the situation where no notice of opposition to the proposed qualification of rights has been filed. In our view, however, there is no logical reason why the position ought to be any different where an opposition has been filed. On the contrary, as we explain below, the function of a qualification of rights is that it effectively serves as a carve-out of the rights accorded in relation to a registered GI. Where the qualification of rights is challenged by the very party whose interests will be adversely affected by the carve-out, it is plain that the burden of proof must fall on the party seeking the carve-out in the first place, ie, the Requestor.

38 The position under r 40 of the GIR (relating to qualification of rights proceedings) may be usefully contrasted to that under r 65 of the GIR (being the prescribed procedure in relation to cancellation proceedings). In this regard, an application for the cancellation of the registration of a registered GI may be made by the registrant, or by “any other person” on certain prescribed grounds (s 52(1) of the GIA). The application must be accompanied by a statement of the grounds relied upon by the person applying for the cancellation (r 65(2) of the GIR), and the Registrar is entitled to ask for evidence in respect of the application (r 65(3) of the GIR). Where the Registrar proposes to allow the cancellation application, the Registrar must publish the proposed cancellation in the Geographical Indications Journal (r 65(5) of the GIR). Crucially, rr 65(6)–65(7) goes on to provide:

39 Following from r 65(7) of the GIR, it is clear that the process for cancellation proceedings is structured such that at the time the Registrar publishes the proposed cancellation in the Geographical Indications Journal for opposition purposes, the Registrar is already satisfied on a balance of probabilities that one or more of the grounds for cancellation is or are made out. In contrast, the inclusion of the Second Phrase in r 40(5) of the GIR is deliberate and fortifies our conclusion above that the Registrar is not obliged, in the absence of any opposition put forth by the registrant of the GI in question, to allow the application for the proposed qualification of rights.

40 Our third reason follows from the nature and function of the qualification of rights process.

41 In this regard, Fonterra highlights that the present action involves opposition proceedings and, citing the example of the opposition proceedings under the ground of s 41(1)(f) of the GIA in Prosecco, submits that it is well established that it is the opponent in opposition proceedings that bears the burden of proof. In this connection, Fonterra points out that the opponent/registrant in qualification of rights proceedings is the party that files evidence in the opposition proceedings first, and is also given the right to file evidence in reply to the Requestor’s evidence: see rr 44(1)(a) and 44(1)(c) of the GIR. Fonterra submits that this is entirely consistent with the opponent/registrant bearing the burden of proof in opposition proceedings.

42 However, the nature of qualification of rights proceedings cannot be analogised to the opposition proceedings in Consorzio di Tutela della Denominazione di Origine Controllata Prosecco v Australian Grape and Wine Incorporated [2023] 2 SLR 509 (“Prosecco”), or indeed to opposition proceedings under s 45 of the GIA. Opposition proceedings under s 45 means an opposition to an application for the registration of a GI. In that context, it is clear that there has been a determination by the Registrar at the time the Registrar accepts the application and causes it to be published in the Geographical Indications Journal. When an applicant applies to register a GI, the Registrar first examines whether the application satisfies the requirements under the GIA: s 43(1) of the GIA. This requires the Registrar to determine, among other matters, whether the indication is one that ought to be refused registration by reason of one or more of the grounds for refusal in s 41 of the GIA. Such grounds include, for instance, that the indication does not fall within the meaning of a GI (s 41(1)(a)), or that the GI name has become generic in Singapore (s 41(1)(e)). Only if it appears to the Registrar that the requirements for registration are met, can the Registrar then accept the application and cause the application to be published in the Geographical Indications Journal for opposition purposes: ss 43(6) and 45(1) of the GIA. Crucially, per s 48(1) of the GIA, where an application for registration of a GI has been accepted and either no notice of opposition is given within the prescribed period or all opposition proceedings are withdrawn or decided in favour of the applicant, the Registrar shall register the GI.

43 In contrast, following from our analysis of rr 40(4)–40(5) of the GIR above, the point to be made is that there is no conclusive determination yet when the proposed qualification of rights is published in the Geographical Indications Journal for opposition purposes. The reason for publishing the proposed qualification of rights in the Geographical Indications Journal is advertisement; in other words, to evoke a response from the party whose interests would be affected by the proposed qualification of rights (ie, the registrant of the GI), and whose views the Registrar would then have the benefit of considering. We reiterate that it is the Requestor in a proposed qualification of rights who is effectively seeking a carve-out of the rights accorded to a registered (or soon-to-be-registered GI). Accordingly, it is this party who has to prove that the carve-out to the rights under the GIA should be made.

44 Finally, Fonterra’s reliance on r 44(4) of the GIR is misplaced. Rule 44(4) states that if the opponent/registrant does not file evidence in support of its opposition, the opponent/registrant is “treated as having withdrawn the opposition”. In our view, where the notice of opposition has been deemed withdrawn, the situation plainly falls within the clear ambit of r 40(5) of the GIR and, as has been discussed above, the Requestor must satisfy the Registrar on a balance of probabilities that at least one of the grounds in s 46(2) of the GIA is made out before the Registrar is obliged to enter the qualification of rights in the register.

45 In conclusion, the party which bears the burden of proof where a proposed qualification of rights is opposed remains the Requestor. In the present case, this means that Fonterra must demonstrate that the ground under s 46(1)(b) read with s 46(2)(b) of the GIA is made out, namely, that “Parmesan” is not a translation of “Parmigiano Reggiano”.

46 Before the Judge and the PAR, there was considerable argument as to whether a faithful approach or a strict literal approach to translation ought to be adopted. We agree that in most cases, a literal translation will be a faithful translation of the term in question. Yet, where there is a divergence between the two, a “translation” under the GIA must refer to a faithful translation which captures the essence of the word or phrase, and not merely to a literal one. To use the examples posed by the PAR, she referred to the terms “terima kasih” and “sama sama” in Bahasa Indonesia. The literal translations of these terms are “receive give” and “same same” respectively, while a faithful translation results in “thank you” and “you are welcome” respectively. We also agree that what is required is a translation of the registered GI as a whole.

47 In our judgment, however, the inquiry is not a binary choice between a faithful and literal translation. What is meant by a “faithful” translation, and does it imply understanding (and whose understanding is relevant)? In this regard, we first set out the parties’ respective positions, before turning to our decision. Prof Llewelyn submits that a “translation” under the GIA means a contextual and locally understood translation; accordingly, consumer perception is relevant. The inquiry is whether the term would be viewed in Singapore by the average consumer of the relevant products to be an unambiguous and non-confusing translation of the registered GI. At the hearing, Prof Llewelyn confirmed that while he is not suggesting that a translation must always be in one of the four official languages of Singapore (ie, Malay, Chinese, Tamil and English), it is most probable as a matter of fact and practice that a translated term would be in one of the four official languages. Furthermore, he submits that placing reliance on dictionary evidence alone is undesirable and unrealistic. Fonterra agrees with Prof Llewelyn’s suggested approach, and adds that the protection of translations of registered GIs under s 4(6) of the GIA constitutes an extravagant extension to the rights of the GI. Accordingly, a translation should, at the very least, be known to the average Singapore consumer to convey the same meaning as the GI in question.

48 The Consorzio submits that a faithful translation should be preferred over a literal one, dictionaries may be considered to shed light on the meaning of words in the English language and that certified translations are not necessary to determine the meaning of the words. The Consorzio further submits that consumer perception is generally irrelevant to the translation inquiry, the only exception being where such perception is captured as part of the history and etymology of the term. Marketplace evidence should therefore be excluded from the translation inquiry.

49 In our view, a “translation” under the GIA cannot be divorced from the fact that it must be a translation known to the average consumer in Singapore. Therefore, how the translated term is known and used locally can have an impact on the inquiry. The function of a GI is, after all, to indicate to consumers that the product originates from a specified region: see s 2(1) of the GIA. Therefore, if a translation is not known or used by the average consumer in Singapore, it is difficult to see how the translated term can be said to indicate to consumers that the product which it describes originates from a specified geographical region. In this sense, the fact that a translation must be known to the average consumer in Singapore also accords with the overarching purpose of the GIA, namely, that of consumer protection. As stated by the then Senior Minister of State for Education and Law, Ms Indranee Rajah, during the Second Reading of the Geographical Indications Bill 2014, GI protection “safeguards the interests of our consumers” and “provides greater assurance that the products bought by our consumers truly carry the characteristics that they are known for, and which are attributable to their geographical origin” [emphasis added](Singapore Parl Debates; Vol 91, Sitting No 15; [14 April 2014] (Indranee Rajah, Senior Minister of State for Education and Law)).

50 We agree that the purpose of protecting translations of registered GIs under the GIA is the same as that of protecting GIs in their native language – to address the unfair trade practice of using in Singapore the GI on a product which may lead the public into the false belief that that product originated from the particular geographic location. This affords a more complete protection and ensures that merchants are unable to circumvent the protection of GIs by simply changing the form and language in which they are presented. It also makes GIs more accessible to people who speak a different language (Singapore Parl Debates; Vol 94, Sitting No 116; [3 February 2020] (Christopher de Souza)). That said, we recognise that speaking a particular language is not necessarily a pre-requisite to usage of a term in that language. To use some common examples from outside the GI context, although not all Singaporeans may be proficient in the Malay language, we posit that the familiarity of certain terms such as “ikan bilis” (dried anchovies) and “teh tarik” (tea prepared in a particular way) in the local vernacular is such that these terms would likely be well known to the average Singaporean consumer.

51 It may be recalled that under the scheme and design of the GIA, translations of a registered GI do not need to be registered at the outset and are protected by default. Where translations are concerned, the determination of whether a particular term is a translation of a registered GI falls to be undertaken on a case-by-case basis, at the time of infringement proceedings under s 4 of the GIA or upon a request for a qualification of rights under s 46 of the GIA: see IPOS Public Consultation at paras 4.4 and 4.15. The Consorzio’s position, if adopted, will lead to the result that any and all translations of a registered GI will be protected under the GIA by default, even if – to give an extreme example – such translation is in a language that is not known to the average consumer in Singapore in the sense that it is hardly heard or used here (eg, Tibetan or Xhosa). In our view, it is difficult to see how a term in a language that is not known to the average consumer in Singapore can be said to convey any particular meaning to the average consumer, such as to be a “translation” protected under the GIA.

52 We clarify that we are here concerned about the average consumer in Singapore, which means Singapore citizens and residents and not those who are merely passing through. In Prosecco, the same point was made in the context of s 41(1)(f) of the GIA: Prosecco at [49]. In the context of the present dispute, we add that the average consumer is not someone with a specialist knowledge of cheese. Nor can it be said that the average consumer of Parmesan or Parmigiano Reggiano cheese must be a member of the expatriate community or just the Italian community in Singapore.

53 The assessment of whether a translated term is known to the average consumer in Singapore must be supported by credible evidence. In our judgment, such evidence may come from a variety of sources and need not be restricted to the meaning of words as reflected in the entries of reputable dictionaries, so long as the sources relied upon can be shown to be relevant and, more importantly, credible to the inquiry in question. For instance, other than dictionaries, evidence of the general usage of the term in Singapore may come from credible tomes (eg, on cheese). It may also come from consumer surveys, although we reiterate the caution that the probative value of such surveys often varies depending on how the survey in question has been conducted. In Prosecco, the Court of Appeal observed (at [69]) that “[c]onsumer surveys can, as one knows, be skewed to reach a certain desired result. Parties should, therefore, when adducing evidence of such consumer surveys, also place before the court evidence of how such surveys were conducted.” In the present case, counsel confirmed that neither party had tendered consumer surveys or other relevant market surveys, hence we do not propose to say more on the subject.

54 To be fair to the PAR and the Judge, dictionaries in the English language remain a helpful starting point for ascertaining the ordinary meanings of words. However, it must be borne in mind that words do not exist in a vacuum, and how a particular word is used and the meaning or meanings which it bears over time may vary depending on the particular context and local conditions in which the word is used. Therefore, dictionaries may not accurately reflect the usage of certain words among the local population in Singapore. We agree with the point made by Prof Llewelyn that dictionary meanings should not be taken to be determinative, because the meaning of words does change over time and, more importantly, the meaning given to words in English dictionaries is not necessarily the meaning which the Singapore populace would ascribe to those words. For instance, a meaning of a word in a particular language or vernacular in Singapore which should be taken into account may not be captured in a foreign published dictionary of that language, because such dictionary entries have been curated by individuals from another country who may not be informed of usage in Singapore.

55 This recognition that foreign published dictionaries may not accurately reflect the usage of a word in Singapore is not without judicial precedent. In Courts (Singapore) Pte Ltd v Big Box Corp Pte Ltd [2018] 5 SLR 312 (“Big Box”), in the context of an application for a declaration of invalidity of a registered trade mark under s 23(1) of the Trade Marks Act (Cap 332, 2005 Rev Ed) (“TMA”), the High Court (per George Wei J) considered (at [70]) that the dictionary definitions of “big box” cited by the applicant in that case did not assist or advance the applicant’s case. Wei J considered that the Oxford online dictionary definition cited by the applicant made clear that the definition provided was “North American informal”, and there was “nothing to suggest that the definition had entered the vernacular in Singapore by the [date of application of the registration of the trade mark (“Application Date”)]”: Big Box at [70]. Furthermore, the Merriam-Webster online dictionary was an American source and, in any case, the date of the copy in evidence was some ten years after the Application Date: Big Box at [70]. More importantly, there was no evidence at all before the High Court to support the view that by 2005 (ie, the Application Date), the average Singapore consumer would have been familiar with the meaning of “big box” advanced by the applicant: Big Box at [71]–[72].

56 At this juncture, we clarify that, given the conceptual as well as historical dissimilarities between GIs and trade marks, the above case is not relied upon to import certain principles of trade mark law into the law of GIs. Rather, it merely illustrates the commonsensical view that the dictionary should not be taken to be the definitive authority on the meaning of a particular word in Singapore when it comes to consumer usage.

57 For the reasons above, we hold that a “translation” for the purposes of s 46(1)(b) read with s 46(2)(b) of the GIA must be a translation known to the average consumer in Singapore. For completeness, we clarify that the relevant time at which this is to be assessed is plainly at the time the request for the qualification of rights is filed. In the present case, Fonterra must thus show, that at least by 16 September 2019, “Parmesan” was not a translation known to the average consumer in Singapore of the term “Parmigiano Reggiano”.

58 The parties do not dispute that what is required is a translation of the term “Parmigiano Reggiano” as a whole. Applying the above approach, we are satisfied that Fonterra has discharged its legal burden of proving that “Parmesan” is not a translation of “Parmigiano Reggiano” for the purposes of the GIA.

59 As a starting point, we deal with the dictionary evidence tendered by the Consorzio. The three dictionary entries which were considered and accepted by the Judge as reputable dictionaries in demonstrating that “Parmesan” is a translation of the Italian term “Parmigiano Reggiano” into the English and French languages are as follows:

(a) the Collins Dictionary extract;

(b) the Larousse Italian-French Dictionary extract and its accompanying translation into English by a qualified translator; and

(c) the Cambridge Italian-English Dictionary extract.

60 As a preliminary matter, it must be ascertained the language in which “Parmesan” is a word, before the question can be asked whether this term (“Parmesan”) is a translation of the term “Parmigiano Reggiano” in a different language. Before this Court, the Consorzio’s position has been that “Parmesan” is a word bearing the same meaning as “Parmigiano Reggiano” in the English, French and even Italian languages.

61 The Consorzio tendered a certified English translation of the Larousse Italian-French Dictionary extract as follows:

However, the Larousse Italian-French Dictionary extract defines the Italian term “Parmigiano” as “Parmigiano (Reggiano)” or “Parmesan m” in the French language. For it to be taken that the average consumer in Singapore would understand “Parmesan m” (in French) to be a translation of the term “Parmigiano Reggiano” (in Italian), it would need to be assumed that the average Singaporean consumer is proficient in French or at the very least familiar with the meaning of the term “Parmesan m” in the French language. There is no evidence before the Court suggesting that it would be appropriate to make this assumption, and we do not find it appropriate to do so.

62 The Collins Dictionary extract states that the term “Parmigiano” in British English means “another name for Parmigiano Reggiano”. In turn, “Parmigiano Reggiano” in British English means “another name for Parmesan cheese”. Lastly, “Parmesan cheese” is then defined as “a hard dry cheese made from skimmed milk, used grated, esp [sic] on pasta dishes and soups”.

63 The Cambridge Italian-English Dictionary extract defines the Italian term “parmigiano (reggiano)” as “Parmesan cheese” in the English language.

64 Taken broadly, the Collins Dictionary extract and the Cambridge Italian-English Dictionary extract may be said to provide some support for the contention that there is equivalence in meaning between the terms “Parmigiano Reggiano” and “Parmesan Cheese” in the (British) English language. However, although Singapore adopts British English as its official working language, these dictionary extracts are meanings compiled by foreign publishers who may not have been informed of the way in which the word has been used in Singapore. Put differently, these dictionaries are not written in (and were not designed to reflect) the vernacular used by locals in Singapore. We therefore reiterate that dictionary definitions should not be definitively relied upon to reflect the usage of certain words among the local population in Singapore, especially where there is cogent evidence to the contrary.

65 In the present case, we are satisfied that sufficient evidence has been adduced by Fonterra to support an inference that consumers in Singapore regard “Parmesan” and “Parmigiano Reggiano” as referring to two different kinds of cheese products distinguishable by origin in that Parmigiano Reggiano originates from a specific region in Italy whereas Parmesan can come from several countries or regions. Fonterra’s evidence is that it “has been for many years” “customary” in Singapore to refer to “a hard, dry, easy to grate, or grated cheese with a sharp, slightly sweet, salty flavour” as “Parmesan” cheese, and that “Singaporean consumers are not accustomed to making a direct link between the hard, dry cheese called ‘[P]armesan’ and the protected GI ‘Parmigiano Reggiano’”. Although Fonterra did not tender any consumer surveys or other market surveys, broadly speaking, it relies in its own words on the following evidence:

(a) First, evidence of at least ten product listings which show that “Parmesan” cheese products are marketed and sold to Singapore consumers in-stores and online with clear indications that these products originate from countries outside Italy. These include “Parmesan” cheese products which have been produced in New Zealand, Australia, Germany, Korea and Japan.

(b) Second, evidence that the online catalogues of groceries and cheese stores in Singapore (such as NTUC FairPrice’s online store) as well as Amazon Singapore categorise “Parmigiano Reggiano” cheese separately from “Parmesan” cheese.

66 Generally speaking, we agree with Mr Chan’s submission that marketing practices can, to some extent, influence consumer awareness of a product on the market. This approach is not novel – in Prosecco, we observed (at [68]) that “[a]dvertising materials may be useful in so far as they provide some evidence as to how the product for which the GI is being registered has been marketed to the consumer in Singapore.” However, the extent to which advertising materials can be said to support an inference of consumer perception must invariably depend on the context at play, including the nature of the product as well as what information exactly it is being assumed that advertising materials convey to Singaporean consumers. Thus, on the facts of the Prosecco case, we concluded (at [73]) that no evidence was before the court such as to render it appropriate to assume that the Singapore consumer would peruse the advertising material and notice that the grape variety “Prosecco” had been used to make the wine of the same name.

67 In the present case, the advertisements and marketing material tendered by Fonterra show how “Parmesan” cheese has been marketed and sold to consumers in stores and on online platforms in Singapore. In our judgment, the differences in the overall visual cues between, on the one hand, the products labelled and sold as “Parmesan” and, on the other hand, those labelled and sold as “Parmigiano Reggiano” are considerable. The former are often labelled as “Parmesan” cheese, with no mention of any association with “Parmigiano Reggiano” cheese. Fonterra tendered at least ten product listings which illustrate this point. To take one example, the product packaging of “Shredded Parmesan Cheese” sold under the “Emborg” brand markets its product as “Traditional and tasty Parmesan cheese – ideal on homemade pastas or in a classic Caesar salad”. Its product description (for example, on the RedMart online store) clearly identifies the product as having been produced in Germany:

68 Likewise, a consumer who peruses the labels of Parmesan cheese sold in wedge or grated form under the “Cowhead” brand would see:

It is at once obvious to the average consumer that what is being offered for sale is “Grated Parmesan” and, in slightly smaller font, “Bold Flavoured Finely Grated Cheese” for the product on the left, and “Parmesan Wedge” for the product on the right. Both products are stated to originate from Australia.

69 On the product listing and packaging for “Fairprice Cheese - Parmesan (Shredded)”, stated to originate from Singapore, there is likewise no association with the words “Parmigiano Reggiano”:

70 That said, it is common knowledge that Parmesan is not only sold in the form of wedges, pre-packaged, sliced or grated, but may also be marketed as an ingredient in sauces or as toppings for other dishes. Counsel for the Consorzio, Mr Sivagnanaratnam Sivananthan (“Mr Sivananthan”), highlighted one instance of “Sacla Whole Cherry Tomato with Parmesan Pasta Sauce 350g” sold on the website , which accompanying product description stated that the Parmesan sauce was “made using Parmigiano Reggiano made only in Northen Italy”:

This may be one instance where the terms “Parmigiano Reggiano” and “Parmesan” have been used in close association or even interchangeably in the advertising material. However, in our view, the weight of the marketing material directed at Singaporean consumers clearly establishes that more products have been sold and labelled as “Parmesan” without any association with “Parmigiano Reggiano”. In this regard, it bears reiterating that the Consorzio itself has not tendered any evidence of consumer perception of the terms “Parmesan” and “Parmigiano Reggiano” in Singapore.

71 Other than the one instance referred to by Mr Sivananthan, “Parmigiano Reggiano” cheese has been labelled and sold as “Parmigiano Reggiano”, with no mention of “Parmesan”:

72 Furthermore, the country of origin or production of “Parmesan” cheese products is often conspicuously displayed in the marketing and advertising material. In addition to the examples cited above, we take for instance “Grated Parmesan Cheese 100g” sold under the “Alba Cheese” brand, which is marketed as “Proudly Australian”:

73 Another example is the product listing for “Meiji Hokkaido Tokachi Parmesan Cheese 80g Japan”, which states, in conspicuous font, “[Made in Japan]”:

74 At this juncture, we observe that, in Singapore, regardless of whether a food product is sold online or in stores, it is common for the country of origin and/or production of the food product to be prominently displayed on product packaging and accompanying product listings. In fact, under Regulation 5 of the Food Regulations (2005 Rev Ed) (“Food Regulations”) promulgated under s 56(1) of the Sale of Food Act (Cap 283, 2002 Rev Ed) it is mandated for information relating to the country of origin of a food product offered for sale in Singapore to be displayed “conspicuously and in a prominent position on the label and shall be clearly legible” (see Regulation 5(3) and 5(4)(e) of the Food Regulations; see also Singapore Food Agency, “A Guide to Food Labelling and Advertisements” <https://www.sfa.gov.sg/docs/default-source/tools-and-resources/resources-for-businesses/aguidetofoodlabellingandadvertisements.pdf> (updated April 2019)):

75 The more pertinent question is this: What is the effect of such information on consumer perception of “Parmesan” cheese in Singapore? In this regard, it bears reiterating that the average consumer is a literate consumer. Quite plainly, he or she would take some notice of the manner in which a food product has been marketed or advertised. He or she may not, depending on the individual priorities and concerns of the consumer in question, ordinarily take notice of the product’s precise ingredient list or its production methods. We also do not go so far as to say that the average consumer would always take notice of where a food product originates from. However, information relating to the country of origin is typically a selling point because it presumably speaks to the quality or authenticity of the product in question, and it is not inconceivable that Singaporean consumers would tend to become familiar with such information through purchasing “Parmesan” cheese over time. Consider, for example, that if a consumer specifically wants to purchase cheese made in Italy, he or she would ordinarily pay attention to such information that is displayed on the product packaging or listed in the product description. Equally plausible is the converse situation, where a consumer wishes to avoid food products made in a particular country, he or she would also pay attention to such information.

76 Therefore, we conclude that the manner in which various “Parmesan” cheese products have been marketed and sold to consumers in Singapore would have influenced the average Singapore consumer to consider that, unlike “Parmigiano Reggiano” cheese which must originate from the Specified Region in Italy, “Parmesan” is not “Parmigiano Reggiano” and can and does, in fact, originate from locations outside Italy.

77 As mentioned above, other than the product packaging and product listings for “Parmesan” cheese products, Fonterra also tendered evidence that the online catalogues of cheese stores and groceries in Singapore as well as Amazon Singapore categorise “Parmigiano Reggiano” cheese separately from “Parmesan” cheese. Fonterra thus submits that traders targeting consumers in Singapore do not regard “Parmigiano Reggiano” and “Parmesan” cheese to be the same products and have categorised these products separately, and this influences consumer perception.

78 For example, a search on the Amazon Singapore online store for “Parmesan” and “Parmigiano Reggiano” leads to completely different results and products:

79 Significantly, it does not appear that “Parmesan” products are brought up in the search for “Parmigiano Reggiano” products on Amazon.sg, and vice versa. We agree that consumers who utilise these online grocers and stores are thus likely to perceive “Parmesan” and “Parmigiano Reggiano” cheese as referring to two different kinds of cheese products.

80 Lastly, we acknowledge at this juncture that Fonterra’s own “Traditional Style Parmesan” sold under the “Perfect Italiano” brand appears to deliberately capitalise on certain Italian associations, such as the colours of the Italian flag. The Consorzio submits that Fonterra is essentially seeking to rely on its own use of ‘Parmesan’ in the marketplace to justify the entry of the requested qualification of rights:

81 Mr Sivananthan also submitted that there have been no statistics on the volume of sales showing how much “Perfect Italiano” cheese has been sold in Singapore. We agree that such information would certainly have been helpful; however, its omission is not necessarily fatal in this case. We have set out above the range of “Parmesan” cheese products marketed and sold to consumers in Singapore, which in our judgment is sufficient for the inference to be drawn that Singapore consumers do not regard “Parmesan” cheese as cheese which must originate exclusively from the Specified Region in Italy.

82 In conclusion, we are satisfied that “Parmesan” is not a translation of “Parmigiano Reggiano” for the purposes of s 46(1)(b) read with s 46(2)(b) of the GIA.

83 For the above reasons, we allow the appeal. The result is that the Request succeeds, and the Opposition is dismissed, and the following qualification of rights should be entered into the Register:

84 Fonterra, having succeeded on its appeal, is entitled to costs here and below, as well as before the PAR. On the quantum of such costs (including disbursements), Fonterra submits that a fair quantum would be the sum of S$83,543.11. The Consorzio proposes a sum of S$113,530 on the basis it succeeds in the appeal. We note that Fonterra’s proposed sum does not appear to have included the costs of the proceedings before the PAR, which parties agreed as being the sum of S$10,530.00 (all-in). Given the complexity of the issues raised in this appeal and below, we fix the costs of this appeal together with AD/OA 20/2023 (being Fonterra’s application for permission to appeal) and the hearings before the Judge and the PAR at S$100,000 (inclusive of disbursements) to be paid by the Consorzio to Fonterra.

85 Finally, we record our deep appreciation to counsel and Prof Llewelyn for their helpful and cogent submissions, which greatly assisted us in the appeal.