The concept of knowledge bases (KBs) was first put forward by Asheim and his colleagues (2005; 2007; 2011) and emphasizes three kinds of knowledge creation, analytical, synthetic, and symbolic KBs, that contribute to a new knowledge distinction. In comparison to old distinctions, such as tacit versus codified knowledge (Gertler, 2003), which are only weakly linked to knowledge generation dynamics, KBs are particularly related to these three kinds of innovation processes, providing a better understanding of the nature of knowledge sharing and innovative activities (Boschma, 2018). Among these, analytical KBs, characterized by formal models, usually characterize innovative science-based projects (Asheim et al., 2011; Davids and Frenken, 2018), as for instance with biomedicine (Ye and Zeng, 2018). In contrast, synthetic knowledge creation is mainly based on experiential learning, such as trial and error. This kind of knowledge is easier to observe in construction and traditional automobile industries (Asheim et al., 2011; Davids and Frenken 2018). The generation of symbolic knowledge is strongly associated with cultural codes or aesthetic elements within cultural industries (Asheim 2007; Klement and Strambach, 2019). Based on this new approach, an extensive body of literature has connected KBs to different fields such as innovation systems, regional innovation policy, institutions, the geography of knowledge, and path dependence (Asheim et al., 2011; Plum and Hassink, 2011; Van Tuijl and Walma van der Molen, 2016; Benneworth et al., 2019; Chen and Hassink, 2020). For example, Asheim and Coenen (2005, p.1180, p.1184) stated that in “a territorially embedded regional innovation system”, the innovation process is mainly based on synthetic knowledge and local buzz between firms, where technological transfer is easier to observe. This stands in contrast to “a regionalized national innovation system”, where analytical knowledge dominates and innovative activities are highly dependent on research institutes and universities. Furthermore, this differs from “a networked regional innovation system”, in which cutting-edge technologies associated with synthetic and analytical knowledge are developed.

Although one kind of KB can reflect the key knowledge generation in some industries, existing studies have indicated that strictly distinguishing KBs from each other in a single industry is unrealistic (Asheim et al., 2017; Manniche et al., 2017). For example, in the eco-building sector, analytical, synthetic, and symbolic KBs can be found (Strambach, 2017). Additionally, in the process of industrial transformation or industrial upgrading, key knowledge creation in the industries may change over time. On this basis, the combinatorial knowledge base (CKB) approach has been suggested and can be viewed as a more advanced alternative.

Combinatorial knowledge bases (CKBs) result from the combination of intra-KBs or inter-KBs, contributing to industrial transformation or cluster transformation (Manniche, 2012; Asheim et al., 2017; Manniche et al., 2017; Plechero and Grillitsch, 2022). Note that the diversified types of KBs in the region suggest a high probability for the emergence of CKBs within the industries or clusters. In comparison to KBs, four features of CKBs can be found. The first concerns the actors carrying knowledge. One kind of KB only refers to one kind of actor, from a scientific or engineering field, or from the field of arts and culture, whereas CKBs may be highly dependent on more agents from different domains. Secondly, a CKB can be understood as a dynamic process, whereas a KB is more static in character. In other words, the evolution of KBs results in CKBs (Boschma, 2018). The third feature concerns the geography. Geographical distance has a bigger effect on synthetic and symbolic KBs than analytical KBs (Boschma, 2018), whereas the space of CKBs varies. Various types of CKBs show different geographical proximity. The fourth feature involves the distinction between codified and tacit attributes. Similar to geographical characteristics, synthetic and symbolic KBs are primarily related to tacit knowledge, and analytical KBs are mostly associated with codified characteristics (Asheim et al., 2011). The diversity of CKBs can lead to some uncertainties in this aspect. In general, the units of observation differ between KBs and CKBs. Research in this area is dominated by two strands (Grillitsch et al., 2017; Marques, 2019; Bennat and Sternberg, 2020). On the one hand, scholars tend to be more concerned with which kind of KB combination is most beneficial to the innovation performance of firms or regions, and with how to facilitate such a combination. On the other hand, transition researchers have long been analyzing combinatorial knowledge dynamics in the process of industrial or cluster transformation (Martin and Trippl, 2015; Zukauskaite and Moodysson, 2016).

Although previous studies are valuable, existing studies on CKBs are mainly used by Western academics with a focus on solving Western problems. There are few studies in the context of China’s economic geography. Nonetheless, there is a need to utilize CKBs in Chinese academic studies for three reasons. The first concerns China’s industrial upgrading and technological revolution. In the current economy, transforming traditional industries based on the combination of new technologies and existing knowledge is a top priority. Disruptive changes and major technological breakthroughs are highly dependent on integrated knowledge creation. As such, they should be explored from the perspective of combinatorial knowledge dynamics. Second, innovation network analysis based on academic papers and patents is challenging. Economic geographers have long been interested in the space and scale of innovation, that is, the typologies of innovation networks, innovation efficiency, and sound innovation environments created by multiple actors based on academic papers and patents. However, previous studies have not comprehensively defined innovation patterns or innovation modes. The third reason concerns state strategy planning. Allocating resources to facilitate emerging industries is key to regional economic development. The CKB approach contributes to an in-depth understanding of innovative resource allocation. Based on these analyses, it is necessary and promising to explore CKBs in China’s economic geography. In this paper, four research directions with uniquely Chinese characteristics are introduced, and concrete cases and scenarios are elaborated. Furthermore, we review the latest approaches to knowledge dynamics and elaborate a research agenda for CKBs in China’s economic geography, offering insights into Chinese-specific phenomena and related strategic planning.

The research prospects of CKBs in China’s economic geography are here introduced in terms of four aspects: the co-evolution of the innovation chain and industrial chain, the geography of innovative activities, innovative entrepreneurship and new path development, and innovation system reconfiguration.

In the last 15 years, on the one hand, the CKB approach offers more nuanced insight into innovation performance and regional transformation. While such insight is valuable, CKBs are primarily used in Western economic geography, with the aim of solving Western problems. So far, CKBs are rarely seen in the context of Chinese transformation. In such a context, different phenomena are observed, which together drive the theoretical reconstruction of CKBs (Gong and Hassink, 2020). In particular, four differences are evident in the transformation process between China (rapid growth with the import, absorption, and application of technology) and Western countries (with a long history of industrialization and technology accumulation). The first difference concerns innovation chains and industrial chains. The variability regarding the coordination of two chains is strongly related to the innovation capacity of the country or the region. In regions with high innovation capacity, innovation is the engine of industrial development. Industrial chains around innovation have prevailed in the last few decades, as their layout was carefully planned. In contrast, in regions with insufficient innovation resources, progressive improvements to the innovation performance of firms based on industrial chains are efficient and easily achievable. As such, it is common to construct innovation chains around industrial chains, and to improve financial systems around industrial chains. The second difference involves the space and scale of innovation. In China, the state plays an important role in the flow of innovative elements at the inter-city and international levels (Howell, 2020b; Hu et al., 2021; Fu and Lim, 2022). For example, the density of inter-city innovation networks within Chinas urban agglomerations is higher than the density outside of urban agglomerations, whereas transnational R&D investment projects coincide with national industrial planning (Gui et al., 2019; Cao et al., 2022). In Western countries, where local buzz and global pipelines prevail, the state is much more concerned with the spatial mobility of labor and labor mobility across sectors. The third difference pertains to entrepreneurship and innovation research in the context of transformation. In China, cross-sectoral entrepreneurial discovery has become a trend in the last few decades, in which three types of entrepreneurs are identified. Entrepreneurs who achieve capital accumulation in some traditional industries are important forces that facilitate emerging industries. Entrepreneurs from internet companies who excel at attracting venture capital are also entering emerging industries, such as manufacturing high-end medical devices and electric vehicles. Entrepreneurs from overseas often have rich professional experience and social ties at the international level. A common characteristic of these three types of entrepreneurs is that they have been engaged in emerging industries and recombinant innovative activities over a relatively short period; this is related to China’s industrial history. In contrast, in Western countries, combinatorial knowledge creation relies heavily on CEOs from family businesses that have been rooted in the industry for a long time. An example is the small-medium enterprise specializing in optical engineering in Germany. These entrepreneurs are often important forces that fuse knowledge innovation and lead world development trends. The fourth difference is that Chinese scholars pay more attention to outcomes led by innovation system reconfiguration, such as the spillover effects of regional knowledge facilities, network density, and innovation efficiency. Western geographers, by contrast, focus on the interaction of multiple actors in the system reconfiguration process (Friedrich and Feser, 2021).

On the other hand, Chinese economic geography is facing a huge challenge in terms of measuring recombinant innovative activities. In the past, academic papers and patents were the main research objects. In the context of transformation, cross-domain R&D and recombinant innovative activities have been the new engine for the industrial revolution and regional economic growth (Grillitsch et al., 2018; Mewes, 2019; Nieth and Benneworth, 2020). As a country that needed to catch up technologically, China has prioritized technological breakthroughs and disruptive innovation since 2015. The central government has pointed out that one of the crucial goals in the next few decades is to develop cutting-edge technologies and apply them in manufacturing sectors. Against this background, it is necessary to examine, analyze, and discuss the process of knowledge combination and the innovative characteristics of advanced manufacturing. The concept of CKBs provides a theoretical lens for this topic (Davids, 2018).

To conclude, in comparison to industrialized countries, emerging economies and developing countries have a shorter history in radical innovation and cross-domain R&D. Therefore, it is worthwhile to analyze how advanced and lagging cities in emerging economies and developing countries utilize innovative sources at different spatial scales to develop CKBs. In this paper, four research directions were introduced in China’s economic geography: the co-evolution of the innovation chain and industrial chain, the geography of innovative activities from the perspective of the complexity of knowledge bases, innovative entrepreneurship and new path development, and system reconfiguration. Such a research agenda will contribute to China’s economic geography by embedding CKBs in a transformational context that is unique to China.