Who funds the production of globally visible research in the Global South?

There exist large disparities globally when it comes to knowledge production with very small proportions of publications from the majority world penetrating the global science system. This paper examines the Scopus data on the funders of publications (co-)authored by individuals based in the Caucasus or Central Asia. The analysis shows that in the conditions of scarce local funding for research, research activity has been fuelled by international funding. The funding has been rather diverse and originated from 98 countries across North America, Europe, Asia, Latin America, the Middle East, and Australia. The two countries most frequently mentioned in the funding acknowledgements are the United States and the Russian Federation, cumulatively accounting for approximately 20% of funding acknowledgements. Different types of organisations have funded globally visible research produced by authors based in the Caucasus and Central Asia, most notably bilateral agencies (68% of all funding acknowledgements), followed by philanthropies (7%), and multilateral organisations (5%). This paper offers evidence to question the viability of the narrative of North–South divisions in the global science system. While higher education and research are shaped within territorially bounded, self-contained, and discrete spaces of nation-states, the global science system rests on the extensive flows of funding and knowledge which extend well beyond the nation state.


Introduction
There exist large disparities globally when it comes to knowledge production. Many of the rich and powerful countries in the Global North have strong capacity to undertake research and generate new knowledge. Research capacity is considerably more limited in the majority world where multiple factors of disadvantage are at play: the institutions and researchers lack autonomy, the funding available for research is scarce, and the working conditions are not attractive. These global disparities in research capacity are exacerbated by the global science system and the dominance of the English language. The publications indexed in Scopus/WoS and published normally in English are considered to be part of global science. The global science system requires that any new knowledge goes through peer review before being published. Peer review can be a daunting process for non-native English speakers who might also lack insider knowledge of how academic journals operate. The status quo, therefore, is that very small proportions of publications from the majority world penetrate the global science system (Chankseliani et al., 2021a(Chankseliani et al., , 2021bDemeter, 2020;Kasozi, 2017;Liverpool, 2021). The common knowledge described in this paragraph misses two crucial pieces of the puzzle.
First, the global science system is a relational system which cannot be understood by following the logic of North-South divisions. Higher education and research are shaped within territorially bounded, self-contained, and discrete spaces of nation-states (Clark, 1983;Enders, 2004;Huisman et al., 2007;Kosmützky, 2015). At the same time, higher education and research are developed through collaborations and funding flows between countries. The calls to study relations, processes, and agencies beyond nation-states (Marginson & Rhoades, 2002;Watson, 2009) date back to the 2000s. As argued by (Collyer et al., 2019), it is 'a serious conceptual mistake to equate global marginality with passivity' as there are intellectuals working in the Global South who engage in research partnerships beyond the borders of their countries and, in some instances, through their research highlight the limitations of universalist theorising (p. 173). Approximately one third of publications from 2019 indexed in the WoS have involved an international collaboration (Chankseliani et al., 2021a(Chankseliani et al., , 2021b. While the literature on international collaborations is expanding, there is a scarcity of evidence on international research funding flows. As there do not exist readily available statistics on research funded from sources outside the corresponding author's country, it is normally assumed that all research is funded from national taxes. It has been argued that richer countries produce more, better quality, and more impactful research output (Allik, 2013). A recent bibliometric study of research output from post-Soviet countries has challenged this assumption, showing that countries that prioritize research, do have more researchers and more publications. Yet, this is where the links between the characteristics of the output and the research funding end. The study found that none of the three measures of the national spending on research (GERD as % of GDP, GERD per researcher, or GERD per capita) are linked with the quality or impact of the research output (Chankseliani et al., 2021a(Chankseliani et al., , 2021b. Second, some countries which are still considered as part of the Global South, such as China, Russia, Brazil, South Africa, India, Turkey, have been actively engaged in various development partnerships, especially within their regions (Kozma et al., 2018;Marginson & Xu, 2021;Shaw & Kabandula, 2020). They have also been funding research nationally and internationally, contributing to the expansion of collaborative research networks (Huang & Huang, 2018). This paper considers these two pieces of the puzzle and focuses on one region-the Caucasus and Central Asia. More specifically, this paper examines the sources of funding for globally visible research produced in Central Asia (Kazakhstan, Kyrgyzstan, Tajikistan, Uzbekistan) and the Caucasus (Armenia, Azerbaijan, Georgia). This region represents a compelling space for studying the international research funding flows. The Caucasus and Central Asia were a part of the Soviet Union for most of the twentieth century.
A key characteristic of Soviet higher education was its seven-decade-long isolation from global influences. Soviet citizens were not allowed to travel abroad, hence, the opportunities for international exchange were extremely limited. At the same time, the Soviet Union was a diverse empire that brought together individuals with different cultures, languages, religions, and traditions of higher education. Within the borders of the Soviet Union, universities, research institutes, academics and students used to engage in exchanges and collaborations. The dissolution of the Soviet Union led to the opening of borders in most of these countries which led to the development of global links, enabling the flows of knowledge and finances. Out of every 10 globally visible publications from post-Soviet countries, seven to eight are internationally co-authored (Chankseliani et al., 2021a(Chankseliani et al., , 2021b. The present paper examines the following research questions in the context of the Caucasus and Central Asia: to what extent do international funding flows support the production of globally visible research? Which countries are the most prominent international funders of research? What are the contributions of different types of international funders-bilateral, multilateral, and philanthropic-to supporting globally visible research?

Research funding
The existing small body of scholarship on research funding can be situated at the intersection of two bodies of literature-a corpus of scholarly writing on research on research and the scholarship on international development that this paper refers to as world development. 1 The studies on research funding within the research on research literature investigate funding allocation/peer review; research assessment; funding priorities and agenda setting; changing landscape of research funding; policy, practical or academic relevance/ effectiveness of funding; productivity and funding; research collaborations. Such literature often focuses either on the funding for specific field of studies (Grant & Buxton, 2018;Tahmooresnejad et al., 2015) and/or a specific national context (Haake & Silander, 2021;Meroka & Ojwang, 2018). There are also studies which take a broader multi-country/ worldwide view (Curry et al., 2020;Thomas et al., 2020;Whitley et al., 2018). Mostly bibliometric studies have examined the acknowledgements of published research. Overall, it has been established that academic publications normally acknowledge six main types of support: moral, access (to samples, materials, facilities, etc.), clerical, technical, financial, and peer interactive. A number of studies examine the characteristics of research funding using the data on acknowledgements (Costas & van Leeuwen, 2012;Crawford & Biderman, 1970;Paul-Hus et al., 2016. A small body of scholarship on research funding within world development literature investigates the sources and impact of funding for research and for higher education, as higher education institutions are producers of research. Traditional lead funding from bilateral (e.g. USA, European countries) and multilateral funders (e.g. the EU, World Bank, IMF) has been supplemented by emerging bilateral donors (e.g. China, Russia, Turkey, Korea), philanthropic organisation, private firms, university associations, international research organisations, and diasporas (Bassett & Maldonado-Maldonado, 2009;Heyneman, 2006;Heyneman & Lee, 2016;Laakso & Hautaniemi, 2014;Lopes & Kararach, 2019;Niemann & Martens, 2021;Nugent, 2016;Sadlak & Hüfner, 2002;Saner & Yiu, 2019;Shahjahan, 2016;Shaw & Kabandula, 2020). Out of these, bilateral, multilateral and philanthropic sources appear to be the most prominent. Bilateral funding is normally defined as assistance from a foreign government agency from a high-income country. Multilateral funders are multilateral institutions which involve three or more governments. Philanthropic funding comes from an organisation whose primary objective is philanthropy.
While the precise impact of research funding in post-Soviet countries has been difficult to establish, it has been argued that the funding from specific multilateral sources such as the OECD and World Bank, philanthropies like the Open Society Foundation or bilateral agencies like the USAID have shaped social and education policies (Dakowska, 2017;Gounko & Smale, 2007). Kubik (2013) also underscores the role of international agencies, including multilateral, bilateral and private actors, in supporting political and economic transformations in post-Soviet countries. The role of donor funding for higher education and research has also been recognised in other parts of the world. In African countries, international funding has impacted the number of trained scholars, the breadth and depth of academic research, and access to libraries and laboratories; it has also enabled international exchanges of scholars and the development of research collaborations (Hydén, 2017). African researchers' publications with a funding acknowledgement, either national or international, have higher average impact and are normally published in journals with higher impact than African researchers' publications without a funding acknowledgement (Kozma et al., 2018). This finding pertains to all collaborative and non-collaborative publications with an author based in Africa. African researchers appear to rely on some local/ regional sources (most prominently, South Africa's National Research Foundation) as well as a variety of international sources, such as the European Union, Wellcome Trust, Bill & Melinda Gates Foundation, and National Natural Science Foundation of China (Kozma et al., 2018). Various funders have their own rationales for funding higher education and research internationally. Mangold (2020) argues that US-based, market-oriented, liberal philanthropic foundations are likely to support other aims through funding higher education and research; in contrast, more state-oriented, corporatist-conservative German foundations might fund higher education for its own sake (Mangold, 2020). Rationales for philanthropic giving in higher education can be explained by biographical, psychological, and cultural factors, including religion (Cascione, 2003).
Furthermore, bilateral aid sometimes tries to promote the commercial and political interests of bilateral donors, serving the purposes of expanding the market and supporting foreign policy which can reinforce the existing power structures and disparities between donors and recipients (Chankseliani, 2021;Ishengoma, 2017).
It is also evidenced that the bulk of international aid for post-secondary education is spent on scholarships to study in donor countries. Such scholarships often fund (post-) graduate degrees with an important research component. In 2015, scholarships constituted 70% of all aid for post-secondary education (UNESCO, 2015). In 2018, the share dropped to around 50% (UNESCO, 2020). Scholarships offer important opportunities to individuals to study abroad and to later contribute to their home countries (Campbell et al., 2020;Chankseliani, 2018;Jamison & Madden, 2021).
While the literature recognises the importance of international funding flows, there is also a clear indication that international funding flows may have limited potential to enhance the research capacities in the Global South due to power asymmetries and imbalanced relationships between collaborators (Ishengoma, 2017). International collaborations do not always support 'homegrown knowledge'; instead they promote the Northern knowledge, the 'cultural and ideological content generated by the major powers' (Kasozi, 2017, p. 88). North-South collaborations can also be detrimental to the development of research capacity of individuals and institutions by at times undermining 'well-intentioned and carefully prepared schemes for building capacity and professionalism that would have enabled domestic institutions to take responsibility for their own development' (Hydén, 2017, p. 22).

Data and methods
The study uses bibliometric analysis of secondary numeric data on the funding sources of published research. The data on the funders of published research was made available from Elsevier's International Center for the Study of Research (ICSR) which is a cloudbased computational platform enabling users to extract and transform the Scopus data. The funder information was extracted from the acknowledgements section as written in the final publication PDF. The dataset encompassed all publication records from 1990 to 2019 inclusive. The ICSR lab filtered for publications that had an author affiliated with an institution in Armenia, Azerbaijan, Georgia, Kazakhstan, Kyrgyzstan, Tajikistan, or Uzbekistan. The dataset included the count of publications by funding acknowledgement (yes/no), funding agency, funding agency's country, publication author country of affiliation, and the year of publication. One caveat here is that some publications were counted more than once in the dataset. These were publications either with authors from multiple countries in this region or with multiple funders.
There was a considerable variation in the proportions of publications with funding acknowledgement from the overall publications linked with authors affiliated with an institution in each country. From all publications with Armenian, Georgian or Kyrgyzstani authors, approximately one-third were linked with a funding acknowledgement (Fig. 1). Of the publications with an author based in Tajikistan, Kazakhstan, or Uzbekistan, approximately one-fifth were linked with a funding acknowledgement. Finally, only 16% of authors based in Azerbaijan who published in a Scopus referenced journal, acknowledged research funding. In terms of actual numbers of publications with funding acknowledgement (Table 1), the publications with authors from Armenia and Georgia were overrepresented. These were followed by publications with authors affiliated with an Azerbaijan-based institution. A considerably smaller number of publications with authors from Central Asia contained funding acknowledgements.
The data cleaning, coding, and analysis was completed by the author who benefited from email communication with the ICSR team for clarifications. In the process of data cleaning, some data points missing the country of the funding agency were removed. These were the funder names such as 'Ministry of Education and Science', or 'National Science Foundation', or 'National Research Foundation' without an indication of the country of the funder. The final dataset included the data on 175,045 counts of funding acknowledgement. The absolute majority (97%) of the counts of funding acknowledgement were from the publications since 2010. The details of data coding are included in the results section, where relevant.
Acknowledgements of published studies have been analysed as paratextual material since the 1960s with Crawford and Biderman (1970) being one of the pioneers; they examined the sponsors of research published in American sociology journals. Machine reading of acknowledgements data has become possible only in the early 21st century when the Web of Science and Scopus started indexing funding acknowledgements. Researchers who have analysed acknowledgement data refer to various limitations of using such data. These limitations mostly relate to the typology of acknowledgements, algorithms for searching these acknowledgements, simple errors and confusion (such as misspelling funder names), cultural and political norms of acknowledging research funding, the tendency to exaggerate the funding from specific sources, and the so-called 'acknowledgement amnesia', i.e. the fact that some support sometimes goes unacknowledged (Costas & van Leeuwen, 2012;Kozma et al., 2018;Paul-Hus et al., 2016;Rigby, 2013). Besides these limitations which apply to any study using the acknowledgement data, there were a number of limitations that the ICSR team highlighted when the data was shared. The first major limitation was the over-representation of US/UK funding and English-language journal articles, and under-representation of non-English language output. When it comes to the data on funders, the ICSR team explained that Scopus has a bias towards US/UK funders as agencies from these countries were the major agencies with whom the data was initially developed. There was also a bias towards acknowledgements in the English language due to the natural language processing algorithms used to extract this information. Another limitation of the ICSR dataset was its limited coverage of the 1990s output. Scopus began in the early 2000s and gradually backfilled earlier publications. Additionally, that was the period when the Soviet Union dissolved and the countries in the Caucasus and Central Asia changed their names. When the names of country affiliations for institutions change, as happened with all institutions in this region in the early 1990s, the labels in the Scopus database can be a mixture of the old and new country names, depending on whether they were cleaned or not. Taking Yerevan University in Armenia as an example, the ICSR team found that looking at the early 1990s some of the publication authors were labelled as from Armenia, USSR, or Russia. Thus, the label of Armenia as author's country of affiliation, might have excluded a number of publications of Armenia-affiliated authors, published in the early 1990s but erroneously labelled as USSR or Russia.

To what extent do international funding flows support the production of globally visible research?
In this study, the national sources of funding encompassed national governments or higher education institutions from a country of publication's author. For example, if a publication had an author from Azerbaijan, a funding source acknowledgement from Azerbaijan was coded as a national source. The database did not include acknowledgements of funding from philanthropies, private firms, non-profits, museums, or hospitals/clinics which were based in the author's country of affiliation. 2 Thus, within each of the countries analysed, the local funders of globally visible research were governments or higher education institutions.
International funders were all those from outside the author's country of affiliation. For example, if a publication had an author from Kyrgyzstan, an acknowledgement of funding from Kazakhstan or Germany was coded as international funding.
The analysis revealed that the absolute majority of funding acknowledgements in the database refer to international sources of funding. Only a very small proportion of funding acknowledgements included national sources of funding. Uzbekistan-based sources were mentioned in 5% of funding acknowledgements of publications with an Uzbekistan-based author. Georgian, Armenian, and Azerbaijani sources were mentioned only in 1% to 2% of funding acknowledgements of publications from these countries. Only one publication with an author from Tajikistan received funding from a local source. The database did not include any articles of Kyrgyzstan-based authors acknowledging a local source of funding. Thus, in the conditions of scarce local funding for research, externally funded research activity has been fuelled by international rather than national funding in the Caucasus and Central Asia. Kazakhstan was the outlier with 29% of funding acknowledgements of publications with a Kazakhstan-based author indicating that research funding had been received from a Kazakhstani source.
The sources of international funding included bilateral donors (e.g. Germany, Japanese government, US, Russian, Chinese, Norwegian, Slovakian, Dutch governments), multilateral agencies (e.g. European Commission, UNDP, Islamic Development Bank, International Organization of Migration, NATO, World Health Organisation), philanthropic organisations (e.g. F.Ebert Foundation, Aga Khan Foundation), and international non-profits (e.g. American Councils). The third section of the results examines the contributions of different types of international funders-bilateral, multilateral, and philanthropic-to the production of globally visible research output in the Caucasus and Central Asia. Before offering some insights into this, the next section looks at the countries where the research funding originates.

Which countries are the most prominent international funders of research?
The USA emerged as the most frequently mentioned funder of research across the Caucasus and Central Asia ( Table 2). The American government funding was mentioned in almost 28% of all funding acknowledgements in 2000. It dropped to 14% of all funding acknowledgements in 2010 and 8% in 2019. In total, across 1990-2019, 14% of funding acknowledgements mentioned an American source of funding. The funding acknowledgements of publications with a Kyrgyzstan-affiliated author mentioned an American source most frequently (22% of all funding acknowledgements with a Kyrgyzstan-based author). The funding acknowledgements of publications with Azerbaijan-affiliated authors mentioned a US source of funding the least frequently (11% of all funding acknowledgements with an Azerbaijani-based author).
The most frequently acknowledged organisation from the United States was the National Science Foundation, followed by the US Department of Energy, and the National Research Council. Various other government agencies, notably the US Agency for International Development, have supported higher education partnerships with local HEIs in the region to develop research capacity, train staff, and improve facilities. One of such initiatives was the establishment of the Innovative Solutions and Technologies Centre (ISTC) at Yerevan State University in 2016. Another recent example is the STEM Higher Education Project in Georgia where the US Millennium Challenge Corporation worked with the Georgian government to establish international university partnerships for modernising STEM education in the country.
Russia emerged as the second most frequently mentioned funder of globally visible research. Russian government funding was mentioned in about 22% of all funding acknowledgements in 2000. It dropped to 10% in 2010 and 8% in 2019. In total, 6% of all funding acknowledgements across all years mentioned a Russia-based source. There was a large variation by the country of author affiliation. Russian funding was mentioned in 24% of acknowledgements from publications with a Tajikistan-based author, in 12% with a Kazakhstan-based author, and in 10% with a Kyrgyzstan-based author. In contrast, only 5% to 6% of funded acknowledgements from publications from Georgia, Armenia, and Azerbaijan received funding from Russia-based sources. Tajikistan is the only country in the region which has a larger share of funding acknowledgements of sources from Russia than from the United States.
The most frequently acknowledged organisation from the Russian Federation was the Joint Institute for Nuclear Research. This was followed by the Russian Foundation for Basic Research, the Ministry of Health of the Russian Federation, the National Research Center Kurchatov Institute, and the State Atomic Energy Corporation ROSATOM. All exclusively state organisations.
Besides the USA and Russia, there were 96 other countries which funded research produced in the Caucasus and Central Asia. Table 2 presents a list of the key countries globally-including post-Socialist countries-where international funding originated; these account for approximately three-quarters of all publications with a funding acknowledgement. The USA and Russia, the most frequent funders of research in the region, were followed by Germany and the UK (Table 2). Agencies from each of these countries accounted for approximately 6% of analysed acknowledgements from the research output produced by academics in the Caucasus and Central Asia. Table 2 also includes multilateral funders which accounted for approximately 5% of funding acknowledgements, with Kyrgyzstanbased authors benefitting from such funding the most (8.6%) and Kazakhstan-based authors the least (4.9%).
Agencies from Brazil, Canada, Switzerland, and Spain were mentioned in approximately 3% to 4% of funding acknowledgements each. These were followed by China. China-based sources of funding accounted for 2% of funding acknowledgements from papers with an author based in one of the three Caucasus nations and 7% from papers with an author based in Central Asia. Unlike the Russia and USA-based sources which accounted for a decreasing share of mentions in research output, the proportion of China-based funding acknowledgements increased from 2% of all funding acknowledgements in 2000 to 4% in 2019.
From post-Socialist countries, Poland was the second most frequent funder of research in the region, mentioned in 2% of funding acknowledgements. Poland was followed by Kazakhstan and Czech Republic, with each mentioned in approximately 1% of funding acknowledgements.
An important takeaway from Table 2 is that the sources of research funding have been extremely diverse. Agencies from North America, Europe, Asia, Latin America, as well as the Middle East and Australia provided research funding to globally visible research (co-)authored by individuals in the Caucasus and Central Asia. The cross-sectional analysis of the data for 2000, 2010, and 2019 shows that the funding sources had been diversifying. While the funding acknowledgements included names of sources from 25 countries in 2000, the publications from 2019 mentioned sources from 78 countries.

What are the contributions of different types of international funders-bilateral, multilateral, and philanthropic-to supporting globally visible research?
Different types of organisation fund research in the Global South: bilateral, multilateral, philanthropic/charity, non-profits, private firms, university associations, international research organisations, and multi-actor engagements. This literature-based typology was expanded when coding the Elsevier/Scopus data to include higher education institutions, government affiliated non-profits, museums, trade unions/professional associations, and hospitals/clinics. The analysis of the Elsevier/Scopus data confirmed that three types of organisation had been the most prominent in supporting research activity in the Caucasus and Central Asia: bilateral, multilateral, and philanthropic.
The most frequently mentioned funding sources were foreign government agencies (Table 3); they were mentioned in 68% of the funding acknowledgements analysed. The share of foreign government funding ranged from 45% in Kazakhstan to 77% in Azerbaijan.
Philanthropies/charities came up as the second most frequently mentioned types of organisation who funded research; these accounted for 7% of funding acknowledgements. Azerbaijan-, Kyrgyzstan-, Georgia-and Armenia-based authors mentioned the philanthropic funding in a larger share (approx. 7-8%) of acknowledgements when compared  to Tajikistan and Uzbekistan-based authors (5%). Some of the most frequently mentioned philanthropies/charities were from North America/Europe. These included: Canada Foundation for Innovation, Leverhulme Trust, Marianne and Marcus Wallenberg Foundation, Hispanics in Philanthropy, A.G. Leventis Foundation, Fondation Partager le Savoir, Welch Foundation, British Skin Foundation, Alfred P. Sloan Foundation, among others. The names of multilateral organisations were encountered in one in twenty funding acknowledgements, with the largest share (9%) of multilateral funding acknowledgements in the publications with Kyrgyzstan-based authors and the smallest share in the publications with Kazakhstan-based authors (5%). The following three multilateral organisations accounted for the bulk of the multilateral funding for research: the European Research Council, the European Regional Development Fund, and the European Commission. Some of the most frequently mentioned programmes were the Seventh Framework Programme, the Horizon 2020, and the Marie Skłodowska-Curie Actions. The most frequent multilateral funders from outside the European supranational agencies included International Atomic Energy Agency, North Atlantic Treaty Organisation, World Health Organisation, and the World Bank Group.
This study coded separately the acknowledgements of funding from higher education institutions located abroad. Such funding was mentioned in approximately 6% of all funding acknowledgements. Looking closely at the cases where a funder was a higher education institution in another post-Soviet country, it can be hypothesised that the research funding originated from the public budgetary allocations in the country where the institution was located. Examples of such cases included acknowledgements of funding from state universities based in Russia, such as Kazan Federal University, Saint Petersburg State University, Tomsk Polytechnic University, or Tomsk State University.
Private for-profit firms and international research organisations were each mentioned in approximately 2% of funding acknowledgements. The remaining types of funding agency-professional associations/trade unions; non-profit, government-affiliated organisations; and non-profit research organisations-each appeared in less than 1% of all funding acknowledgements (Table 3).
When it comes to the two main funding sources-the US and the Russia-based organisations-important differences emerge. There were 1062 various US-based sources of funding acknowledged in publication but only 42 Russia-based organisations. The USbased organisations spanned across all types while the Russia-based organisations were almost exclusively either government organisations or higher education institutions.

Discussion
The countries in the Caucasus and Central Asia chose divergent paths of economic, social, and political development (Chankseliani, 2022b). Irrespective of paths chosen, higher education and research policy discourses across most countries in the region emphasise global norms, global reputation, and global competitiveness (Chankseliani et al., 2021a(Chankseliani et al., , 2021bChankseliani & Silova, 2018). Global is becoming a condition in which individuals, institutions, and countries aspire to act. Yet, the investments in R&D are not aligned with these aspirations. The Gross Domestic Expenditure on R&D (GERD) ranges between 0.10 and 0.13% of GDP in Tajikistan, Kyrgyzstan, Kazakhstan, and Uzbekistan. Azerbaijan, Armenia, and Georgia spend between 0.18 and 0.30% of their GDP on R&D (Chankseliani et al., 2021a(Chankseliani et al., , 2021b. Prior research has also shown that former Soviet countries are not key contributors to global science, as they cumulatively produce about three percent of the global output included in the WoS. This is comparable to the cumulative research output from all countries on the African continent. The countries in Latin America have a slightly larger share (5%) of publications in the global domain (Chankseliani et al., 2021a(Chankseliani et al., , 2021b. Hence, former Soviet countries can be situated in what literature refers to as 'Global South.' This paper has analysed funding acknowledgements of publications included in WoS. This analysis has shown that a very small proportion of globally visible research (co-) authored by academics based in the Caucasus and Central Asia was funded from their respective local sources. Between 1 and 5% of funding acknowledgements from publications with authors based in Georgia, Armenia, Azerbaijan, or Uzbekistan mention local sources. Kazakhstan is the outlier with 29% of funding acknowledgements from publications with a Kazakhstan-based author indicating that research funding was received from a Kazakhstani source. There are no acknowledgements of such sources in funded publications with authors from Kyrgyzstan and Tajikistan. 3 Kazakhstan being the outlier here can be explained by a relatively high public investment in higher education and research in Kazakhstan in the last decade or so, when compared to other countries in the region. Kazakhstan has implemented a variety of policies which reconfigured the higher education and research organisational and funding landscapes Lovakov et al., 2022). These policies contributed to the sharpest rise in the production of globally visible research by Kazakhstani-based researchers since 2010, when compared to other post-Soviet countries (Chankseliani et al., 2021a(Chankseliani et al., , 2021b. The absolute majority of publications with a research funding acknowledgement had received funding from international sources. One caveat to consider here is that what authors acknowledge as funding sources, in these countries and elsewhere, is normally the funding allocated for a specific research project that the research output stems from. However, research project funding is one form of the national funding allocated to research. There are also annual public budgetary allocations, which are not designated for a specific research project. Such public allocations are not normally mentioned in funding acknowledgements of research output. The majority of globally visible research in these countries is produced either at universities or research institutes (Lovakov et al., 2022). Like in many other parts of the world, both types of institution receive public funding that sustains their educational activities and some limited research activities. Thus, individuals employed at these institutions who are not in receipt of research grant income, might still engage in research and publish, while being paid from the public purse. In the Caucasus and Central Asia, the funding for globally visible research is geographically diverse and ranges from North America to Russia, from Europe to Asia, from Australia to Latin America. The regional funding flows from other post-Socialist countries have also been notable. The research funding from the philanthropic sector and from multilateral organisations has been overshadowed by the funding received from foreign government agencies. 68% of analysed acknowledgements indicated that the funding was received from a foreign government agency. The acknowledgements of philanthropies/charities (7%) and multilateral organisations (5%) were significantly less frequent in the publications analysed.
The funding from the USA has been most frequently mentioned in the globally visible research output from this region. The USA is followed by Russia. The USA's dominance as the key source of funding for global research emerging from these countries can be explained by the USA's leading role in providing development assistance in this region. Moreover, the USA's global public diplomacy spending in the region has been relatively high. Four of the countries (Uzbekistan, Kyrgyzstan, Georgia and Kazakhstan) are among the top 22 highest recipients of the global public diplomacy spending of the US government; the other three (Armenia, Tajikistan, and Azerbaijan) are among the top 63 on the list of 179 countries (U.S. Advisory Commission on Public Diplomacy, 2020). The global public diplomacy spending of the US government covers various international exchange programmes, some of which explicitly support research. Most notable of these are selected programmes under the Fulbright umbrella. This spending also covers the funding for twenty-two American Overseas Research Centers globally, including American Research Institute of the South Caucasus focusing on the funding of research on and about Armenia, Azerbaijan, and Georgia.
The former imperial and colonial power-the Russian Federation-uses a variety of partnership, diplomatic and warfare tools to retain and strengthen its political power in the region (Cameron & Orenstein, 2012;Chankseliani, 2021;Krickovic, 2014;Saari, 2014), and funding research appears to be one of those tools.
Despite the prominence of American and Russian funding in the acknowledgements, one take away from the analysis of the sources of funding is that the sources of support are far from bipolar. The American and Russian funding sources cumulatively account for approximately 20% of funding acknowledgements. Researchers in the Caucasus and Central Asia seem to have benefitted from financial support from 98 counties, including funding from Brazil, Spain, and Australia which are not key international development funders for the region. Whitley et al. (2018) highlight the potential implications of the diversification of funding sources in the European context which might also apply to this region. They discuss how funders not encouraging the diversity of research objectives or approaches might lead to the homogenisation of research. The narrowing of objectives or approaches might be particularly problematic, they note, in the contexts where public research funding is scarce. This can be an important topic of future investigation in the Caucasus and Central Asia.
The flows of funding are normally accompanied by the flows of knowledge as international funding flows are often part of international collaborations. Knowledge creation thrives on research collaborations (Mouton et al., 2018). As shown elsewhere, the majority of globally visible publications emerging from this region are produced through interactions across borders (Chankseliani et al., 2021a(Chankseliani et al., , 2021b. Kyrgyzstan and Georgia have about seven in ten publications co-authored internationally. These are followed by Kazakhstan, Tajikistan, and Armenia, with about six in ten publications with international coauthors. In Azerbaijan and Uzbekistan, every second publication in the global domain is internationally co-authored (Chankseliani et al., 2021a(Chankseliani et al., , 2021b. International collaborations are sometimes unjust or unsustainable (Moyi Okwaro & Geissler, 2015). At the same time, international collaboration and interdependence, understood broadly, can be drivers of domestic change (Gilardi, 2012). International funding flows and international collaborations can have considerable impact on research capacity in the Caucasus and Central Asia. Research capacity is all about the presence of adequate expertise, motivation, and opportunity (McIntyre & McIntyre, 2003;Murray et al., 2009). In the context of scarce national funding and a lack of critical mass of individuals in various fields of research locally , international collaborations and international funding availability are likely to positively impact the motivation to do research, the development of expertise, and the expansion of opportunities to conduct, write up, and publish research. International funding and international collaborations can also contribute to the inclusion of local researchers into global research communities.
Thus, the production of globally visible knowledge in the Caucasus and Central Asia has not been an endogenous process contained within the territorial boundaries of nationstates. In the context of research being extremely poorly funded nationally, research funding from various international sources has made possible and shaped research productivity in the Caucasus and Central Asia. This study offered evidence to question the viability of the narrative of North-South divisions in the global science system and in the context where research production is relationally shaped by funding flows which extend well beyond the nation state.