National and International Dimensions of the Triple Helix in Japan: University-Industry-Government versus International Co-Authorship Relations

Introduction

• Earlier studies treated international collaboration and Triple Helix relations separately, partly because identifying and cleaning institutional addresses and economic sectors requires substantial data processing.The authors note that even assigning institutional addresses within a country is non-trivial.Source: Introduction · p. 2
• The Japanese dataset attributes publications to university, industry, and government sectors and records international co-authorship, enabling three- and four-dimensional mutual-information analysis.The authors use it to specify interactions between national and international levels.Source: Introduction · p. 3
• The research asks whether relations among system dimensions reduce network uncertainty and which dimensions contribute most to that effect.The authors describe negative uncertainty as an indicator of synergy.Source: Introduction · p. 3
• Canadian data serve as a background check on whether main trends persist across different data types, not as the basis for a direct comparative study.The Japanese and Canadian datasets differ in their source and included publication types.Source: Introduction · pp. 3–4

Methods

• The method uses Shannon entropies and mutual information to quantify relationships between publication distributions; mutual information is zero for independent distributions and positive otherwise.The paper describes negative network-level uncertainty as configurational information that cannot be attributed to one contributor alone.Source: Methods · pp. 4–5
• For university, industry, and government, the three-dimensional indicator combines their individual, pairwise, and joint entropies and may be positive, negative, or zero.A negative value indicates reduced uncertainty in the configuration.Source: Methods · p. 5
• The four-dimensional indicator extends the calculation by adding foreign co-authorship and its entropies and interactions.The foreign dimension is denoted f.Source: Methods · p. 5
• Because the number of foreign papers without a Japanese address is unavailable, the analysis uses all papers with at least one non-Japanese address as a proxy.The total paper count is increased by this proxy for normalization in the international case.Source: Methods · p. 5

Data

• University-only papers fell from 69.4% in 1991 to 45.3% in 2004 as university scholars became more involved in collaborations.The share of papers with university addresses rose to 85.2% in 2004 after declining from 80.5% to 79.5% during the 1980s.Source: Data · p. 6
• Internationally co-authored papers and papers co-authored by university research and public-sector laboratories both increased from 4.0 (± 0.1)% in 1981 to 14.6 (± 0.1)% in 2004.Source: Data · p. 7
• Industry-only papers declined from 10.4% in 1991 to 3.5% in 2004, while university-industry co-authorship also decreased during the 1990s by less than one percentage point.The figure excludes co-authorship series below 3% of total publication volume.Source: Data · p. 7, Figure 1 · p. 7
• Triple Helix relationships in the Japanese publication system decreased after 1990, which the authors associate with declining university-industry co-authorship.The paper also reports that industry was the second-largest producer of academic papers until 1995, while industrial-government address combinations remained low.Source: Data · p. 8

a. The Japanese publication system

• Figure 2 shows a trend breach around 1995; the system’s loss of synergy is attributed not only to declining industry participation in the national portfolio but also to eroding university-industry co-authorship relations.Source: a. The Japanese publication system · p. 9
• The reported mutual-information trend is reflected in the national system with some delay.Source: Figure 2 · p. 9, a. The Japanese publication system · p. 10
• University-government mutual information remained approximately stable during 1993–2004, while industry-government mutual information was an order of magnitude lower throughout the period.Source: a. The Japanese publication system · p. 10

• By around 2000, Japanese national university-industry-government mutual information approached zero, while international participation increasingly reduced uncertainty from 1995 onward.Source: Figure 4 · p. 11, b. Inclusion of international… · p. 12
• Canada’s share of internationally co-authored university-addressed papers was 10% larger than Japan’s in the early 1980s, widening to 22.1% in 2004.Source: Figure 5 · p. 13, c. Comparison with Canadian data · pp. 13–14
• Only-government or only-industry publications were 10% more prevalent in Japan, while industry-academia co-authorship remained higher there than in Canada.Source: c. Comparison with Canadian data · p. 14
• Canada was more internationalized, but Japan had stronger national Triple Helix relations; 2004 industry-academia co-authorship was 5.4% in Japan versus 1.3% in Canada.Source: c. Comparison with Canadian data · p. 14
• Canada’s university-industry mutual information stopped declining after 1990, but university-government relations kept eroding; by 2004, the latter reduced uncertainty by about half its 1981 level.Source: Figure 6 · p. 15, c. Comparison with Canadian data · p. 15

Conclusions

• The international dimension can be endogenized in the Triple Helix model as a fourth dimension, extending the dimensionality of Shannon-type entropies.Source: Conclusions · p. 17
• Internationalization among Japanese academic scholars in 1992–1995 is linked to opening the national system to the international dimension.Japan is described as atypically weakly internationalized for an OECD member.Source: Conclusions · p. 17
• The international dimension is identified as a source of synergy during the period under study (1980–2004).Source: Conclusions · p. 17
• The authors expect similar synergies to have emerged since 1990 in other Asian countries and Eastern European accession countries.Source: Conclusions · p. 17
• Mutual-information measures indicate transitions in a publication system through its relevant subdynamics, using transmissions between and among the helices as the operationalization.Source: Conclusions · p. 17