Introduction

This study aims to fill a void in the bibliometric analysis of research internationalization in the Social Sciences and Humanities (SSH), namely the fact that enlightenment literature has not been studied in this respect. Internationalization in the SSH has mainly been shown through the analysis of the increasingly common multi-national co-authorship of journal articles (Benavent-Pérez et al. 2012; Kyvik 2003; Leydesdorff et al. 2014). Using a novel application of barycenter representations, Verleysen and Engels (2014) illustrate the internationalization of academic book publishing for several SSH disciplines and for three book publication types (monographs, edited books and book chapters) as practiced at Flemish universities. However, no study has undertaken a broader analysis in which both academic and enlightenment book publications have been included. In part, this is due to the limited availability and reliability of bibliographic databases containing book references (Sivertsen and Larsen 2012; Gorraiz et al. 2013). Meanwhile, evidence of the importance of book publications in many SSH disciplines continues to accumulate (Piro et al. 2013; Nederhof et al. 2010). In particular the fact that the share of book publications does not seem to decline, at least not in the Humanities (Adams and Testa 2011; Engels et al. 2012), calls for the study of internationalization using book publication data.

In the present article, an analysis is made of research internationalization in the SSH based on the geographic center of weight, i.e. the barycenter of places of publication of monographs, edited books and book chapters. The 20,403 bibliographic references used here are authored by researchers affiliated with at least one of the five universities in Flanders (Belgium), for which comprehensive publication data since the year 2000 have been collected in the VABB-SHW database (Engels et al. 2012). The analysis of places of publication of books sheds light on fundamental aspects of research internationalization in the SSH. The conscious choice by an author for a publisher (and consequently for a place of publication) determines to a considerable extent whether his or her book publication will be readily available to its intended readership at home and/or abroad. This element of active publisher selection is especially valid for authoring a monograph or an edited book, but likely less so for a book chapter, as a chapter author is dependent on the choice of publisher made by the book editor (Verleysen and Engels 2014). From the perspective of the publisher, books must both hold relevance and be accessible (in terms of contents and form) for a specific target audience if they are to be marketed successfully (Thompson 2005). Hence in a context of growing appeal for academic authors worldwide of prestigious, often multi-national, publishing houses (e.g. Oxford University Press) and of globalization of the academic book publishing trade (Thompson 2005; Giménez-Toledo and Román-Román 2009; Giménez-Toledo et al. 2013), the place of publication of a book is relevant from an information dissemination perspective.

In terms of publication types and subtypes, the output of the SSH is markedly heterogeneous. This is, among other elements, explained by the multiplicity of readerships of SSH publications. The most specialized part of the output in terms of subject-matter is likely consulted mainly by academic colleagues, whereas another segment, the enlightenment literature, is aimed at a wider audience of interested readers (Hicks 2004, 2013). As research on the discipline of History shows, it would be naive to suppose that the sets of ‘academic’ and ‘enlightenment’ SSH literatures and their respective readerships are in every respect disjoint. Enlightenment literature may be used and/or cited by academic researchers, and a non-academic audience may show interest in specialized academic publications (Stieg Dalton and Charnigo 2004; Stieg Dalton 2008; Verleysen and Engels 2012). Nevertheless, the distinction between the two types of literature remains valuable to characterize the variety of SSH publishing and its multi-facetted role in both academia and society at large. As clear disambiguation of the two types of SSH literature in publication data is not straightforward, we here use the nearest proxy in the data from Flanders: the distinction between peer reviewed and non-peer reviewed book publications as decided upon by a panel of academic experts (see “Materials and methods”). Peer review is widely accepted as an imprimatur of scientific or scholarly authenticity (Ziman 1968) and therefore the best available way to distinguish between the academic and the enlightenment literature.

Using a basic indicator, namely the places of publication barycenters of monographs, edited books and book chapters published between 2002 and 2011 by scholars affiliated with Flemish universities, we show in this article that, contrary to the academic peer reviewed literature, the non-peer reviewed enlightenment literature remains domestic to a large extent. This is even more so for the Social Sciences than for the Humanities, where both literatures appear to be rather close to each other.

Materials and methods

This study is based on the places of publication of books as registered in the VABB-SHW database (Flemish Academic Bibliographic Database for the SSH, or VABB-SHW) (see http://www.ecoom.be/en/vabb). The VABB-SHW retrospectively collects bibliographic information from the year 2000 onwards on all SSH publications by authors affiliated with at least one of the five universities in Flanders, Belgium. Since 2010, the VABB-SHW is used to calculate part of each university’s share in the University Research Fund (‘Bijzonder Onderzoeksfonds’ or BOF), some 150 million euro yearly granted by the Flemish government to finance basic research. Both peer reviewed and non-peer reviewed publications are registered in the database, but only the peer reviewed ones are taken into account for the allocation of the University Research Fund. The decision whether or not a publication is considered peer reviewed has been entrusted by the Flemish government to an independent Authoritative Panel (‘Gezaghebbend Panel’ or GP) of academic experts. Book publications are for the most part not assessed individually, but at the publisher level. To this end, in 2010 the GP adopted the so-called ‘Level-2’ list of reputed academic publishers with established peer review procedures previously identified in view of the construction of the Norwegian national publication database CRIStin (Sivertsen 2010). After 2010, the Flemish list of approved publishers was expanded selectively. In most recent years, steps have been undertaken to make individual book publications and book series eligible for inclusion as well (Engels et al. 2012; Verleysen and Engels 2013a; Verleysen et al. 2014).

At the ISSI conference 2013 (Verleysen and Engels 2013b) and in (Verleysen and Engels 2014) we introduced an adaptation of (Rousseau 1989), and first used barycenters for places of publication of books. We applied the method for analysis of internationalization to 4,098 peer reviewed monographs, edited books and book chapters published in the 2002–2011 timeframe. To analyze evolution over time, we compared two subperiods, 2002–2006 and 2007–2011. The present article contrasts this previous result with additional data on 16,305 commercially available non-peer reviewed book publications. Both peer reviewed and non-peer reviewed book publications belong to seven Humanities and five Social Science disciplines. In the VABB-SHW database, publications are classified as belonging to one or more disciplines based on departmental author affiliation. For the 12 disciplines analyzed here, the VABB-SHW contains a sufficient number of both peer reviewed and non-peer reviewed book publication references (minimum 50 publications per discipline per type for each of the two subperiods). For determining barycenters, all places of publication were identified as available in the VABB-SHW. For 8 % of the records, the database contains more than one place of publication. In those cases, the first one mentioned has been retained.

A barycenter is defined as C = (C 1, C 2)

$$ C_{1} = \frac{{\sum\limits_{j = 1}^{n} {{m_{j}}{L_{j,1}}} }}{M};\,C_{2} = \frac{{\sum\limits_{j = 1}^{n} {{m_{j}}{L_{j,2}}} }}{M}\quad{\text{with}}\,M = \sum\limits_{j = 1}^{n} {m_j} $$

where L j  = (L j,1, L j,2) is the location of the jth element in the system (the place of the jth vertex of a regular n-gon), m j is the contribution of the jth element, and m j /M is the relative contribution of the jth element.

Barycenters of places of publication of books are geometrically located within a standardized polygon, where each vertex represents one geographic location. Here we seek to determine the publication share of four (aggregated) geographic locations, which together account for over 99 % of the places of publication of the book publications studied. They are: 1° Flanders, 2° the rest of continental Europe (defined as the EU-27 except the United Kingdom plus its acceding or candidate members Croatia, Iceland, Montenegro, Serbia, the FYR of Macedonia and Turkey; plus Albania, Belarus, Moldavia, Norway and Switzerland), 3° the United States and 4° the United Kingdom. In this case the polygon is a square. The center of this square is the origin (0,0) of an X- and a Y-axis, and would be the barycenter location if the system were completely in balance—that is, if each of the four geographic locations represented an exact same number of publications. Each vertex lies at the same distance of 1 to the origin. Consequently the four vertices each have two geometric coordinates: (1,1), (1,−1), (−1,−1) and (−1,1), respectively. Barycenter coordinates (C 1, C 2) within the polygon are determined by calculating a weighted average of each of the four vertices’ two coordinates according to the number of book publications per geographic location. Consequently, the relative proximity of a barycenter to one geographic location (vertex) is indicative of that location’s share in the total of the book publications. Internationalization is then measured by geometric distance of a publication barycenter from the domestic location Flanders, where the authors of the books and chapters included in this study have a university affiliation. For further elaboration on methodology, we refer to (Rousseau 1989; Jin and Rousseau 2001; Verleysen and Engels 2014).

In the results section, we first present the barycenter locations for both peer reviewed and non-peer reviewed book publications in the seven Humanities (i.e. Art History, History, Law, Linguistics, Literature, Philosophy, Theology) and five Social Science disciplines (i.e. Educational Studies, Economics and Business, Political Science, Psychology, Sociology) for the 2002–2011 period. As in previous Studies (Engels et al. 2012), we include Law as a Humanities discipline, although this differs from the classification in the Frascati manual and the CERCS. Law will therefore be separately commented upon. Following this, we compare the distribution of publication languages for peer reviewed and non-peer reviewed books. By this, we aim to further contribute to distinguishing various readerships of the SSH book literature. Finally, the analysis will focus on internationalization as a process. To this end, we compare the percentual shares for each of the four aggregated geographic locations and the main publication languages for two subperiods, 2002–2006 and 2007–2011. A Chi square (χ 2) test of goodness of fit is performed on the absolute values to determine whether results for 2007–2011 are significantly different from those for 2002–2006. As our analysis is limited to four values, there are three degrees of freedom; χ 2 reaches significance when it exceeds 7.81. Rejection of the null hypothesis (no change) is traditionally at p < 0.05. In order to more accurately assess the significance of the observed change, the use of exact p-values is recommended (Schneider 2013).

Results and discussion

Internationalization patterns and SSH readerships

Peer reviewed and non-peer reviewed book literatures show different patterns as regards the degree of internationalization. Figure 1 illustrates this for the seven Humanities disciplines.

Fig. 1
figure 1

Barycenter locations for peer reviewed (filled triangle) and non-peer reviewed (filled diamond) book publications in the Humanities and seven of its main disciplines

Full size image

Figure 1 clearly shows that the non-peer reviewed Humanities book literature is more domestic in terms of places of publication than its peer reviewed counterpart. This is true for all seven Humanities disciplines: barycenters for non-peer reviewed books are without exception located within the quadrant assigned to the location Flanders. For the more internationally published peer reviewed books, five barycenters are located in that same quadrant—though closer towards the vertex assigned to the UK. Two barycenters (Linguistics PR and Law PR) are situated in the adjacent quadrant assigned to continental Europe. Of the 14 barycenters, the one for ‘Law peer reviewed’ is located closest to the center of the polygon, indicating that the difference between the publication shares of the four locations is smallest in this subset. For peer reviewed book literature, Law is the most internationalized Humanities discipline. In this Law is exceptional and in fact takes more of a Social Science profile, given that its non-peer reviewed books are also the most domestically published. This contrast can be explained by the fact that, regarding subject-matter, non-peer reviewed law publications are mostly concerned with Belgian law, whereas peer reviewed published books and chapters mostly comment on European and other international law and their impact on and interaction with the Belgian legal system. Law is indeed a special case among the SSH disciplines as practiced at Flemish universities, since its journal publication patterns are archetypical of those of the Humanities and its main disciplines (Engels et al. 2012). In fact, the result presented here indicates that Law, in terms of publication patterns, is neither a Humanities nor a Social Sciences discipline. More in general, differences between disciplines regarding barycenter location and type of subject matter are related to a partially different readership of the book literature. Among the seven Humanities disciplines, Literature and Theology show the smallest geometric distance between their two respective barycenters. In those disciplines subject matter and the peer reviewed character of a book determine to a lesser extent whether it will have mainly a national or an international readership. As regards the other disciplines, it is notable that for Art History, History, and to a lesser extent Philosophy and Literature, the more international orientation of the peer reviewed publications essentially results from more UK publications as the barycenters per discipline are right below each other. This differs somewhat from the position of Linguistics and Theology for which we find the barycenters of the peer reviewed literature especially more to the right of the barycenters of the non-peer reviewed literature, i.e. more strongly oriented towards continental Europe.

Compared to the Humanities, the differences in barycenter locations for peer reviewed versus non-peer reviewed book publications are considerably more outspoken for the Social Sciences. Figure 2 represents the places of publication barycenters for the book publications in five Social Science disciplines.

Fig. 2
figure 2

Barycenter locations for peer reviewed (filled triangle) and non-peer reviewed (filled diamond) book publications in the Social Sciences and five of its main disciplines

Full size image

In the Social Sciences, the distance between the two respective barycenters for peer reviewed and non-peer reviewed book publications is on average greater than in the Humanities. As also observed for the Humanities, barycenters for non-peer reviewed book literature are all located in the quadrant assigned to Flanders, but in the Social Sciences those for peer reviewed literature are, with one exception, firmly located in the UK quadrant. Only the barycenter for ‘Educational Studies peer reviewed’ shows a dominant share of continental European places of publication.

All this is an indication that the Social Science book literature serves more clearly separated readerships with distinct information needs than is the case for the Humanities, where the barycenters of the peer reviewed and the non-peer reviewed literature mostly do not lie as far apart. From this perspective, two types of Social Science book literature can be discerned: on the one hand, there is a largely domestically published, non-peer reviewed book literature concerned with issues such as socio-economic policy, education and national politics, and aimed at a broad readership of policymakers, interest groups, ngo’s, common citizens and, sometimes, academic researchers with a national focus alike. This literature is the ‘enlightenment literature’ mentioned in the introduction. On the other hand, the Social Sciences generate a specialized academic literature which is mostly published in the UK and aimed at the community of international academia involved in (comparative) empirical, conceptual, and theoretical research. Undoubtedly to some extent a similar segmentation of book readership exists in the Humanities as well, but the Humanities’ barycenter locations suggest more of a continuum both in terms of places of publication and intended readership.

This interpretation of a partial divide between book readerships in the SSH and its relation to the places of publication is supported by the language distribution for non-peer reviewed versus peer reviewed book publications in the VABB-SHW. Whether or not a book is published in the common local language (for Flanders: Dutch), or in the universal language of international academia, English, evidently makes a difference with regards to accessibility of a literature for its intended readership (Hicks 2004). Although English is read and understood by many people in Flanders outside of academia as well, the use of English as a communication (publication) language between native Dutch speakers in Flemish/Belgian society is uncommon. Inversely, not many people outside of the Low Countries readily read or understand Dutch.

In the Humanities, over two-thirds of non-peer reviewed book publications are published in Dutch. English accounts for 19 % of these publications. French and German, official languages of Belgium mainly spoken in the south and the east of the country, account for 9 and 3 % respectively. In contrast, for the peer reviewed literature English is clearly dominant with 67 % of all publications. At the level of the Humanities disciplines the distribution is mostly similar, although there is some variation. In particular, Law sticks out as the discipline with the highest % non-PR in Dutch and the lowest % of PR in Dutch (Table 1). Thus, if Law were not to be considered as a Humanities discipline, the contrast in Fig. 3 would be less pronounced, yet still very clear.

Fig. 3
figure 3

Language distribution for non-peer reviewed and peer reviewed book publications in the Humanities (H)

Full size image

In comparison with the Humanities, the linguistic gap between non-peer reviewed and peer reviewed book literatures is wider for the Social Sciences (Fig. 4).

Dutch is the predominant publication language (62 %) for the non-peer reviewed publications, but represents only 7 % of the peer reviewed publications as a large majority of these are in English (91 %). This further illustrates that the intended readership of peer reviewed book publications in the Social Sciences is primarily an international one. In the Social Sciences as well, differences in language distribution between individual disciplines are noticeable, though they are smaller than in the Humanities (Table 2).

The overall picture of a largely international, English-language academic peer reviewed book literature, as opposed to a domestic enlightenment one published in Dutch is apparent for both the Humanities and the Social Sciences. In terms of the distribution of publication languages as well, this gap is more outspoken for Social Sciences disciplines than for their Humanities counterparts.

Internationalization as a process

Regarding the evolution of internationalization over time, our results show additional contrast, both between peer reviewed and non-peer reviewed book publications and between the Social Sciences and the Humanities. For the non-peer reviewed book publications Table 3 shows the percentual shares for the four geographic locations in the two subperiods 2002–06 and 2007–11, as well as the results of the test for goodness-of-fit.

Table 1 Shares of publication languages for peer reviewed and non-peer reviewed book publications in Humanities disciplines (2002–2011)
Full size table

Overall, the picture is one of stability. For the Social Sciences there clearly is no meaningful or statistically significant change. For the Humanities, however, the χ 2 result does indicate a different distribution for the 2007–11 subperiod, although the actual percentual changes are never greater than +1.1 % (for the UK). Hence for the Humanities too we mainly observe stability in terms of places of publications of non-peer reviewed book publications. At the level of individual disciplines (details not shown in the table), some variability can be observed. For example, the share of the domestic location Flanders has slightly increased for five disciplines, belonging to the Social Sciences or the Humanities (Law +1.44 %; Philosophy +7.23 %; Political Science +2.37 %; Psychology +4.98 %; Theology +2.9 %), but has decreased for the seven remaining disciplines (Art History: −8.22 %; Economics and Business −2.84 %; Educational Studies −1.1 %; History −4.3 %; Linguistics −6.23 %; Literature −7.04 %; Sociology −5.29 %).

Table 2 Shares of publication languages for peer reviewed and non-peer reviewed book publications in Social Sciences disciplines (2002–2011)
Full size table
Table 3 Shares of Flanders, Continental Europe (Con. EUR), United Kingdom (UK) and United States of America (USA) among non-peer reviewed book publications (2002–2006 and 2007–2011)
Full size table

On the whole the non-peer reviewed book literature appears stable domestic in character. This contrasts with the ongoing internationalization of the peer reviewed book literature. In (Verleysen and Engels 2014) we present the same approach towards analyzing the process of internationalization for the peer reviewed book publications by comparison of data for the subperiods 2002–06 and 2007–11. For the SSH as a whole, we observed an outspoken process of internationalization, with a clear drop in the publication share of Flanders (−13.32 %). In the Humanities the publication center of weight is shifting away from Flanders (−18.3 %) and moving fast towards continental Europe (+10.31 %) and the UK (+6.82 %). For the more internationally oriented Social Sciences, the already elevated share of the UK increases further still (+7.88 %), to the detriment of the other three locations. At the level of individual disciplines, all of the 12 SSH disciplines show a drop in the share of the domestic location Flanders. The change is greatest for Art History (−29.05 %); History (−21.2 %); Literature (−27.7 %), Philosophy (−27.97) and Theology (−24.56 %). It is smallest for Economics and Business (−3.7 %), Educational Studies (−1.52 %), Law (−4.6 %), Linguistics (−4.89 %), Political Science (−4.22 %), Psychology (−1.64 %) and Sociology (−1.03 %) (Verleysen and Engels 2014).

Fig. 4
figure 4

Language distribution for non-peer reviewed and peer reviewed book publications in the Social Sciences (SS)

Full size image

Internationalization measured by places of publication is also reflected in change as regards publication language. In terms of publication languages, the academic peer reviewed book literature evolves towards more use of English in the Humanities. The non-peer reviewed book publications, however, remain mostly in Dutch. Tables 4 and 5 present the shares of publication languages for peer reviewed and non-peer reviewed book publications in the Humanities and the Social Sciences in the periods 2002–2006 and 2007–2011 respectively.

Table 4 Shares of Dutch, English, French and other languages among non-peer reviewed book publications (2002–2006 and 2007–2011)
Full size table
Table 5 Percentual shares of Dutch, English, French and other languages among peer reviewed book publications (2002–2006 and 2007–2011)
Full size table

For the non-peer reviewed enlightenment literature, there is stability in the language distribution. The p value for the Humanities is marginally significant, but the actual change in share of a publication language does not surpass 1.5 %. By contrast, for the academic peer reviewed books, more change has occurred. As expected on the basis of the results on shifting places of publication, change has been greatest in the Humanities, which increased their use of English by almost 10 %. In the Social Sciences, which already published in English by and large during the first subperiod, a further modest decline of the use of Dutch has occurred, to the benefit of other publication languages.

All in all, knowledge on the geographic center of weight of places of publication offers an insightful perspective on internationalization of book publishing in the SSH. As we have argued, this perspective pertains most obviously to the broadening of dissemination channels for SSH book literature. Looking at places of publication frames internationalization of book publishing in the interaction between academic authors seeking out a broader readership, and their publishers striving to bring academic content to the (international) market place. At the same time, there are likely consistent links between the place of publication and other intrinsic aspects of research internationalization. Two of these elements related to places of publication are authorship/editorship and collaboration in book publications; the results in Verleysen and Engels (2014) relating to different barycenter locations for the three book publication types already point in the direction of an intrinsic link with the place of publication. One way to take these issues forward could be a combined analysis of, on the one hand, places of publication, and, on the other, of international co-authorships of book publications. Such an analysis could yield further insight into the precise relationship between academic authors, their publishers and the places of publication of academic books in the SSH.

Conclusion

The results of our analysis imply that the gap between peer reviewed and non-peer reviewed SSH book literature is widening in terms of places of publication. This goes hand in hand with the growing use of English by Flemish affiliated SSH researchers for their academic book publications. From the perspective of multiple SSH book literatures catering to distinct readerships and published by dedicated publishers, the results in this article therefore suggest a process of academic and enlightenment literature growing further apart. The distance between readerships is currently greatest in the Social Sciences and in Law, although the Humanities do show a stronger evolution towards publishing peer reviewed books internationally, especially by publishers in the UK. Moreover our results suggest that the use of peer review as a criterion for distinguishing academic from enlightenment book literature, is more valid for the Social Sciences and less so for the Humanities, where the boundaries between different literatures, academic versus enlightenment and international versus domestic, remain less clear cut.