1. Preliminary remarks
2. Likelihood of confusion between trademarks in EC case law
3. The European Court of Justice in Medion AG v. Thomson
4. The experience in Italy regarding likelihood of confusion in the field of composite trademarks.
1. Preliminary remarks
The evaluation of the likelihood of confusion between trademarks marks is a necessary step for assessing the likelihood of confusion in terms of articles 5.1 of EC Directive 89/104, letter b) and 8 of CTM Reg. 40/94.
The decisions of the Court of First Instance and of the Court of Justice on this matter have a long-consolidated structure, which relies on a bunch of statements which are promptly reproposed in all the discussed cases.
The model resulting from the EU rulings is built on the following steps :
a) it starts with the statement that the protection of the mark is tied to the risk of confusion ;
b) it continues with the statement that the risk of confusion requires an overall evaluation which takes into consideration all the pertinent factors of the case in hand ;
c) it adds that this evaluation implies a certain interdependence between the factors which enter into consideration and, in particular, between the similarity of trademarks and that of designated products or services ;
d) lastly, it explains that the perception of the trademarks made by the average consumer of the product in question plays a determinant role in this overall evaluation of the risk of confusion .
Once this premise is made, it is possible to enter the evaluation of the likelihood of confusion between products, which is an abstract one referring to the registration class, and lastly to consider the appearance of the marks in question .
It is precisely this second and last tier of the judgement of likelihood of confusion which the considerations below are aimed at and, in particular, at the difficult subject of the appearance of composite marks, whether they result from one or more name-related or figurative elements.
The European Court of Justice recently ruled on this difficult subject in the Medion v Thomson case delivering a highly innovative decision.
The inquiry into the community jurisprudence also provides an occasion for a more generalised reflection on the judgement of likelihood of confusion between composite marks viewing the significant experience in Italian case law.
2. Likelihood of confusion between trademarks marks in EC case law
Until the recent decision in Medion v Thomson case, the principle governing the definition of likelihood of confusion that was constantly confirmed by the lower courts in their rulings and upheld by the Court of Justice recited as follows: “the evaluation of similarity between two trademarks must be based on the overall impression produced by the same, in particular in consideration of their distinctive and dominant elements.”
This statement does recur in almost all the rulings involving a likelihood of confusion issue, precisely for the judgement of likelihood of confusion between marks .
Nevertheless, following an overall consideration of the rulings, it appears clear how this statement of principle has an intrinsic inconsistency and basic ambiguity which the EU judges did not overlook for long .
The ambiguity of the formula resulted from the tension that exists between the “overall impression” produced by conflicting trademarks compared to “in particular in consideration of their distinctive and dominant elements”, as parameters which can equally direct the judgement of likelihood of confusion between trademarks .
First of all it is necessary to underline that while the first reference to the overall impression leads to an evaluation made in the memory of the hypothetical consumer in the presence of only one of the two marks and based on factors which, at least regarding the perception of the earlier trademark, can also be effected by its use, the second reference to the single distinctive and dominant elements seems to require a simultaneous analysis of the two marks and certainly refers to the abstract value and role of each single element the mark is composed of. This is the first aspect of contradiction within the European formula.
But then, and even more importantly, almost everyone acknowledges that placing priority on the overall impression parameter that a trademark produces as the guiding criterion of the judgement of likelihood of confusion leads to an outcome which is at the opposite of the one which results if only the single distinctive elements within a mark are considered.
It might be true that two trademarks which have common components can easily deliver a similar overall impression but it is also true, as experience shows, that two trademarks can produce a very different overall impression from each other despite presenting simultaneous common distinctive components, as the Thomson case proves. This consideration must be kept in mind when the protectable limits of the various marks which a trademark is composed of are to be determined, within the limits of an operation ex ante which hardly looks entirely convincing.
Just think of the paradigmatic case of marks which have a common part, with the addition of diversifying elements which consist of the name of the manufacturer/distributor, of the general trademark or of other words . If in these cases the overall impression conveyed is undoubtedly different from all points of view (visual, phonetic and above all conceptual), it is not easy to deny that the common distinctive component should lead to confirming the likelihood of confusion between marks.
In the end it appeared clear that the parameters formulated by the EU jurisprudence up to the Medion case (i.e. the overall evaluation and the concurrent account for the single elements), could not be combined together in a single criterion, as had been done in the much repeated formula, but pointed towards two different aspects of the judgement of the likelihood of confusion between marks. In other words, the two constructions were not easily referable to unity within a single stereotypical formula.
At times the ambiguity here uncovered could be traced in some rulings and noticeably in the important decision of the Matratzen  case.
Here the Court seemed to have fully acknowledged the problem raised when it wrote (point 33) that: “a composite trademark can be considered similar to another trademark, identical or similar to one of the components of the composite trademark, only if the latter constitutes the dominant element of the trademark… This occurs when this component by itself can dominate the image of this trademark that the public has in its memory…;” nevertheless, the Court continued (point 34), “it is necessary to point out that this approach does not mean taking into consideration only one component of a composite trademark and comparing it with another trademark. Instead it is necessary to make the comparison by examining the trademarks in question considering each one as a whole…this does not exclude that the overall impression produced by a composite trademark can be dominated by one or more of its components.”
The reasoning of the Court seemed to assume that in the case of a composite trademark the identifying capacity would normally result from one (or some) of its elements and that the judgement of likelihood of confusion thus need to be conducted accounting only for these dominant component. I think that this is an inconclusive approach to the extent that sometimes the components in composite trademarks can be placed, in terms of their characterising capacity, all on the same level, being thus impossible to make a ranking among them that in turns reflects on their level of protection .
So what is the prevailing aspect in consideration of the various results of each of the reviewed approaches?
Actually, these two parameters do not lend themselves to always being harmonised within a single criterion, but point in two different directions.
EU decisions, even before the Medion ruling, did not remain insensitive to this ambiguity which is witnessed by the many rulings of the CFI which, after having identified the various distinctive components of the composite trademark, had correctly acknowledged their protection even if coupled with further additions in subsequent trademarks, thus finally disregarding the overall impression parameter which probably would have led to excluding interference . Just like in other cases where, ruling on the opposite, the prevailing of an overall impression approach led to neglecting to protect some components despite being acknowledged as autonomously distinctive .
Another anomaly involved in the EU reasoning lies in the fact that the overall impression criterion based on the consideration of the trademark distinctive single elements is promptly cited even when the compared trademarks are simple trademarks, for example purely name-related trademarks by definition composed of a single distinctive element (and not of various elements, which could lend themselves to be ranked based on their distinctive character).
3. The European Court of Justice in Medion AG v Thomson
In the ruling in question  the EU judges directly faced the subject here discussed and, revealing the innate contrast in the rule according to which “the evaluation of similarity between two trademarks must be based on the overall impression produced by the same, in particular in consideration of their distinctive and dominant elements”, they reproposed it in an amended version precisely tailored on a trademark composed of two names, both characterising.
The case involved the trademarks Life and Thomson Life, where the first mark is reproduced in its entirety within the second, with the addition of the manufacturer’s trademark (Thomson). The critical point was precisely that of the judgement of likelihood of confusion between trademarks, since the two products resulted to be identical.
Under this profile, during the case it appeared clear the contrast between the so-called “Pragetheorie”, or overall impression theory widespread in German practice, face an analytical approach sustained by the owner of the earlier trademark and aimed at enhancing the distinctive position of his mark Life within the later Thomson Life trademark.
As correctly demonstrated by the defence of the two parties, the application of the overall impression theory would have led to denying the likelihood of confusion between the two marks, given that the impression carried by the second mark was undoubtedly different from that of the first. Nonetheless, the first mark maintained its characterising autonomy in the composite trademark, even if not constituting its dominant component, and denying it would have led to withdrawing protection of a trademark each time were it to be reproduced as such in a later mark next, for example, to the manufacturer’s general trademark.
The Court showed to care for this latter argument and in making its ruling reduced the role of the overall impression criterion where, as in the case in hand, even if the reproduced trademark were not able alone to cause this impression and thus did not constitute the dominant element of the second mark, it nevertheless had an autonomous distinctive position. Substantially the protectability within the composite mark was acknowledged of those among its components which possessed an own and significant distinctive capacity.
The steps in the Court’s ruling appear to be much the same taken in previous cases when evaluating likelihood of confusion: it is reiterated that the protected function of the trademark is that of indication of origin, that the protection is against the risk of confusion, to then explain that the risk of confusion must be the subject of an overall evaluation which, in terms of similarity of the trademarks, must be based on their overall impression, in particular in consideration of the distinctive and dominant elements of the same trademarks.
This overall impression though may be dominated by one or more components of the composite trademark, which are then the subject of a specific protection but, and this is the core of the explanation provided by the EU judges, the protection of the autonomous component must be recognised if such component “maintains an autonomous distinctive position in the composite mark, even without constituting its dominant element" (see point 30 of the ruling).
In fact even in this case the use of a distinctive element within a later composite mark does not exclude the possibility of producing a risk of confusion leading the public to believe that the products come from at least associated companies (see point 31 of the ruling).
The Court resorts to the example of the well-known trademark to which the owner had added a pre-existing trademark: in order to confirm the infringement of the latter, an analytical outlook must be used, since on the other hand, it is clear how the composite mark would otherwise result dominated by the well-known trademark, excluding thus confusion between the two marks in terms of their overall impression.
A brief reflection must still be made on the conclusion reached by the Court that the analytical emphasis on the distinctive component, even if not dominant, is imposed by the necessity of avoiding a risk of confusion.
In reality, as clearly demonstrated by the example of the well-known mark proposed above, the problem does not seem to be strictly related to the risk of confusion, at least in the sense of a risk of confusion between sources of origin as it is currently intended.
In fact it may be doubtful, even in the case heard by the Court, that the problem of the trademark owner was (only) that of avoiding the confusion of the Life mark with the later Thomson Life one. Undoubtedly there is an interest of the owner of any trademark to defend its mark from appropriations by others which can lead the consumer to create an association between the marks, even in the form of creating any sort of connection between them having the form of a “link” which does not necessarily refer to the linking between the sources of origin.
This broad approach to the risk of confusion has already been upheld by the Court of Justice and should be carefully considered also in the judgement of likelihood of confusion between marks where the presence of one component in a composite mark can lead the consumer to build up a connection relating to one of the various components of the message that the trademark is the carrier of.
4. The experience in Italy regarding likelihood of confusion in the field of composite trademarks
By analysing the significance of the Medion ruling in the context of the experience in Italy, it seems to me that the position of the Court of Justice may sound somehow familiar in that it is consistent with a statement that is frequently found in rulings on the subject and according to which the protection of a composite trademark covers all the elements that compose it as long as they show an individual distinctive character , with the consequence that “appropriation is also the inclusion, in a composite trademark, of the name-related or emblematic element which characterises the previously registered trademark, even if the composite trademark is composed of other elements which differentiate it from the previous trademark”.
According to Italian case law “two trademarks composed of the same mark can be confused even when one of them includes the business name of the owner” ; and “the likelihood of confusion between trademarks…is not excluded by the concrete application of the different name of the producers on the product” .
Thus, it would seem that in Italy there is no doubt that all the single distinctive elements within a composite mark may benefit from an autonomous protection regardless of the different overall impression conveyed by the conflicting trademarks, due to the presence of differentiation elements and to the product presentation . Exactly as the Court of Justice stated in the Medion case.
The point had long been clear in the Italian literature with specific reference to the subject of composite trademarks where it was observed that by qualifying the likelihood of confusion of trademarks as illegal, the judicial system acted in a preventive manner given that:
“…confusion may de facto not occur, for example because one or the other of the two products is identified even by means of another distinctive mark (another trademark or characterising presentation of the product), which fully meets the function of differentiation, but the use of the similar trademark is still illegal, because the law, to avoid as strongly as possible the confusion between products, has qualified likelihood of confusion as illegal, without placing any relevance on concomitant circumstances, which may decrease (without excluding it) such likelihood of confusion; and without doubt, a situation of likelihood of confusion is abstractly created by using two similar trademarks even if connected with other distinctive means (trademark or method of product presentation), because the latter may not fulfil its function of differentiation” .
A brief mention must also be made of an aspect that has been underlined various times by the rulings which have affirmed the protectability of the single elements, that is the reference to the right of the trademark’s owner as a property right .
This means that the exclusive right over the trademark is an absolute right regardless of the concrete method of its infringement and covers all the cases where the trademark is reproduced by a third party on products interfering with those for which it was registered, causing a potential confusion .
All of these considerations have recently led to wondering whether, with the case of composite marks, the principle which governs the evaluation of the likelihood of confusion between trademarks is fully applicable. According to this the signs ought to be compared not in an analytical manner, with a detailed examination of each single component of the mark, but with a unitary and synthetic approach taking its salient elements into consideration .
The fact that the above principle requires a cautious application did not go unnoticed, based on the significance of the analytical examination , as effectively highlighted by the Court of Justice in the Medion case.
The decision of the Court of Justice also bypasses the main criticism that could be put forth against the approach of the autonomous protection of the single elements of a composite trademark, which is that of tout cour protectability of all the distinctive elements which are part of a composite mark, even though they may appear completely marginal .
Composite trademarks composed of various name-related or figurative elements are frequently found in the field, with more than one element having a distinctive capacity which nevertheless misses a significant role in terms of characterisation of the mark. This role is essentially played only by some, or only by one, of these elements.
The mechanical application of the analytical perspective discussed here above, which did not occur in the Medion case, may cause some completely unsatisfactory results from the standpoint of the general interest as it would seem to legitimise that within a trademark composed of twenty elements, with a large and impressive one and nineteen small and marginal ones, it would be sufficient for confirming the likelihood of confusion between trademarks if only one of the nineteen marginal elements was used in a second mark.
In order to get around this clear inconsistency, the Court placed priority on the role of the autonomy of the distinctive position of the various mark elements (see point 30).
More precisely, the Court confirmed that, in judging the infringement of a simple mark by a subsequent composite mark that contains it, it is necessary to consider the autonomy of the distinctive position of the previous mark in the subsequent one, regardless of the circumstance that the earlier mark is also dominant in the overall impression of the second mark. The Court thus considered the dominance of a mark inside a composite mark along with the emphasising of its autonomous distinctive position relating to its distinctive character, which is fully in accordance with the fundamentals of trademark law.
This approach is consistent with the criteria of the so-call “core of the mark” one (cuore del segno) adopted by a stream of national decisions , which aims for the purposes of protection at the selection, among the various distinctive mark components revealed following an analytical dissection, of the ones and only which most characterise it, thus safeguarding the subsequent marks which adopt one or more of the marginal ones .
Thus the problem becomes actually more that of a correct definition of the category of composite trademarks.
In fact, in relation to composite marks whose elements may lack in distinctive capacity, either in themselves or in relation to the context in which they are inserted, there would be no reason to consider the problem of the autonomous protectability of the single components nor to resort to rules other than those commonly followed for simple marks which emphasise the canon of synthetic judgement.
On the contrary, within an actual composite trademark composed of various elements all bearing the same distinctive character, it does not appear possible to isolate the core of the mark precisely because all the single elements are core, potentially calling for analytical consideration and protection .
Based on what is proposed and on the ruling in the Medion case, the category of composite trademarks might shrink if compared to the size currently envisaged even by the EU jurisprudence , in light of the necessity to reserve the protection resulting from the principle of protection of the single distinctive components confirmed by the Medion case for those marks which actually consist in a bunch of autonomous trademarks. And this appears not to be a merely terminological question since the judicial qualification of the trademark in hand, which nevertheless could only occur following an examination of the concrete case, would result in a type of protection, that of the single elements, which is affordable only within restricted limits and would otherwise appear disproportionate when compared to the interests of the trademark owner, on the one hand, and of third party market operators, on the other.
The criterion of analytical judgement of the distinctive components of the composite trademark adopted in the Medion case seems to me the only one which allows protection regardless of the concrete way of their insertion within the trademark of a possible infringer, by way of their combining, for example, with the indication of the producer, with a third party trademark or, in general, with other marks which in an overall appraisal could lead to excluding a general impression of similarity between the two marks .
Thus not only does the emphasis of the analytical level in comparing the marks seem consistent with the need for abstract protection of the mark , but it also seems to be more in line with the need to protect the message that the trademark is the carrier of which, in the case of composite trademarks, may not be single, but breakable down into its different pieces .
Further developing the reasoning, common and autonomously protectable portions within different marks, regardless of the context of insertion planned by the third party, may produce a risk of association according to art. 1.1, letter b) IPC, at least from the viewpoint of those who sustain that association is independent from a risk of confusion for the public , as well as an appropriation of the suggestive values intrinsic to a well known trademark (art. 20.1 letter c, IPC) where there is no doubt that the issue is not the one of confusion even in the case of similar products .
 I refer, without claiming to be complete due to the number of the rulings, to CFI 23 October 2002, case T-6/01 (Matratzen Markt Concord with figurative part and Matratzen); CFI 23 October 2002, case T-104/01 (Miss Fifties with figurative part/Fifties); CFI 12 December 2002, case T110/01 (Saint Hubert/Hubert); CFI 3 July 2003, case T-129/01 (Bud/Budmen); CFI 9 July 2003, case T-156/01 (Giorgi with figurative part/Giorgio Aire); CFI 9 July 2003, case T-156/01 (Giorgi with figurative part/Giorgio Beverly Hills); CFI 4 November 2003, case T-85/01 (El Castillo with figurative part/Castillo); CFI 25 November 2003, case T-286/02 (MOU, Kiap MOU); CFI 18 February 2004, case T-355/02 (FLEX with figurative part/CONFORFLEX); CFI 3 March 2004, case T-355/02 (Sir with figurative part/ZIRH); CFI 17 March 2004, case T-183/02 and 184/02 (Mundicolor and Mundicor); CFI 31 March 2004, case T- 20/02; (HAPPIDOG/HAPPY DOG with figurative part ); CFI 22 June 2004, case T-185/02 (Picasso and Picaro); CFI 22 June 2004, case T66/03 (GALA/Galaxia with figurative element); CFI 30 June 2004, case T-317/01 (EURODATA TV/M+M EUROdATA).
 Cf. ECJ 29 September 1998, C-39/97 (Canon); ECJ 22 June 1999, C-342/97 (Lloyd)
 Cf. ECJ 11 November 1997, C-251/97 (Sabel); ECJ 29 September 1998, C-39/97 cited; ECJ 22 June 1999, C-342/97 cited; ECJ 22 June 2000, C-425/98 (Mode Trademark).
 Cf. the rulings cited in note 3 above.
 Cf. ECJ 11 November 1997, cited; ECJ 29 September 1998, cited.
 The model is substantially the same whether likelihood of confusion is judged on for the purposes of evaluating the novelty of the mark or its infringement.
 In addition to the ECJ preliminary rulings concerning interpretation of EC directive 89/104 cited in the previous notes and starting with ECJ 11 November 1997, C-251/97 (Sabel), in opinion point 23, the following rulings of the CFI concerning the registration of EU trademarks have to be added: CFI 23 October 2002, case T-6/01 (Matratzen), point 75; CFI 23 October 2002, case T-104/01 (Miss Fifties), point 34; CFI 12 December 2002, case T110/01 (Hubert), point 48; CFI 3 July 2003, case T-129/01 (Budmen), point 45; CFI 9 July 2003, case T-156/01 (Giorgio Aire), point 64; CFI 9 July 2003, case T-156/01 (Giorgio Beverly Hills), point 39; CFI 4 November 2003, case T-85/02 (Castillo), point 39; CFI 25 November 2003, case T-286/02 (Kiap MOU), point 38; CFI 18 February 2004, case T-355/02 (C), point 43; EU, point 43; CFI 3 March 2004, case T-355/02 (ZIRH), point 42; CFI 17 March 2004, case T-183/02 and 184/02 (Mundicor), point 77; CFI 31 March 2004, case T- 20/02 (Happy Dog), point 40; CFI 22 June 2004, case T-185/02 (Picaro), point 53; CFI 30 June 2004, case T-317/01 (M+M EUROdATA), point 48; CFI 22 June 2004, case T66/03 (Galaxia), point 23; CFI 24 november 2005, case T-3/04 (KINJI SPA/ KINNIE), point.
 This impression is confirmed by a reading of a ruling which examines the possibility of registering a composite mark under the profile of its distinctive character. The Court states that: “…art. 7 no. 1 lett. b) CTM regulation makes no distinction between different categories of trademarks. The evaluation criteria of the distinctive character of trademarks composed of figurative elements or by a combination of name-related and figurative elements are not thus different, from those applied to other categories of trademarks. In addition, if it involves a trademark composed of various elements (hereinafter: the “composite trademark”), it is necessary to consider it as a whole for the purposes of evaluating its distinctive character. However, this is not incompatible with a later examination of the various elements the trademark is composed of” (CFI 5 December 2002, case T-91/01 BiolD trademark point 27 and 36 of the opinion).
 The point was exactly acknowledged by the ECJ in the Thomson ruling. Cf. § 3 below.
 An example taken from reality is precisely the one of Life/Thomson Life. Completely imaginative examples can be produced, such as: Toblerone Suchard/ Toblerone Ferrero, but the case law has also addressed similar cases such as those of Cass., 14 July 1987, no. 6128, in Giur. ann. dir. ind., 2101, where the trademarks were Mistero e Mystère de Rochas, respectively; Court of Milan, 4 March 1999, 1999, p. 977 ss, where Agip disputed the well-known trademark composed of a dog with six legs against a competitor who had reproduced it identically on a t-shirt, only changing the words “agip self-area” with “acid self-area”.
 CFI 23 October 2002, case T-6/01.
 Cf. the examples in note 10 above.
 See CFI 25 November 2003, case T-286/02, points 43-45 (Mou-Kiap Mou); CFI 23 October 2002, case T-104/01 (Miss Fifties with figurative part compared to Fifties); here the composite trademark was the previous one, as well as in CFI 18 February 2004, case T-10/03, (Flex, with figurative part, and Conforflex); instead in CFI 31 March 2004, case T-20/02, Happy Dog, the composite trademark was the later one; CFI 4 November 2003, case T-85/02 (Castillo); CFI 30 June 2004, in case T-317/01 (Eurodata TV, M+M EUROdATA). However the Court sticked to the ‘overall impression’ rule and confirmed that: “…the existence within the marks of other verbal elements makes the overall impression provided by each mark be different” (point 72 of the last cited ruling); CFI 24 November 2005, case T-346 (Arthur, Arthur et Felicie); CFI 24 november 2005, case T-135/04 (Online Bus- BUS).
 See CFI 23 October 2002, cited (Matratzen); CFI 9 July 2003, in case T-156/01 and CFI 9 July 2003, case T-162/01 (Miss Giorgi and Giorgi Line compare to Giorgio Aire and Giorgio Beverly Hills); CFI 24 November 2005, case T- 3/04.
 ECJ October 6, 2005, case C- 120/04.
 See, among the most recent, Supreme Court 22 February 1994, no. 1724, in Giur. ann. dir. ind., 3021/7; Supreme Court 25 October 1978, no. 4839, in Giur. ann. dir. ind., 1014/2; Court of Appeal Bologna, 29 May 2002 4438/1; Court of Rome, 30 July 2001 4360/2; Court of Aquila, 21 July 1998 3937/2; Court of Bassano del Grappa, 15 December 1990, 3032/2; Court of Appeal Florence, 9 September 1994 3252/4; Court of Milan, 5 March 1990, 24521/2; Court of Appeal Turin, 27 July 1989 2438/1; Magistrate’s Court Milan, 13 December 1988 2356/2; Court of Naples, 8 January 1988, in Dir. e giur., 1990, page 142; Magistrate’s Court Rome, 23 April 1987, in Giur. ann. dir. ind., , 2165/2; Court of Rome, 28 September 1983 1689/4; Court of Appeal Naples, 22 December 1982, 1982, page 669; Court of Modena, 28 April 1982 1543/2; Court of Catania, 23 May 1978, 1054/2; Court of Appeal Milan, 28 November 1972, 207/1; Court of Appeal Milan, 22 September 1972, 171/4. Aggiornare
 The principle is taken from Supreme Court 16 October 1969, no. 3343, in Mass. Giust. Civ., 1969, page 1698, following Supreme Court, 27 July 1938, in Giur. It. Mass., 1938, 806. Accordingly Supreme Court, 24 May 1969, no. 1833, in Giur. It. Mass., 1969, 763; Supreme Court 11 May 1982, no. 2929, in Giur. ann. dir. ind., 1469: “The inclusion of the element which characterises a previously registered trademark in a composite trademark constitutes encroachment, even if the later composite trademark is composed of other elements which differentiate it from the previously registered trademark,” also published in Foro. It., 1982, I, 1891; Supreme Court 14 July 1987, no. 6128, in Giur. ann. dir. ind., 2101: “infringement of a previously used or registered trademark may exist even if the reproduction is inserted in a composite trademark.” In same sense see also Court of Milan, 17 November 1980, in Riv. dir. ind., 1982, II, page 42 ss; Court of Milan, 18 May 1978, in Giur. ann. dir. ind., , 1052/1; Court of Milan, 31 January 1977, in Giur. ann. rep. sist., 1977, page 977. The principle dates back in time: Court of Milan, 15 December 1947, in Rass. Propr. ind. lett. e artistica, 1948, page 174 ff., 177; Court of Turin, 21 January 1949, page 125; Court of Appeal Milan, 18 May 1942, in Mon. Trib., 1942, page 355.
Reciprocally, infringement takes also place by way of a partial imitation of a composite trademark involving its characteristic parts, i.e. those which serve a specific individualising and differentiating function (Supreme Court 3 June 1968, no. 1680, in Giur. It. Mass., 1968, 598). See also Court of Verona, 26 March 2003, in Giur. merito, 2003, I, page 1970.
 Court of Appeal Milan, 30 January 1996, in Giur. ann. dir. ind., 3461.
 Court of Appeal Milan 22 September 1972, cited; Court of Milan, 30 December 1999, in Riv. dir. ind., 2001, II, page 265; more recently cf. Court of Modena, 28 January 2003, in .
 This involves emphasising the likelihood of confusion between the marks, i.e. the risk of confusion, and not the confusion which is actually produced. This the content of art. 20.1 letter b) IPC (IPC stands for Intellectual Property Rights Code, enacted with legislative decree 30/2005). The principle is much older if the Supreme Court as early as 12 April 1937, in Giur. It., 1937, I, 1, 1122, 378, confirms that: “For the purposes of a trademark infringement the fact that the claimed imitation gave rise to product confusion or not is irrelevant, since confusion is only relevant for the purposes of unfair competition.” On this point see also SENA Confondibilità fra segni e confondibilità fra prodotti o servizi nella giurisprudenza comunitaria, cited, pp. 203-204.
 See AULETTA, Somiglianza dei marchi complessi, in Riv. dir. comm., 1953, II, 253 ff., in part. 259.
 See Supreme Court 14 July 1987, no. 6128, cited; Supreme Court 11 May 1982, in Foro It., I, 1891-1892; Supreme Court 16 October 1969, no. 3343, in Foro It. Rep. 1969, item Marchio, no. 63-64.
 In an old precedent (Court of Milan, 15 December 1947, cited), it is stated that “a point has already been settled in the literature and law, for which the right to the exclusive use of the trademark by its owner is similar to a property right on the whole of the characteristic and composing elements of the trademark and on each of these, whether they are of an emblematic nature or name-related elements, and it must be ascertained if the reproduction, or even imitation of one or more of these characteristic, emblematic or name-related elements, bears in itself the possibility of confusing effects occurring”.
 See Court of Appeal Bologna, 29 May 2002, cited. The principle is common in many rulings.
 Cf. RICOLFI, I segni distintivi. Diritto interno e comunitario, Turin 1999, pp. 130-131; ID, in Diritto industriale. Proprietà intellettuale e concorrenza, Turin, 2001, pp. 136-137; BIGLIA, Alcuni problemi sull’affinità fra prodotti e sul giudizio di confondibilità fra marchi, in Riv. dir. ind., 1982, II, pp.42 ff., 47.
The idea goes back to an opinion formulated by Rotondi according to which: “…In the hypothesis of a registered trademark, the owner has the right to the exclusive use of each single emblematic or name-related element included in the trademark, and to use it even separately. Thus the reproduction or imitation of only one is enough to be considered, without doubt, the specific illegal act which is constituted by infringement. Only one limitation can and must be made, i.e. in relation to the partial reproduction or imitation which involves those elements of the trademark which in themselves do not have a character of innovation…” ROTONDI, Osservazioni sull’elemento della confondibilità, cited, 1128.
Favoring an ‘overall evaluation’ approach: FERRARA Jr., La teoria giuridica dell’azienda, Florence 1945, page 256; SORDELLI, Inesistenza della nullità parziale nell’ipotesi di marchio complesso: osservazioni sull’<<oggetto>> del marchio, in Riv. dir. ind., 1961, II, 213 ff., 221-222; AGHINA, Sui marchi complessi, in Riv. dir. ind., 1965, II, page 85 ff., 99.
 However it must be clarified that the meaning of the rule of autonomous protectability of the single components has never been the one of permitting an exclusive right to cover just any component of the trademark, unrelated to the evaluation of its originality. It was the idea of the author to whom the opinion is attributed that the right to the trademark “…est totum in toto, ed totum in aliqua parte” (BOSIO, Trattato dei marchi e segni distintivi di fabbrica secondo la legge italiana e il diritto internazionale Della contraffazione e della concorrenza sleale, Turin 1904, pages 424-425), who writes that: “… the crime of infringement, in its various forms, occurs even for the partial use of the trademark by others, with the understanding that the part which is appropriated must be the essential and integral part of the trademark” (emphasis added). ROTONDI, Osservazioni sull’elemento della confondibilità, cited, 1130, expresses the same concern.
 For some precedents see Supreme Court, 22 February 1994, no. 1724, in Giur. ann. dir. ind., 3021/7; Court of Verona, 26 March 2003, cited; Court of Asti, 29 July 1998, 3834/1; Court of Appeal Milan, 30 June 1995 3428/2; Court of Bologna, 8 November 1993 3080/2; Court of Bassano del Grappa, 15 December 1990, 3032/2; Court of Appeal Catanzaro, 15 December 1989 2459/4; Court of Naples, 8 January 1988, in Dir. e giur., 1990, page 142. Aggiornare
 When the marginal components are, even if scarcely, distinctive, the result of the judgement aimed at searching for the “core of the mark” coincides with the judgement that, in comparing the marks, isolates the single elements with a prevalent distinctive capacity in the sense of attracting the consumer’s attention. Cf. Supreme Court 26 April 1979, no. 2417, cited; Supreme Court 26 May 1969, no. 1833, cited; Supreme Court 3 June 1968, no. 1680, cited; Supreme Court 23 April 1965, no. 718, in Giust. Civ., 1965, l, page 1846, in particular 1848; Supreme Court 15 June 1957, no. 2290, in Giur. It. Mass., 1957, 513; Supreme Court 26 July 1932, cited; Court of Milan, 15 November 1971, in Giur. ann. dir. ind., 42/2.
 The concept of the composite trademark as a group of autonomous trademarks could have an impact on its nullity. Based on the principle of unity, the literature which has dealt with composite trademarks has primarily concluded in the sense that partial nullity is not admitted. On the other hand former trademark law, and currently the c.p.i. (see the articles 25 and 27), do not provide for partial nullity, which has neither been recognised by case law. See Supreme Court 22 February 1961, no. 414 in Riv. dir. ind., 1961, II, p. 213, with an assenting note by SORDELLI, Inesistenza della nullità parziale nell’ipotesi di marchio complesso, cited, page 221; Court of Appeal Lecce, 27 July 1972, cited; see page 1136; Court of Turin, 19 February 1965, in Riv. Dir. Ind., 1965, II, 85 ff., with an assenting note by AGHINA, Sui marchi complessi, cited. Note that, as long as some components are judged not protectable because they lack in distinctive character, the question of ascertaining their nullity seems to be a merely formal one.
If the complex mark is composed of marks with autonomous distinctive capacity, the declaratory judgement of nullity of one component could let the remaining mark survive. This possibility was acknowledged by the Court of Appeal Rome, 19 April 2004, with reference to the ‘Mario Valentino Costa Smeralda’ and ‘Mario Valentino’ trademarks, soon to be published in Giur. and. dir. ind., 2004.
 When only one element is judged as protectable, the trademark is basically simple even if it appears composed of other insufficiently distinctive elements. This was not the case in Medion/Thomson, where the Court of Justice considered the trademark as composite as it consisted in the juxtaposition of two autonomously distinctive components.
 Imagine also the use of a different form of presentation of the product, even in the presence of a confusingly similar trademark, but also the case of the use of only the name-related or emblematic part of a composite mark (for example the Lacoste trademark) or that of a single component of a name-related mark composed of words with independent meaning (Dolce e Gabbana, Viktor and Rolf).
 See SENA Confondibilità fra segni e confondibilità fra prodotti o servizi nella giurisprudenza comunitaria, cited, pp. 203-204.  In this light it must be considered that it is with reference to name-related composite marks that the plurality of distinctive components generally corresponds to a plurality of messages. On the other hand, in the case of name-related and figurative composite marks the message is often a single one. In the latter case, the use of one of the two trademark components could however lead to an association (regardless of the fact that the component was also registered as such). In this connection see, with different outcomes, Court of Milan, 12 July 1999, in Giur. ann. dir. ind., 4017, related to the crocodile in the Lacoste trademark and Court of Milan, 4 March 1999, 3897, related to the six-legged dog of Agip.
 On the notion of risk of association compared with risk of confusion see the different opinions in GALLI, Funzione del marchio ed ampiezza della tutela, cited, p. 188 ff.; and MANSANI, La nozione di rischio di associazione fra segni nel diritto comunitario dei marchi, in Riv. dir. ind., 1997, I, p. 133 ff.; ID. La funzione di indicazione d’origine del marchio nell’ordinamento comunitario, Milan 2000, pp. 67-69.
The ECJ confirmed that the risk of association does not represent an alternative notion to that of confusion, but is part of it when the consumer may feel that the products bearing confusing marks come from sources which are somehow connected to each other by some kind of relationship. Thus, not any type of connection between the two marks does seem relevant for the ECJ, even in the subform of a mere mental association, when there is no risk of confusion at least in the broad sense described above. See MANSANI, Recenti orientamenti della Corte di Giustizia CE in materia di marchi, in Studi parmensi, 2002/1, pp. 43 ff., 44-45.
 As confirmed by the ECJ in the Adidas case which interpreted the protection under art. 5.1 letter c) of EC Directive 89/104 (corresponding to art. 20.1 letter c, IPC) in terms of a general need of protection of the well-known trademarks suggestive messages, which goes well beyond the specific case of the use of a well-known trademark on non similar products. See GALLI, L’allargamento della tutela del marchio e i problemi di Internet, in Studi parmensi, 2002/1, pp. 17 ff., 29 ff.