Publication Patterns in the Humanities: Generational Shifts and Changing Research Agendas

1 Publication Patterns in the Humanities: Generational 1 Shifts and Changing Research Agendas 2 Nicolas Robinson-Garcia1, Wenceslao Arroyo-Machado2*, Elvira González-Salmón1 and 3 Daniel Torres-Salinas1 4 5 1Unit for Computational Humanities and Social Sciences (U-CHASS), EC3 Research Group, 6 University of Granada, Granada, Spain 7 2Center for Science, Technology and Environmental Policy Studies (CSTEPS), School of Public 8 Affairs, Arizona State University, Phoenix, AZ 85004, USA 9 10 * Corresponding author: [email protected] 11 Abstract 12 The humanities have distinct publication practices compared to the sciences, with books and 13 local language literature being essential. This study aims to identify and analyze the 14 publication patterns of humanities scholars in Spanish-speaking countries, revealing unique 15 publication behaviors and fostering diverse perspectives rather than linear knowledge 16 progression. We analyzed the publication histories of approximately 40,000 scholars from 17 1950 to 2021 using data from the Dialnet database. By identifying archetypal publication 18 profiles, we explored their distribution across generational cohorts and research topics. Our 19 findings reveal substantial generational shifts towards journal-centric publication patterns 20 probably influenced by bibliometric-driven evaluation systems. The also show a relation 21 between publication patterns and research topics. This highlights the need for more inclusive 22 assessment practices that recognize the diverse nature of humanities scholarship. We 23 contribute to ongoing discussions on promoting bibliodiversity in research assessment and 24 the potential impacts of current policies on the humanities. 25 Keywords 26 humanities, publication patterns, bibliodiversity, research evaluation, Spanish-speaking 27 countries, generational shifts 28 29 30 31 32 33 34 35 36 37 38 39 40 This is a preprint and has not undergone peer review 41 The final published version may differ 42 10.5281/zenodo.17279808 | Version v3 43 2 The humanities are distinct from other sciences. Unlike other disciplines, books remain an 44 essential format of scientific publishing1–3, and their target audience is often more scattered 45 and diverse4,5 with non-English literature playing an important role in its dissemination6. 46 Citation patterns differ as well7: older works are more likely to still be cited5, and research 47 primarily involves exploring new perspectives and fostering the coexistence of competing 48 knowledge rather than following a linear process of development characterized by big 49 breakthroughs8. Furthermore, solo research is still common9. These disciplinary specificities 50 have led to the emergence of diverse valuation regimes,– that is, differing logics, practices 51 and standards by which research quality and relevance are judged 10,11. As a result, research 52 evaluation remains a highly controversial issue12,13. This lack of consensus is related to the 53 social structure of these fields, generally organized around schools of thought or ‘tribes’14 54 which compete, contrasting with the ‘paradigmatic’ organization of the Natural and Exact 55 Sciences15. These differences have sometimes been interpreted as weakness of 56 humanities16,17, rather than merely different research styles. 57 58 Given the particularities of the humanities, a body of literature has been built within the fields 59 of scientometrics and research evaluation over the last twenty years trying to better 60 understand its dynamics18. For this, studies have focused mainly on three issues: differences 61 on publication patterns, database coverage limitations 19–22, and the validity of citations as a 62 proxy for impact or quality8,23 proposing a range of alternative indicators3,24,25. 63 64 This study is framed within the first stream of literature, that is, the understanding of publication 65 patterns within the Humanities. Our goal is to identify types of humanists based on their 66 publication patterns and understand the factors underlying the differences between these 67 types. We want to understand to what extent humanists tend to publish in a diverse range of 68 outlets. Is it common to all fields? How is a journal-centric evaluation culture affecting their 69 publishing habits26? What is the role played by language27, database indexing28–30, scope31,32 70 or research topic28,31,33? 71 72 To this aim, we analyze publication patterns of individual scholars during their complete 73 academic career. We examine the publication history of 39,753 scholars from Spanish 74 speaking countries who started publishing from 1950 onwards up to 2021. We consider around 75 1.2 million publications in all outlets and languages, studying 13 different fields from the 76 humanities. Our dataset is extracted from Dialnet (https://dialnet.unirioja.es/), a specialized 77 bibliographic database focused on the social sciences and humanities, maintained by the 78 Dialnet Foundation in collaboration with over 80 Spanish and Latin American universities and 79 public institutions34,35. Dialnet indexes diverse publication types with a strong focus on 80 Spanish-language outputs, aiming to reflect national and regional scholarly production in the 81 Spanish-speaking world often overlooked by international citation indexes. The uniqueness of 82 this data lies not only on the richness of publication types and languages, but also by the fact 83 that author profiles are manually curated and regularly updated by Spanish university, public 84 and special libraries related to the Dialnet Foundation through a consortium agreement. This 85 allows us to track their complete publication history, identify their starting publishing date, 86 compute their career length or differentiate between outputs indexed and non-indexed 87 publications in mainstream bibliometric databases (i.e., Scopus and Web of Science). 88 89 Based on this dataset, we built eight variables for each scholar, defining their publication 90 patterns by publication type, language, and database indexing. We then computed an 91 3 archetypal analysis36 per discipline. This allows us to identify extreme-type configurations of 92 humanists based on their publication patterns. A hierarchical clustering analysis allowed us to 93 group archetypes across fields, identifying six distinct publication profiles in the humanities. 94 We then computed distance measures between each scholar and archetype to measure their 95 resemblance to them. Furthermore, we assigned scholars to their most similar archetype and 96 explored differences between archetypes in terms of generational cohort, and research topic. 97 Topics were identified by applying the Leiden community detection algorithm37 to a co98 occurrence network based on keywords extracted from publication titles. Then, we computed 99 the Jaccard distance between archetypes across fields in order to identify differences in the 100 topics studied by humanists based on their publication profiles. Further details are provided in 101 the Data and Methods section. 102 103 We contribute to current literature in several ways. First, we analyze publication patterns at 104 the author level, revealing the extent of variation and contributing to the debate on the diverse 105 literatures within the humanities9,38. These patterns may not be only disciplinary but also 106 generational, as the rise of bibliometric-driven evaluation systems and the “Publish or Perish” 107 culture has incentivized shifts in what, how, and where humanists publish39,40. Second, we 108 offer novel insights into current research policy discussions on promoting bibliodiversity in 109 research assessment and the potential risks of policies that alter publication habits26,41. Finally, 110 we examine the relationship between research topics, language and geographic outreach — 111 an under-researched issue often mentioned to resist changes in publication patterns within 112 the humanities 6,31. 113 Results 114 Distribution of scholars by field and general publication trends 115 We explore the publication patterns of 39,753 humanists distributed among 13 research fields 116 (Fig. 1). Historians, philologists, and philosophers represent half of the population (50.2%), 117 whereas Geography, Paleontology and Cultural Studies account for less than 5% of the 118 population (a). This distribution is modified slightly when looking at productivity differences 119 among fields (b). The proportion of outputs from the three largest fields increases slightly to 120 52.7%, with History representing more than a quarter of the publications (26.8%), that is 4.3% 121 higher than expected. Literature also produced more publications than expected compared to 122 other fields, with a difference of 3.0 points. Conversely, Philosophy constitutes 12.4% of the 123 workforce but accounts for only 10.2% of all publications. 124 125 126 127 128 129 130 131 132 133 134 4 Fig. 1 Descriptives of population of scholars and outputs by field. 135 136 137 The distribution of career lengths and first publication years of scholars across fields reveals 138 significant variations. In terms of career length (c), Literature and Geography stand out with 139 the highest mean values (21.7 and 21.2 respectively), indicating the presence of more senior 140 scholars in these fields. By contrast, scholars from fields such as Translation Studies, 141 Anthropology, and Music have shorter career lengths on average, indicating a younger 142 academic workforce. Regarding generational differences (d), most scholars began their 143 academic career during the 1990s and 2000s; however notable differences exist across fields. 144 Translation Studies and Anthropology show the most recent mean entry points into academia, 145 with means around 2003 and 2002, and medians both in 2005. On the other hand, Literature 146 and Geography have the oldest mean first publication years, around 1995, indicating a more 147 established cohort of scholars. 148 149 Regarding publication patterns (e), articles are the predominant publication type across most 150 fields, particularly in Paleontology (79.2%) and Music (78.1%). Philosophy also shows a high 151 proportion of articles (62.0%). In contrast, fields like Philology and History, while still favoring 152 articles (47.4 and 49.7%, respectively), show a larger presence of book chapters (31.6 and 153 31.4% respectively). The prevalence of books is particularly notable in Literature (16.8%) and 154 Philosophy (11.7%). Fields like Anthropology and Archaeology exhibit a balanced distribution 155 between articles and chapters, with Translation Studies showing 46.8% articles and 31.3% 156 chapters, Philology with 47.4% articles and 31.6% chapters, and Geography with 46.6% of 157 articles and 29.8% of chapters. In terms of publication language (f), Spanish is the dominant 158 language in most fields. It accounts for 89.8% of publications in Literature, 82.6% in Cultural 159 Studies and 79.9% in Philosophy. In contrast, Paleontology stands out with a significant 160 proportion of outputs in English language publications (44.5%). English also plays a notable 161 role in fields like Translation Studies (30.0%) and Philology (20.8%). Other languages such as 162 5 Catalan, French and Portuguese are also present but to a lesser extent. Catalan is particularly 163 present in Philology (9.2%) and Geography (7.6%). 164 Humanists’ publication profiles 165 For each scholar, we defined eight variables to describe their publication patterns to better 166 understand their preferred outlets. Six of these variables relate to publication types: four relate 167 to authored publications (journal articles, books, chapters and proceedings), and two with 168 edited publications (edited conference proceedings and edited books). The other two variables 169 relate to the geographic scope of the output. One relates to the number of non-English 170 publications and the second to the number of publications indexed in mainstream international 171 scientific databases (i.e., Scopus and Web of Science). Further details on the data processing 172 and methodological design are provided in the Data and Methods section. 173 174 Using these variables, we applied an archetypal analysis to identify the types of publication 175 profiles exhibited by humanists. Archetypal analysis36,42 is a methodology used to identify 176 extreme points in a multidimensional dataset that best represent the diversity of the data. 177 Unlike traditional clustering methods, which group similar data points into clusters, this method 178 identifies extreme examples within the data, providing a convex hull that encompasses the 179 dataset. Furthermore, it provides a similarity measure, called the α-score, which indicates how 180 similar each case is to the identified archetypes. The number of archetypes is defined by 181 following an elbow criterion after plotting the screeplot of a residual sum of squares (RSS) 182 analysis (Appendix, Fig. A1-14). Hence, the number of archetypes identified varies per 183 discipline, with 8 disciplines exhibiting three different archetypes and 5 disciplines exhibiting 184 2. In order to identify similarities across fields, we conducted a hierarchical clustering analysis 185 of all archetype-discipline combinations. In this way, we are able to identify commonalities and 186 differences of profiles across and within disciplines. 187 188 Overall, we have identified six profiles which are spread across the humanities. Figure 2A 189 displays a heatmap showing the normalized values of the eight publication variables used to 190 define each archetype. Archetypes are sorted and grouped based on hierarchical clustering. 191 The corresponding fields and archetype numbers are shown on the right side of the y-axis. To 192 enhance readability and engagement, we gave each cluster of archetypes or profile, a 193 descriptive name that reflects its core characteristics. Figure 2B shows the distance between 194 these profiles, displaying how they cluster together based on their publication patterns. These 195 are the following: the bridger, the cosmopolitan, the local chronicler, the sage, the polymath 196 and the participant. By naming these profiles, we aim to provide a clearer understanding of 197 the diverse publication practices in the humanities and highlight the generational and cultural 198 shifts influencing these patterns. 199 200 201 202 203 204 205 206 207 208 6 Fig. 2. Clustering and spatial distribution of archetypes in the humanities: (a) Hierarchical 209 clustering and heatmap based on publication profiles, (b) MDS visual representation of 210 disciplines and archetypes reflecting their relational distances. 211 212 213 The most widespread profiles are those of the participants and the local chroniclers, which are 214 both present across 11 of the 13 disciplines analyzed. In the case of the participants, the two 215 fields in which they are absent are Language & Linguistics, and Literature. This profile is 216 characterized by publishing in national languages mainly. We name them participants as they 217 tend to participate in collaborative works with book chapters as their preferred publication 218 venue. However, researchers exhibiting a participant publication profile will occasionally use 219 any of the other publication venues analyzed, except for journal articles indexed in Web of 220 Science or Scopus. In the case of the local chroniclers, these are present in all disciplines 221 except for Language & Linguistics, and Paleontology. They are characterized by publishing 222 mainly journal articles in national language (i.e., non-English), and only occasionally books or 223 in indexed journals. 224 225 The third most common profile is that of the cosmopolitans. Named like that due to the high 226 proportion of publications which are placed as articles in mainstream journals. Cosmopolitans 227 are present in 7 disciplines, these are: Philology, History, Archaeology, Arts, Translation 228 Studies, Anthropology and Philosophy. Next, we observe two publication profiles which are 229 only present in two disciplines respectively. These are the bridgers and the polymaths. 230 Bridgers are scholars who primarily publish journal articles, engaging with both national and 231 mainstream communities by publishing in both, in national and English languages, as well as 232 in indexed and non-indexed journals. These scholars are present in the disciplines of 233 Language & Linguistics, and Paleontology. On the other hand, polymaths tend to use a wide 234 variety of publication venues, similarly to the profile of the participants, although with a lesser 235 emphasis on book chapters. And again, they publish in non-English languages and in non236 mainstream journals. These scholars are present in Language & Linguistics, and Literature. 237 Finally, we observe one profile which is distinctive of just one of the disciplines in the 238 Humanities, that is the sage, which is only present in Literature. These scholars are 239 characterized by publishing mainly monographs in national language, although occasionally 240 also publishing journal articles. 241 7 242 Overall, we observe how publication choices are varied within the humanities, with different 243 publication patterns observed within and across fields. While many similarities exist among 244 some of the identified profiles, others are unique to specific disciplines, reflecting the richness 245 and diversity of publication habits. 246 Generational differences on publication patterns by field 247 Next, we hypothesize the existence of different profiles within disciplines as a reflection of a 248 generational shift, where publication patterns respond to the context in which they are 249 produced. Phenomena such as bibliometric-driven evaluation systems, a Publish or Perish 250 culture, and a generation of digitally native scholars who may prioritize novelty and shorter 251 communication formats over longer, more reflective works, could be driving these differences 252 across cohorts. 253 254 We explore this in Figure 3, which plots the share of scholars by publication profile across 255 disciplines over time. The x-axis represents scholars’ first publication year which is used as a 256 proxy for their generational cohort, ranging from 1960 to 2020. The data reveals that 257 generational shifts are taking place both, in terms of dominant archetypes within disciplines, 258 and the emergence of new publication profiles. Similarities can be found in the fields of History, 259 Anthropology, Archaeology and the Arts, dominated by the local chronicler and the participant 260 profiles, but with an incipient growth of the cosmopolitan profile among younger scholars. 261 Meanwhile, the other two profiles, the participant and the local chronicler, remain cross262 generational, suggesting a more stable presence in these fields across different cohorts. A 263 similar trend is observed in Philology and Philosophy, where the cosmopolitan profile has 264 become more prominent among newer generations. However, these show that the local 265 chronicler profile is adopted by older scholars, with a marked decline especially in Philology 266 since the late 1990s, indicating a shift from publishing in collective monographs. 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 8 Fig. 3 Generational shifts in publication profiles across disciplines. 288 289 290 In Cultural Studies and Geography, we observe a co-existence between the participant profile, 291 characterized by publishing mainly book chapters, and the local chronicler, which prioritizes 292 journal articles in national language. In the case of Geography, we observe a U shape in the 293 distribution of participants, whereas in Cultural Studies, there is an even split which oscillates 294 over the study period. But there is no dominant pattern across fields, reflecting the richness 295 and diversity of the humanities in terms of publication patterns. In the case of Music, for 296 instance, the pattern is somewhat the reverse, with the local chronicler being a cross297 generational profile and the participant profile being adopted by generations from the 1990s 298 and 2000s, although declining in the case of younger generations. 299 300 Paleontology is a unique case, showing distinct patterns compared to other disciplines. Here 301 the bridger—who combines publishing in national language with mainstream journals—, is the 302 dominant profile, with a smaller share of participant profiles are present across multiple 303 generations. Other distinct cases include Language & Linguistics, and Literature. In the case 304 9 of the former, two profiles are present: the polymath, who tends to use a diversity of publication 305 venues and formats, and the bridgers. Scholars showcasing the polymath profile are cross306 generational, while there is a clear rise of scholars with a bridger profile among recent 307 generations, pointing towards an effort to reach wider and more international audiences. 308 Finally, Literature is the only discipline exhibiting the sage profile, characterized by scholars 309 who primarily publish books. However, this profile is associated with older generations of 310 scholars. In contrast, newer generations in the field tend to exhibit a polymath and a local 311 chronicler profile, suggesting a shift in terms of publication practices. The dominance of the 312 local chronicler profile still suggests the importance of national-language journals within the 313 field over mainstream journals. 314 315 Fig. 4 Thematic landscape and comparison of research portfolios by archetype in History, 316 Philology, Philosophy and Arts., which represent the four largest fields in our dataset. 317 318 The relation between research topics and publication patterns 319 Next, we explore whether publication profiles are associated with broad tendencies in topic 320 engagement. Since archetypes are derived from patterns in publication, comparing their 321 thematic focus allows us to assess how these publication behaviors relate to research topics. 322 We translated non-English titles into English and extracted keywords using the ChatGPT 323 API43. We then applied the Leiden community detection algorithm37 to group keywords into 324 clusters or topics. We identified a total of 1,437 topics (see Supplementary material to read 325 the list of topics) which were clustered into eight major areas. Based on these topics, we 326 created a vector for each archetype-discipline combination, representing the distribution of 327 records among topics. A detailed description of the methodological approach followed is 328 provided in the Data and Methods section. 329 16 community detection algorithm to further cluster topics into 8 major fields. From the general 581 map, an overlay was created for each archetype within each discipline, showcasing the 582 research fronts where they publish the most. 583 Thematic profile similarity 584 Finally, we represented each humanist profile as a vector, where each entry corresponds to 585 the total number of publications in a specific research topic. Pairwise distances between these 586 topic vectors were then computed using the Jaccard distance, which quantifies dissimilarity 587 by comparing shared and unique topics between profiles. The Jaccard distance 𝑑𝑗 between 588 two sets 𝐴 and 𝐵 is defined as: 589 𝑑𝑗(𝐴, 𝐵)= 1 − |𝐴∩𝐵| |𝐴∪𝐵| 590 591 where |𝐴 ∩ 𝐵| is the size of the intersection (i.e., the number of shared topics) and |𝐴 ∪ 𝐵| is 592 the size of the union (i.e., the total number of unique topics across both profiles). 593 Subsequently, for each field, we applied a one-dimensional Multidimensional Scaling (MDS)53 594 analysis to visualize how profiles align based on thematic similarity across. This analysis 595 allowed us to assess whether the observed thematic patterns at the field level were consistent 596 across the broader dataset. 597 Funding Declaration 598 This work is part of the COMPARE project (REF: PID2020-117007RA-I00) funded by the 599 Spanish Ministry of Science (Ref: MCIN/AEI/10.13039/501100011033 FSE invierte en tu 600 futuro). Nicolas Robinson-Garcia is funded by a Ramón y Cajal grant from the Spanish Ministry 601 of Science and Innovation (REF: RYC2019-027886-I). Elvira González-Salmón is funded by 602 a FPU grant from the Spanish Ministry of Universities (FPU2021/02320). We would like to 603 thank the Dialnet Foundation for providing access to the data. 604 605 References 606 1. Engels, T. C. E., Istenič Starčič, A., Kulczycki, E., Pölönen, J. & Sivertsen, G. Are book publications 607 disappearing from scholarly communication in the social sciences and humanities? Aslib Journal 608 of Information Management 70, 592–607 (2018). 609 2. Giménez-Toledo, E., Mañana-Rodríguez, J. & Sivertsen, G. Scholarly book publishing: Its 610 information sources for evaluation in the social sciences and humanities. Res Eval 26, 91–101 611 (2017). 612 3. Zuccala, A. & Robinson-García, N. Reviewing, Indicating, and Counting Books for Modern 613 Research Evaluation Systems. in Springer Handbook of Science and Technology Indicators (eds. 614 Glänzel, W., Moed, H. F., Schmoch, U. & Thelwall, M.) 715–728 (Springer International Publishing, 615 Cham, 2019). doi:10.1007/978-3-030-02511-3_27. 616 4. Sivertsen, G. Patterns of internationalization and criteria for research assessment in the social 617 sciences and humanities. Scientometrics 107, 357–368 (2016). 618 5. Hammarfelt, B. Beyond Coverage: Toward a Bibliometrics for the Humanities. in Research 619 Assessment in the Humanities (eds. Ochsner, M., Hug, S. E. & Daniel, H.-D.) 115–131 (Springer 620 International Publishing, 2016). doi:10.1007/978-3-319-29016-4_10. 621 17 6. Kulczycki, E. et al. Multilingual publishing in the social sciences and humanities: A seven-country 622 European study. Journal of the Association for Information Science and Technology 71, 1371– 623 1385 (2020). 624 7. Wang, J. Citation time window choice for research impact evaluation. Scientometrics 94, 851–872 625 (2013). 626 8. Ochsner, M., Hug, S. E. & Daniel, H.-D. Four types of research in the humanities: Setting the stage 627 for research quality criteria in the humanities. Research Evaluation 22, 79–92 (2013). 628 9. Nederhof, A. J. Bibliometric monitoring of research performance in the Social Sciences and the 629 Humanities: A Review. Scientometrics 66, 81–100 (2006). 630 10. Robinson-Garcia, N., Costas, R., Nane, G. F. & van Leeuwen, T. N. Valuation regimes in 631 academia: Researchers’ attitudes towards their diversity of activities and academic performance. 632 Research Evaluation 32, 496–514 (2023). 633 11. Reale, E. et al. A review of literature on evaluating the scientific, social and political impact of social 634 sciences and humanities research. Research Evaluation 27, 298–308 (2018). 635 12. Toledo, E. G. Why Books are Important in the Scholarly Communication System in Social Sciences 636 and Humanities. 2, 6 (2020). 637 13. Hammarfelt, B. Recognition and reward in the academy: Valuing publication oeuvres in 638 biomedicine, economics and history. Aslib Journal of Information Management 69, 607–623 639 (2017). 640 14. Becher. Academic Tribes and Territories: Intellectual Enquiry and the Cultures of Discipline. (Open 641 University Press, Buckingham, 2001). 642 15. Kuhn, T. S. The Structure of Scientific Revolutions. (University of Chicago Press, 1996). 643 16. Sokal, A. D. A physicist experiments with cultural studies. Lingua franca 6, 62–64 (1996). 644 17. Linmans, A. J. M. Why with bibliometrics the Humanities does not need to be the weakest link. 645 Scientometrics 83, 337–354 (2010). 646 18. Verleysen, F. T. & Weeren, A. Clustering by publication patterns of senior authors in the social 647 sciences and humanities. Journal of Informetrics 10, 254–272 (2016). 648 19. Hicks, D. The difficulty of achieving full coverage of international social science literature and the 649 bibliometric consequences. Scientometrics 44, 193–215 (1999). 650 20. Sivertsen, G. & Larsen, B. Comprehensive bibliographic coverage of the social sciences and 651 humanities in a citation index: an empirical analysis of the potential. Scientometrics 91, 567–575 652 (2012). 653 21. Glänzel, W., Thijs, B. & Chi, P.-S. The challenges to expand bibliometric studies from periodical 654 literature to monographic literature with a new data source: the book citation index. Scientometrics 655 109, 2165–2179 (2016). 656 22. Archambault, É., Vignola-Gagne, É., Côté, G., Larivière, V. & Gingrasb, Y. Benchmarking scientific 657 output in the social sciences and humanities: The limits of existing databases. Scientometrics 68, 658 329–342 (2006). 659 23. Hug, S. E., Ochsner, M. & Daniel, H.-D. Criteria for assessing research quality in the humanities: 660 a Delphi study among scholars of English literature, German literature and art history. Research 661 Evaluation 22, 369–383 (2013). 662 24. Torres-Salinas, D. & Moed, H. F. Library Catalog Analysis as a tool in studies of social sciences 663 and humanities: An exploratory study of published book titles in Economics. Journal of Informetrics 664 3, 9–26 (2009). 665 25. Hammarfelt, B. Using altmetrics for assessing research impact in the humanities. Scientometrics 666 101, 1419–1430 (2014). 667 26. Hammarfelt, B. & de Rijcke, S. Accountability in context: effects of research evaluation systems on 668 publication practices, disciplinary norms, and individual working routines in the faculty of Arts at 669 Uppsala University. Res. Evaluat. 24, 63–77 (2015). 670 27. van Leeuwen, T. Bibliometric research evaluations, Web of Science and the Social Sciences and 671 Humanities: a problematic relationship? Bibliometrie - Praxis und Forschung 2, (2013). 672 18 28. López Piñeiro, C. & Hicks, D. Reception of Spanish sociology by domestic and foreign audiences 673 differs and has consequences for evaluation. Research Evaluation 24, 78–89 (2015). 674 29. Ossenblok, T. L. B., Engels, T. C. E. & Sivertsen, G. The representation of the social sciences and 675 humanities in the Web of Science—a comparison of publication patterns and incentive structures 676 in Flanders and Norway (2005–9). Research Evaluation 21, 280–290 (2012). 677 30. Kulczycki, E. et al. Publication patterns in the social sciences and humanities: evidence from eight 678 European countries. Scientometrics 1–24 (2018) doi:10.1007/s11192-018-2711-0. 679 31. Kulczycki, E. & Korytkowski, P. Researchers publishing monographs are more productive and 680 more local-oriented. Scientometrics 125, 1371–1387 (2020). 681 32. Chavarro, D., Tang, P. & Rafols, I. Interdisciplinarity and research on local issues: evidence from 682 a developing country. Research Evaluation 23, 195–209 (2014). 683 33. Rafols, I., Ciarli, T. & Chavarro, D. Under-reporting research relevant to local needs in the South: 684 Database biases in rice research. in The Transformation of Research in the South Policies and 685 outcomes (eds. Arvanitis, R. & O’Brien, D.) (Éditions des archives contemporaines, 2019). 686 34. Mateo, F. Producción científica en español en humanidades y ciencias sociales. Algunas 687 propuestas desde Dialnet. Profesional de la información 24, 509–516 (2015). 688 35. Arroyo-Machado, W. & Robinson-Garcia, N. Towards a Science of Humanities: How Big Data can 689 Solve the Limitations of Scientometrics. In A. Gallego Cuiñas & D. Torres-Salinas (eds.). 690 Humanities and Big Data in Ibero-America: Methodological issues and practical applications. De 691 Gruyter. 75–92 (De Gruyter, 2023). doi:10.1515/9783110753523-006. 692 36. Eugster, M. J. A. & Leisch, F. From Spider-Man to Hero — Archetypal Analysis in R. Journal 693 of Statistical Software 30, 1–23 (2009). 694 37. Traag, V. A., Waltman, L. & van Eck, N. J. From Louvain to Leiden: guaranteeing well-connected 695 communities. Scientific Reports 9, 5233 (2019). 696 38. Hicks, D. The Four Literatures of Social Science. in Handbook of Quantitative Science and 697 Technology Research 473–496 (Springer, 2005). 698 39. Cañibano, C., Vilardell, I., Corona, C. & Benito-Amat, C. The evaluation of research excellence 699 and the dynamics of knowledge production in the humanities: The case of History in Spain. Sci 700 Public Policy 45, 775–789 (2018). 701 40. Knöchelmann, M. Formal authorship in the wake of uncertain futures: the narrative of publish or 702 perish in the humanities. Research Evaluation 33, rvae044 (2024). 703 41. CoARA. The Agreement. Coalition for Advancing Research Assessment 704 https://coara.eu/agreement/the-agreement-full-text/ (2022). 705 42. Cutler, A. & Breiman, L. Archetypal Analysis. Technometrics 36, 338–347 (1994). 706 43. OpenAI. ChatGPT. (2023). 707 44. Frandsen, T. F. & Nicolaisen, J. Intradisciplinary differences in database coverage and the 708 consequences for bibliometric research. Journal of the American Society for Information Science 709 and Technology 59, 1570–1581 (2008). 710 45. Jiménez-Contreras, E., De Moya Anegón, F. & López-Cózar, E. D. The evolution of research 711 activity in Spain. Research Policy 32, 123–142 (2003). 712 46. Buela-Casal, G. Reflections on the Accreditation System of tenured University Professors in Spain. 713 Psicothema 19, 473–482 (2007). 714 47. Hicks, D., Wouters, P., Waltman, L., de Rijcke, S. & Rafols, I. The Leiden Manifesto for research 715 metrics. Nature 520, 429–431 (2015). 716 48. R Core Team. R: A Language and Environment for Statistical Computing. R Foundation for 717 Statistical Computing (2021). 718 49. Van Eck, N. J. & Waltman, L. Software survey: VOSviewer, a computer program for bibliometric 719 mapping. Scientometrics 84, 523–538 (2010). 720 50. Eugster, M. J. A. & Leisch, F. Weighted and robust archetypal analysis. Computational Statistics 721 & Data Analysis 55, 1215–1225 (2011). 722 51. Kolde, R. pheatmap: Pretty Heatmaps. (2019). 723 19 52. Rose, S., Engel, D., Cramer, N. & Cowley, W. Automatic Keyword Extraction from Individual 724 Documents. in Text Mining 1–20 (John Wiley & Sons, Ltd, 2010). 725 doi:10.1002/9780470689646.ch1. 726 53. Davison, M. L. & Sireci, S. G. 12 - Multidimensional Scaling. in Handbook of Applied Multivariate 727 Statistics and Mathematical Modeling (eds. Tinsley, H. E. A. & Brown, S. D.) 323–352 (Academic 728 Press, San Diego, 2000). doi:10.1016/B978-012691360-6/50013-6. 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 20 Appendix 766 767 Fig. A1 Screeplot of the residual sum of squares (RSS) of Humanities. The red dot represents the 768 value chosen for the analysis. 769 770 771 772 Fig. A2 Screeplot of the residual sum of squares (RSS) of History. The red dot represents the value 773 chosen for the analysis. 774 21 775 Fig. A3 Screeplot of the residual sum of squares (RSS) of Philosophy. The red dot represents the 776 value chosen for the analysis. 777 778 779 Fig. A4 Screeplot of the residual sum of squares (RSS) of Philology. The red dot represents the value 780 chosen for the analysis. 781 782 22 783 Fig. A5 Screeplot of the residual sum of squares (RSS) of Arts. The red dot represents the value 784 chosen for the analysis. 785 786 787 Fig. A6 Screeplot of the residual sum of squares (RSS) of Archaeology. The red dot represents the 788 value chosen for the analysis. 789 790 791 23 792 Fig. A7 Screeplot of the residual sum of squares (RSS) of Language & Linguistics. The red dot 793 represents the value chosen for the analysis. 794 795 796 Fig. A8 Screeplot of the residual sum of squares (RSS) of Literature. The red dot represents 797 the value chosen for the analysis. 798 799 24 800 Fig. A9 Screeplot of the residual sum of squares (RSS) of Music. The red dot represents the value 801 chosen for the analysis. 802 803 804 Fig. A10 Screeplot of the residual sum of squares (RSS) of Anthropology. The red dot represents the 805 value chosen for the analysis. 806 807 25 808 Fig. A11 Screeplot of the residual sum of squares (RSS) of Translation Studies. The red dot 809 represents the value chosen for the analysis. 810 811 812 Fig. A12 Screeplot of the residual sum of squares (RSS) of Geography. The red dot represents the 813 value chosen for the analysis. 814 815