关键词: 创新, 创新系统, 发展历程, 研究展望

Abstract:  In the late 1980s, Freeman (1987) and Dosi (1988) and a group of earlier innovation scholars studied the comparative advantages of many countries, which laid a foundation for the research of innovation system. Innovation system has become an important research field, especially related to academics and policy makers increasingly. Some international organizations such as OECD (Organization for Economic Cooperation and Development) have adopted this method in their analysis and conducted some studies based on the innovation system, which have influenced the practice of innovation system construction in many OECD countries and developing countries.
At present, the research field of innovation system has become increasingly large and complex. However,there are large disputes about the distinction, boundary, function and classification method of innovation system among scholars. On the 30th anniversary of the concept of innovation system being proposed, this paper discusses some changes of the innovation system that has been created and applied including: national innovation system (NIS), regional innovation system (RIS), sectoral innovation system (SIS) and technological innovation system (TIS). According to the evolution process of innovation system research, this paper analyzes the theoretical origin of each innovation system method, examines the most cited case studies, analyzes the spatial boundary and analysis unit, discusses the basic components function and classification method, with the aim to clarify the theoretical system of innovation system research, and on this basis, puts forward further research each innovation system research method needs.
The research of national innovation system(NIS) can be used to compare the innovation capacity of each country comprehensively, and can furtherly and deeply understand how innovation occurs systematically, the driving and obstacle factors of innovation, and the influence of innovation on the future, economic status and science and technology of a country in the increasing global competition, rather than just focusing on the figures of past economic performance.
The research of regional innovation system(RIS) argues that the preferred spatial level of regulatory intervention is regional rather than central, expanding the current discussion on national innovation system: innovation system can be found at regional (or even sub-regional) levels as well as national and global levels.The necessity of establishing new spatial boundary of innovation system has been proved, and it is meaningful to discuss regional or local technology system.
The research of sectoral innovation system(SIS) , which complements the concepts of NIS and RIS, believes that the organization (space) of innovation is influenced by factors related to industrial and technological rather than national or regional factors. SIS mainly addresses the issues of what are the main network characteristics of innovators, what factors are responsible for the transformation of industries, and what are the relevant national institutional frameworks.
The research of technology innovation system (TIS) focuses on the construction of dynamic system and industry formation of emerging (clean technology) industries, which is a supplement to the industry innovation system (SIS) method and the embodiment of innovation system research in the field of technology (clean technology).
This paper finds that the theoretical origin of each innovation system method, the most cited case studies, the spatial boundary and analysis unit, the basic function of component and classification method of four the innovation system approaches all have the characteristics of generality and historical continuity, as well as the differences of the specific research objects, which are manifested in the further refinement of the research objects from country, region, industry to technology.
This paper points out the shortcomings of the current innovation system research:
First, RIS and NIS methods have been criticized for providing a rather static concept of innovation and paying too much attention to space. By presetting national or regional boundaries, the NIS and RIS concepts cannot fully capture the activities of organizations, networks, and institutions that develop at the supranational level. As a result, there is a lack of clear understanding of how they affect the geographically rooted dynamics of innovation. In addition, because RIS is difficult to give a clear definition with clear applications, its concepts seem to be a mixture of different sources and are often limited to high-tech and/or manufacturing industries.
Second, SIS scholars have developed elaborate industrial classifications that construct innovation processes based on technological systems and trajectories. This approach allows for the development of rigorous analytical frameworks, but it also draws strong criticism for their technical bias. In particular, SIS studies increasingly underestimate the importance of more distributed institutions, non-corporate actors, and the impact of informal institutions on the innovation process. Moreover, it tends to focus on long-term industry dynamics in existing manufacturing sectors (such as semiconductors, automobiles, pharmaceuticals, telecommunications, machine tools, etc.), with few explanations for the emergence of new industries and technologies.
Third, most of the empirical works in TIS set a priori system boundary at the national level, and limit the analysis to the clean technology industry, which is considered to be a group of coherent industries with similar technical trajectories. Therefore, it has attracted criticisms that the empirical applications pay too much attention to space and ignore the differences of innovation process between different industrial contexts. In the future, TIS research needs to extend the system boundaries at the national level to international and global geographic regions, and apply the analytical framework to other industries. In addition, it is necessary to study the nature of different stages of development of TIS, such as the formation stage, in order to evaluate the relative advantages of different systems.
This paper proposes further research directions of innovation system research: continuing to develop innovative system methods, specific directionality and normative orientation including four abilities: self-reflection capacities, bridging and integration capacities, anticipation capacities and experimentation capacities. Further research on innovation systems should also integrate the latest methods within the framework of innovation system analysis, the latest trend of which is called "Global Innovation System (GIS)". On the other hand, the current empirical research on innovation system is mainly targeted at developed countries such as Europe, America, Japan and OECD countries, while there are few studies on emerging economies represented by China. In the future, it is necessary to strengthen the research on these countries to furtherly prove the applicability of innovation system theory in different contexts, especially to strengthen the empirical research and application in China.

Key words:  innovation, innovation system, development history, research prospect