Mapping Judicial Dialogue Across National Borders: An Exploratory Network Study of Learning from Lobbying Among European Intellectual Property Judges

This paper looks at dialogue and collective learning across borders through personal networks of judges. We focus on judges participating in the Venice Forum, bringing together European patent judges involved in institutional lobbying for the construction of a European Patent Court. Empirical observation shows that personal networks of discussion with foreign judges, reading of their work and references to their decisions do exist in this milieu and can be mapped. Our network study shows that judges from some European countries are more active in this dialogue than judges from other countries. The learning process is driven, to some extent, by a small subset of super-central judges who frame this dialogue and can be considered to be opinion leaders in this social milieu. We measure a strong level of consensus among the judges on several controversial issues surrounding the procedure of a possible future European Patent Court. But strong differences between them remain. Dialogue and collective learning do not, by themselves, lead to convergence towards a uniform position in these controversies.

what others have done. But it is difficult to analyse. If you are not very well informed about the ins and outs of German law and its applications in that country, you cannot analyse a German decision properly. The most useful thing is (…) to go sit down with the German judges and participate in their decisions, see how they do it, and then compare: would we have done it the same way or not? This is more effective than just reading.' Looking at the context in which dialogue and learning take place across borders is thus crucial. This context is itself structured by differences, debates and competitions of many kinds, including regulatory competition, in ways of carrying out fundamental tasks, such as building legitimacy, creating coherence and respecting plurality. 1 Our approach in this paper will not deal with these general issues. Instead, in order to understand dialogue and the process of learning across borders, we focus on a case in point and a specific dimension of this context: social exchange among judges. We ask what is the relationship between actors' positions in the structure of international social exchanges among national court judges and their learning behaviour? Are stands taken by individual judges on various issues or controversies in their fields of specialty correlated with their position in this structure? Is this structure highly centralized or rather highly polarized, and what are the consequences of each situation for these processes? These questions lead to an additional issue, that of the capacity of such networks to contribute to the creation of a collective and international frame of reference for the judiciary, including but not exclusively high court judges, for example for judges specialized in the same substantive areas and following each other's work. As part of this wide issue of learning and influence, we wish to map such networks by describing, in an exploratory way, 'How do judges look at the work of other judges at the international level? ' We thus focus on invisible aspects of judicial internationalisation rather than the more visible legal phenomena (e.g. judicial decisions, legislation, or even formal and official associations such as that described by Claes and de Visser. 2 Indeed, essential to this purpose is the mapping of the various means and resources of which courts avail themselves to retrieve and exchange information beyond their own borders. Occasionally such means are formally institutionalised, e.g. in the form of electronic databases or periodic meetings for the exchange of information. Often however they are opaque and implicit, utilized in such a natural and subconscious fashion that one must actively work to identify them. We therefore map out various ways in which judges conduct dialogue with their foreign peers. We also analyse the relevant results and make systematic inferences about the general tenor and character of transjudicial dialogue. Our mapping can thus help study usually invisible 'transjudicial borrowing ' and 'transjudicial dialogue' . 3 In particular, based on previous work on the balance between cooperation and competition characterizing collegiality, 4 we add a third dimension to Claes and de Visser's approach to this dialogue. Instead of stressing conversation and deliberation, we put more emphasis on institutional entrepreneurship in which judges can be involved in a bottom-up collegial, regulatory process underlying pluralism, 5 professional and political, as well as regulatory competition.

Research method
We use social network analysis to map a social network of judges, including High Court judges, that describes the ways in which they look at the work of other judges across borders and learn from each other. Social network analysis is a method that reconstitutes the system of social exchanges among members of a predefined population. In our case, the exchanges are exchanges of information and advice. We would like to explore questions such as 'Where do High Court judges get their information (about decisions made elsewhere) from? Who are the judges talking to, or seeking advice from? Whose work abroad are they aware of? Who cites whom, if at all?' . Each of these questions would help reconstitute a separate network; each network can be analysed separately or jointly with others.
In this context, learning does not take place as it would at school. It happens in relation to work performance and to common projects raising controversies when judges from different countries and legal cultures have different perspectives on such projects. In addition to the reconstitution of an international social network among judges, we would like -at this exploratory stage -to look at more dependent variables such as position taken in such controversies. This would provide examples of the effect of existing social networks on ongoing debates in learning and influence within this population.

Population and fieldwork: the example of international patent judges
Given the difficulty of asking network questions to justices in general, we used the following strategy 6 for the reconstitution of some of these networks. We focus on an association of specialized judges who meet and exchange across borders, at the European level. We were able to locate the Venice Forum, an international association of strongly committed Intellectual Property (IP) judges, in particular patent judges (35 judges comprising a mix of judges from highest courts and lower courts). They meet annually in Venice, convened by the European Patent Lawyers Association (EPLAW) together with the European Patent Office (EPO) and the Intellectual Property Judges Association (IPJA). 7 It is also a heterogeneous set of judges in the sense that the structure of their career differs in various European countries and legal cultures. Europe has currently 27 jurisdictions and 27 kinds of patents decisions.
The Venice Forum judges represent courts involved in IP litigation. One has to be invited by EPLAW in order to participate. There are country delegations, usually a couple of judges from each country, except for the largest countries or the more active ones in the field. Pre-selection is mainly carried out by EPLAW, whose president picks out the judges who will come and contribute. Four Dutch judges were present in Venice in 2009 whereas in the Netherlands there are about 10 to 12 such judges. They are the same 'kind' of judges from all countries: they are not representatives, but rather activist judges concerned with patent issues and about the emergence of the European Patent Court (EPC). Some come to participate, together with EPLAW, in shaping the procedure of the future EPC; others accept the invitation because they want the voice of their country to be heard.
Indeed, the main background feature of this meeting is the long and tortuous process of creation of the European Patent Court. The meeting itself was part of an ongoing process of lobbying Brussels for the construction of the EPC and definition of its procedure. This lobbying was seen by corporate lawyers and judges alike as important for the European economy, in which entire industries are built on business models with patents at their core. As we shall see, a small number of countries are pushing harder than others in the lobbying process.
The social construction of a European Patent judiciary as a bottom-up process begins with transnational interactions and dialogue between these judges who can be considered to be 'institutional entrepreneurs' . This set of actors is therefore quite specific, closed and selective. We were not allowed to attend their meeting per se but we could see that fierce discussions could take place -although no-one 6 Another strategy is to look at citation networks among judges. However, that is not necessarily much easier, especially because judges do not always have the same citation practices in all countries. We know for example that French judges do not cite foreign statutes or cases for reasons of their own (for example they allegedly do not want to cite statutes that might become obsolete). 7 The European Patent Lawyers Association (EPLAW, formerly EPLA) is a professional association of patent lawyers, with a registered office in Brussels. Its stated object is 'the promotion of efficient and fair handling of patent litigation in Europe and the strengthening of ties between well established patent lawyers in Europe'. In a conference organized by the EPLAW in October 2005 in Venice, a panel of judges adopted a resolution calling for a single European patent court. In a second conference, called the Second Venice Forum and jointly organized by the EPLAW and the European Patent Office in 2006 in San Servolo, an island in the Venetian Lagoon, a panel of judges approved a Resolution adopting guidelines for the Rules of Procedure of the European Patent Court to be set under the European Patent Litigation Agreement (EPLA), which has yet to enter into force. Since 2001, the EPLAW has also taken a series of resolutions relating to European patent law.
left shaking their head. 8 We carried out the survey nevertheless because we hoped that the heterogeneity of this subset of judges was sufficient to bring out regularities in the pattern of exchanges and show the diversity of approaches towards learning -even if this learning is inextricably linked to controversy with respect to the choice of procedure to be implemented at the EPC. Many patent judges in this association have more or less frequent bilateral relations, exchanging views and problems. Fieldwork to reconstitute the network that is created by these relations was carried out in Venice in October 2009. We took advantage of the existence of this annual meeting in Venice in order to carry out our interviews. We submitted this survey to all participating judges at the European Judges Forum (San Servolo Conference, 30-31 October 2009). At the beginning of our fieldwork, we did not know if individual patent judges would be prepared to answer our network questions. We received a relatively open attitude in general, once the judges knew that the interviewers were scholars. The survey received a 100% participation rate. The questionnaire is presented in Appendix 1.

Maps of dialogue across borders as measured by social networks
Mapping flows of information among judges identified by their country was carried out using social network analysis and qualitative face-to-face interviews. The reconstitution of the network used sociometric questions listed in Appendix 1. We also asked the judges about their preferences in controversial issues regarding the construction of a harmonized procedure for the European Patent Court. In order to understand the learning and influence processes that drive internationalization of law, we then link position in the network and preferences with respect to the controversy.
We find that there is a hierarchy among various forms of network learning across borders. Three networks were measured among IP European judges at the Venice Forum: 'personal discussion network' , a 'reading other judges' work' network and an 'explicit reference to other judges' decisions' network. Reconstitution and analysis of these networks shows that the discussion network is denser than the reading network, which is denser than the explicit reference network. There is much more activity in direct personal discussion with colleagues across borders (for example at events such as the Venice Forum), than with actual reading of their work (decisions and articles); there is quite little explicit reference to other foreign judges' decisions in these judges' own decisions. In other words, at this stage, learning across borders occurs more through discussion than through reading, and more through reading than through explicit reference to work of other judges (which is not allowed in some countries).
Some countries are more active in the learning process among this set of judges: Dutch, UK, German, Italian and French judges are the most active in the personal discussion network and in the reading network. UK, German and Dutch judges are the most active in the explicit reference network. Dutch, UK and German judges display the highest activity in all three networks (reading, discussion and explicit reference). Italian and French judges also display high activity in the reading and discussion networks, but not in the explicit reference network. Note that explicit reference to work of other judges is not permitted in some countries (for example France).
The analysis of these networks also shows that learning across borders depends on the existence of highly active and central judges in a core-periphery structure in which the periphery is itself fragmented. A small subset of highly central judges in this system can be identified as opinion leaders among their colleagues. Indeed some networks are more centralized than others. The central judges facilitate learning through personal discussions, through reading other judges' work, and through explicit citation of other judges' work. This highly central subset of opinion leaders also increases transparency by referring to decisions of foreign courts. Colleagues recognized as opinion leaders come from different countries depending on the kind of network: French, Dutch, Italian, German, UK and Swiss judges are among the most central in the personal discussion network. Dutch, German, and UK judges are among the most central in the reading network. UK, German and Dutch judges are the most central in the explicit reference network. Opinion leaders who are central in all three networks tend to be UK, German and Dutch judges.

Figure 1a Visualization of the personal discussion network among European IP judges belonging to the Venice Forum
Legend: Individual judges are identified by their country of origin and countries are colour coded. Circle size represents each judge's centrality. Highly central judges (identified by the symbol *) are those most cited by colleagues (i.e. number of citations above the 90th percentile in at least two networks). To protect anonymity of respondents, individuals who are single representatives of their country in this group are clustered in the white (X) subgroup.

Figure 2 Visualization of the direct reading network among European IP judges belonging to the Venice Forum
Legend: Individual judges are identified by their country of origin and countries are colour coded. Circle size represents each judge's centrality. Highly central judges (identified by the symbol *) are those most cited by colleagues (i.e. number of citations above the 90th percentile in at least two networks). To protect anonymity of respondents, individuals who are single representatives of their country in this group are clustered in the white (X) subgroup.

Figure 3 Visualization of the explicit reference network among European IP judges belonging to the Venice Forum
Legend: Individual judges are identified by their country of origin and countries are colour coded. Circle size represents each judge's centrality. Highly central judges (identified by the symbol *) are those most cited by colleagues (i.e. number of citations above the 90th percentile in at least two networks). To protect anonymity of respondents, individuals who are single representatives of their country in this group are clustered in the white (X) subgroup.

Dialogue and controversial issues among IP judges
In order to measure a possible outcome of this learning process, we look at whether or not there is consensus among these judges with respect to controversial issues concerning patents and procedure for the future EPC: assessment of inventive step, determination of scope of protection, involvement of technical experts, and personal rule applied to define what a patent is. When assessing inventive step, most judges (76%) declare that they -explicitly or tacitly -apply the same method as the Examining Divisions and Boards of Appeal of the EPO (problem-and-solution approach). Almost all judges (91%) think that decisions of foreign courts in relation to the same patent with which they deal are indeed relevant. A strong majority (67%) also declares that they do refer in their own decisions to decisions of foreign courts. When determining a scope of protection, a majority of judges (64%) considers that the statements of the applicant during the grant procedure before the European Patent Office does play a role. In the mind of 65%, this could only lead to a limitation of the scope of protection. A majority of judges (60%) involves independent technical experts to report on inventive step. Only 49% involve independent technical experts to report on scope of protection. Almost all judges say that the parties have the opportunity to comment on the reports of the experts.
For choosing between two different rules to apply when defining patents, 46% of the Venice Forum judges consider patents are exceptions to the freedom of copying. Therefore, they think that the validity of patents and their scope of protection should be critically assessed. 27% of the judges support another definition, for which patents are granted as a reward to the contribution of the inventor to technological development, with the implication that the patent holder is entitled to a mild assessment of the validity of the patent and a broad scope of protection. 21% hesitate. 74% are personally quite happy with their choice.
In sum a high level of consensus exists among Venice Forum IP judges on several issues (assessment of inventive step, problem-and-solution approach, relevance of foreign decisions), but strong differences remain: there is a heterogeneous use of judicial discretion in balancing 'patent as exception' and 'patent as reward' . There is great diversity with respect to narrow versus broad scope of protection. A real risk exists that differences among the practices of these judges could lead to diverging decisions. This diversity leads to the question of whether learning across borders leads to uniform positions among judges and to the role of opinion leaders in this harmonization.

Dialogue, networks and controversies
The analysis shows that, on average, opinion leaders in this learning process have positions different from less central IP judges with respect to five issues. Figure 4 illustrates these differences.

Figure 4 Comparison between average positions of opinion leaders versus average positions of all other judges with respect to five issues
It shows that highly central judges tend to use the problem-and-solution approach less systematically than all others. They refer to foreign courts in their own decisions (with exceptions, for example, in France, Spain, and Romania) more than all others. They consider that 'statements of applicant during grant procedure play a role when determining the scope of protection' less than all others. They involve independent technical experts to report on inventive step less than all others. They involve independent technical experts to report on scope of protection less than all others. One may therefore hypothesize a future convergence of judges in this network towards a UK-German-Dutch position if judges follow their highly central colleagues (opinion leaders identified above). A final map ( Figure 5) represents judges perceived by their peers as closest to a future EU uniform position, i.e. as reflecting in their discourse the future European norm.

Figure 5 Judges perceived by their peers as closest to future EU uniform position with respect to patents
Legend: Countries are coded by colours, and circle size represents each judge's centrality. In this graph, highly central judges (identified by symbol *) are those most cited by colleagues (i.e. number of citations above the 90th percentile in at least two networks). To protect anonymity of respondents, individuals who are single representatives of their country in this group are labelled X. Figure 5 shows that the Venice Forum judges perceive among themselves at least six judges, from five countries, as colleagues whose positions on controversial issues are likely to reflect the future EU uniform position. In this network, these judges are Dutch, German, British, Italian and Swiss. Some of them are already central in the social networks mapped above. Others have different ideas with respect to 'harmonization' of positions in Europe, stressing 'circulation' instead of convergence, even if this leads to diverging decisions. Thus, in 2009, convergence towards consensus on the EU uniform position in this network still remained uncertain. We infer from this structure that learning through networks across borders does not necessarily, by itself, lead to convergence of perspectives and uniform positions among IP judges in Europe. As a consequence, our analysis provides some understanding of the process of the social construction of a common frame of reference among them but cannot yet predict the future European norm which will come out of this process.

Discussion and conclusion
This descriptive paper tracks dialogue and collective learning across borders through personal networks of judges in a case in point: patent judges brought together at an annual meeting by corporate lawyers involved in institutional lobbying for the construction of a European Patent Court with a specific procedure. Empirical observation shows that personal networks of discussion with foreign judges, reading of their work and reference to their decisions do exist and can be reconstituted with a degree of approximation. Our network study shows that judges from some countries are more active in this dialogue than judges from other countries. The learning process is driven, to some extent, by a small subset of super-central individuals who frame this dialogue and can be considered to be opinion leaders in this social milieu. We measure a strong level of consensus among these judges on several controversial issues surrounding the procedure of a possible future European Patent Court. But strong differences between them remain. Dialogue and collective learning do not, by themselves, lead to convergence towards a uniform position in these controversies. For example, on average, opinion leaders in this group have different positions from most judges with respect to several issues. In particular, opinion leaders, even if they increase transparency by referring to decisions of foreign courts, are not always considered by their peers to be closest to a future EU uniform position. Dialogue across borders does not mechanically create consensus. Our descriptive approach is not meant to answer the question why some individual judges and countries are more central than others in the learning process that is intertwined with a regulatory one. Many hypothetical reasons could be listed. One such reason might be that leading legal systems, or strong traditions in legal entrepreneurship in the business sector, structure that network. Another may be that some countries may invest in the patent judiciary more than other countries for institutional, economic and/or political reasons. For example, France has fewer judges specialized in intellectual property than other countries simply because judges in the French judiciary are usually generalists rotated from one specialty to another every three years. Thus highly specialized, stable and central British, German or Dutch judges would tend not to have comparable French alter egos in this network because the structure of judicial careers is different in France. A third reason could be that some senior individuals with staying capacity in this rather closed collegial milieu and with high centrality become particularly important as institutional entrepreneurs, and skilled at helping other colleagues in the network reach a certain level of comfort with specific solutions to common problems -a kind of judgment associated with learning by Bell. 9 Further research should improve the list of hypotheses and test them systematically. This would help disentangle various effects hidden behind the importance of specific individuals in the networks examined here, since these individuals represent various legal orders, legal systems, and positions in their countries and in Europe. This would also bring new insights into the effect of context on the learning process.
Our empirical network study of this milieu was also meant to evaluate the feasibility of this approach and provide tentative substantive results. Methodologically, we were able to identify several limitations for this kind of research. If 100% of the judges present in Venice in October 2009 responded to our questionnaire, it is because interviews were carried out face to face with our team of highly skilled colleagues. An online questionnaire, that was meant to extend the study to the judges from the 2008 Venice Forum, failed miserably. As a result, for example, German judges are under-represented in the survey compared with the number of German judges usually attending the meetings. One opinion leader, who was invited and expected at the meeting, did not come and did not participate in the survey.
If transnational dialogue and collective learning do take place among this specific set of IP judges, it is in fact inseparable from two phenomena. First, corporate lawyers, who represent corporate clients, play an important role in gathering the judges because they have an interest in knowing their judges personally, in getting used to their ways of making decisions, and in trying to get a few messages across to promote the interests of their clients. The network examined here was gathered by invitation from corporate lawyers (EPLAW) with help from EPO and IPJA. This in itself both makes measurement possible and raises measurement problems. The IP domain is one in which judges are most heavily courted by large industrial interest groups with patents at the core of their business model. Therefore several well-known IP judges in Europe never come to the Venice events because they consider them to be part of a political process in which they should not participate. Much learning takes place in such settings nevertheless.
Second, in our specific case in point, internationalisation in IP law, and business law in general, can be considered to be more advanced than in other areas of law, and more exposed to institutional changes. Indeed, the judges themselves are involved in a lobbying process in which they are institutional entrepreneurs in the construction of the European judiciary (a European Patent Court), with strong incentives to exchange with their peers and compete for influence on behalf of their country and/or on behalf of their own ideas and sense of professionalism. Some of them come because they fear that if they did not, the voice of their country would not be heard; in a context in which large IP cases are taken by multinational companies to Germany, the UK and the US, they think that it matters whether or not their country is represented in this setting. It remains to be seen whether this very specific process of dialogue and learning across borders -in which the political stakes and process are very present and visible -can be extended to other areas of law. Generalization of results obtained from this set of strongly committed IP judges to the population of all judges making IP decisions in Europe is therefore not established. So we believe that it would be useful to carry on examining bottom-up collegial and judicial dialogue across national borders in its context, while paying attention to the economic, social and political phenomena that drive it.

Introduction
This survey is part of a project run by the Hague Institute for Internationalization of Law. We are interested in the extent to which judges handling cases related to intellectual property participate in the internationalization of law. The project is carried out with the help of the University of Paris-Dauphine, the University of Utrecht and EPLAW. HiiL's website and Highest Courts project can be found at the following internet address: <http://www.hiil.org/research/main-themes/highest-courts/>.
Your responses are strictly confidential. They will be entirely anonymized.

Mapping the flow of information among European IP judges
The following set of questions is concerned with communication among judges in Europe. They will help us map the flow of information among them. Please look at the following lists of colleagues of yours who are invited to Venice this year or were invited last year.
* READ WORK: Would you go through the first list and check the names of colleagues whose work (decisions, articles) you have directly read?
* DISCUSSION: Would you go through the second list and check the names of colleagues with whom you have personally and directly discussed IP matters? * EXPLICIT REFERENCE: Would you go through the third list and check the names of colleagues whose work you refer to explicitly in your own decisions?
* UNIFORM: Would you go through the fourth list and check the names of colleagues who you think are usually closest to what could become the uniform European position with regard to intellectual property issues.

Issues
The next few questions are about your personal views on intellectual property related issues.

The assessment of inventive step
When assessing inventive step, do you -explicitly or tacitly -apply the same method as the Examining Divisions and Boards of Appeal of the European Patent Office (problem-and-solution approach)?

Yes No
Are decisions of foreign courts in relation to the same patent you are dealing with relevant? Yes No Do you refer in your decisions to decisions of foreign courts? Yes No