Robert Elgie and the nature of political science

I argue that Robert Elgie's late political leadership duology makes a remarkable contribution to the debate on political science methodology and the philosophy of social science. It demonstrates that a wide range of methodologies are consistent with scientific realism can speak to the same research question. I draw out implications for the methodology and organisation of political science.


Introduction
This piece is a memorial essay on a less-noticed aspect of Robert's voluminous work. It is not a personal memorial, although I hope my affection is obvious. Neither is it a review essay that tries to summarise Robert's career. Instead, I will place two late, great books in the context of Robert's career and the future of our subject. It is an interpretation, which I will hope will be interesting for some readers. I make no claim that Robert would agree with most, or even any, of it. This is not something I would have dared to attempt as a contribution to a festschrift! institutionalist with a reputation and relevance way beyond France or even semipresidentialism. This trajectory is a classic comparative politics story, which could be seen to map on to the almost archetypical case of Arend Lijphart. He established his reputation with a country-case study; went on to argue that important elements of its politics were shared by many other consociational systems; and became a leading student of all democratic regimes, including majoritarian countries, which operated according to a logic that was the opposite of consociationalism. Robert's career was not as linear as Lijphart's. He maintained a productive synergy between France, semi-presidentialism, and wider questions in political science throughout his career (Elgie 2001;Elgie and McMenamin 2005;Baturo and Elgie 2019). His early book on political leadership, although formally a textbook, established a reputation for bold and lucid synthesis of many literatures, while still a young man (Elgie 1995). He remained committed to the study of French politics, through his editorship of this journal and co-editorship of the Oxford Handbook of French Politics (Elgie et al. 2016).
Lijphart appears to have been generally consistent in his approach to the collective project of comparative politics and, while he made methodological contributions earlier in his career, showed little interest in its philosophical underpinnings. He published on the comparative method as a practitioner, rather than a philosopher of social science. Robert again is less straightforward than Lijphart in relation to philosophy of science. Robert is regarded as a first and foremost an institutionalist and there are many others better qualified than I am to discuss his achievements in that respect (Passarelli 2021). Here, I want to reflect on Robert's relationship with the nature of the collective political science project. It is a less prominent, but potentially discipline-transforming, part of his legacy.

Robert Elgie as a story of shifting paradigms
Since paradigms were not an explicit subject of Robert's written work, there is necessarily some speculation in this section. Nonetheless, the language and methodology of his work provide strong indications of an underlying paradigm. I can also bolster the account with some reminiscences. Until relatively recently, paradigms were a subject Robert talked about, rather than wrote about.
Robert wrote his PhD at a time when the study of French politics in the UK, was still very much part of the traditional British school of political studies. It was eclectic, but perhaps more reminiscent of contemporary history, than sociology or economics. Informal as it may have been, there was a strong institutionalist bent to this school, led by scholars like Vincent Wright and Jack Hayward. There was little sign of Michel Foucault and not much evidence of Anthony Downs either. Robert's work was notably structured: you might say his PhD effectively leveraged variation on the independent variable of legislative-executive relations to explain variations in the role of the prime minister (Elgie 1993: 5, 185-187). Nevertheless, for many years his work remained case-based and strictly qualitative. So, his first (proto-)paradigm was the British School.
The nineteen-nineties were the time of the institutionalist turn in political science. Indeed, institutionalism could be combined with other major approaches to the study of politics. Hall and Taylor famously distinguished sociological, historical, and rational-choice institutionalism (Hall and Taylor 1996). The British School could fairly comfortably join the historical-institutionalist movement, but by the late nineteen-nineties Robert committed himself to rational-choice institutionalism. Robert's methodological leitmotiv was that formal institutions are observable and this suited the positivist thrust of the rational-choice theorists. However, it was also a career move. In order to gain recognition from those doing cutting-edge research at the world's leading universities, rational choice was the rational choice. It was, at least for a period, virtually the paradigm of the American Political Science Review. The rational-choice theorists had almost a full-blown paradigm and many wished to establish its hegemony and eradicate pre-scientific competitors. Robert signed up to a course on game theory. He began reading classics of institutional economics. (After all, at Oxford he had read Politics, Philosophy, and Economics, not history or French.) Robert's paradigm shift coincided more or less with his appointment as Paddy Moriarty Professor of Government and International Studies at Dublin City University. His inaugural was mostly based on his article on the European Central Bank, rooted in economic theory (principal-agent frameworks and credible commitments) and economic methodology (the construction of legal and behavioural indices to measure institutional independence) (Elgie and Thomson 1998;Elgie 2002). He used it to define the study of politics as the study of institutions.
Soon after joining DCU, Robert tried to convince all staff and research students to read King, Keohane, and Verba's (decidedly positivist, but not necessarily rational choice) Designing Social Inquiry. (Some readers can imagine how that went. Gary Murphy can still recite the whole text backwards.) He continued to use the credible commitment approach, but now to study agencies and regulators. He and I wrote a piece for Public Choice that is presented in the stark normal-science language of that journal (Elgie and McMenamin 2008). We replicated the work of other scholars then solved the puzzle of how to extend the dominant theory to new contexts by adding a new variable. In this period, Robert often spoke about "best practice" and "the industry standard". We would identify the dominant technique and then try to master and apply it. Around this time our then colleague, Francesco Cavatorta, came back from an ECPR conference saying that when a scholar had found out he worked at DCU, he said, "Oh, so then you must know, Robert Elgie, the God of Semi-Presidentialism". This was also about the time, Robert abandoned the more hard-core rational-choice literature, for almost two decades of spectacular success leading a research programme on semi-presidentialism, which inspired not just a plethora of junior scholars, but also established political scientists such as Cheibub, Samuels, Shugart, and others.
Robert was empirically and theoretically innovative, but, so far, he had been, like most of us, methodologically derivative and structurally conventional. Indeed, he had often been insecure about methodology. I remember him baulking at the suggestion that he replace me as the teacher of research design to our PhD students. (Robert was not at all reluctant to do as much teaching as the rest of us; he genuinely doubted his own ability.) Some years later, I remember he made some references to being exhausted due to "having to read so much constructivism", but did not hint to me at the transformation to come. He volunteered to present to our school seminar and the title, I think, was just something about semi-presidentialism. I was surprised to hear fifteen minutes about William James, scientific realism, and American pragmatism, followed by a more familiar empirical section about institutions. I think the seminar went well, but Robert turned up at my office ten minutes after the seminar finished saying, "Have I made a fool of myself?", "Have I taken a wrong turn?", "What did you think?". I think I muttered something positive, and maybe something about having read more of Henry, rather than William, James and not being very qualified to give an assessment.
Robert's pair of books for the Palgrave Political Leadership series are remarkable singly, even more remarkable as a set, and even more remarkable again, with some knowledge of Robert's professional and paradigmatic history. Robert could be scathing about constructivism and had a particular antipathy towards sociology. He had committed himself to the positivist project of the rational-choice school. Although, along with much of the political science establishment, he had resiled from the hegemonic pretensions of this school, he was at the apex of a hyper-successful career in mainstream comparative institutional politics. The books have the characteristic Elgie lucidity and thoroughness. His early book on political leadership was also broad, but here was breadth, depth, bold innovation, and generosity. The different literatures are placed in an overall meta-framework, but each is also presented on its own terms and he writes eloquently of the case for each of the paradigms for the study of political leadership. He admits that, philosophically and ultimately, it is not possible to adjudicate between the paradigms, but requests that scholars explicitly identify in which paradigm they are writing to they can be judged by its standards (Elgie 2015: 185). In the second book, he obeys his own advice and sets out his philosophy of social science. Robert articulates a version of scientific realism, in which there is a very complex social world of mind-independent entities, associated with causal processes (Elgie 2018: 15) and in which it is possible for humans to gain at least some knowledge of these entities and their associated causal processes (16), albeit within a very narrow "horizon of predictability" (Elgie 2018: 17). He then uses Piercean pragmatism to connect this philosophy of science to a diverse, but coherent, methodological approach (Elgie 2018: 18-22). He manages methodological triangulation among approaches (QCA, case study, experiments, inferential statistics) that are often seen as incompatible. The man who, not so many years previously, did not feel confident to teach new PhD students, achieved a multi-methodological tour de force (with a just a little help from David Doyle on experiments).

Taking paradigms seriously
Robert's aim was to make the case for the study of observable institutions. He did so in a way that is embedded in the literature on leadership, the philosophy of social science, and social science methodological developments of the last three decades, in way that others could not imagine, never mind emulate. As in many great books, the hints, asides, and implications are almost as profound and interesting as the explicit arguments. I would like to mention a few that have occurred to me here. In doing so, I am thinking out loud.

Foundational disclosure
Robert was frustrated at pointless inter-paradigmatic debates, which he thought could be ended by foundational disclosure. If scholars were to be explicit about the their ontological and epistemological foundations, criticism could be restricted to that which shared the philosophical foundations of the author. Indeed, if everybody was clear about their paradigms then everybody could save a lot of time by not even reading those from other epistemological and ontological worlds. Would this require a detailed and tedious ontological and epistemological statement at the beginning of all research works? (Not all of us have Robert's energy-the amount of reading behind his foray into the philosophy of science is terrifying!) Since most of us do not go through paradigm shifts between every article, could we just post a philosophical position online for reviewers and readers, to be updated if and when appropriate?

One paradigm, one department?
Indeed, if we are working in different paradigms, we're living in different worlds (Kuhn 2012(Kuhn [1962: 111), and should not pretend we are studying more or less the same subject in more or less the same way. Should we immediately divide our university departments, scholarly associations, and journals into separate paradigms? Robert was quite allergic to the word sociology. In retrospect, I think he did not so much deny the legitimacy or quality of sociological thought. He just did not want the confusion and tension that the admission of another paradigm into our structures and networks would cause. Regardless of how political sociology is, it had developed an incommensurably different vision of the study of politics to Robert's view that centred on observable institutions.

Methodological multi-competence?
Political science requires more skills and more flexibility than many other disciplines. PhD graduates are usually expected to have theoretical and empirical skills, usually with a substantive specialism too. PhD graduates are also usually required to have some minimum competence in both qualitative and quantitative methods. Moreover, most political scientists have to gain some degree of mastery over legal, policy, and contextual details too. For comparative politics, an area specialism and language(s) are usually still expected, although the comparative politics in which Robert began his career has, to some extent, been divided into comparative politics and area studies branches. If all that sounds like a jack-of-all-trades rather than an academic expert, we can add in the portfolio of methods used by Robert in his Political Leadership duology, plus data science.

The organisational consequences of the paradigmatic peace?
For the practise of political science, Scientific Realism is a thin framework, not a thick paradigm. It allows, and indeed justifies and encourages, much diversity, for example, Robert's embrace of Piercean Pragmatism. Indeed, the scientific realist position downgrades the importance of paradigmatic thinking in the social sciences. No thick theory can, or should, aspire to hegemony, but neither can any school evade criticism or debate by playing the incommensurability card. There are few better ways of ending a conversation than declaring incommensurability. It is obvious that not all political scientists accept a scientific realist philosophy of social science. Some constructvists deny the existence of reality and believe our observations to be theory-determined. Some positivists think we can make observations independent of our theories. However, Robert hints that most of us would accept a scientific realist position that our observations are theory-laden and we can assess their practical adequacy because there is a political reality outside of our minds to which we have some access. I share this intuition, but it is hard to get much further, as so few political scientists make explicit statements and paradigmatic professions are not always borne out by research practises. Some constructivists make highly relativistic statements in the abstract, but then appear to argue about evidence in a way that seems to be somewhat realist. Many positivists employ universalistic terms and use the language of natural science, but then make tentative claims about very small portions of the political universe, more in the keeping with a scientific realist practise than the normal-science aspirations they appear to cherish. Methodological multi-competence is not a reasonable proposition for the individual, but it is for a team and this is one of the great strengths of natural science. Co-authorship has become much more common, although the science-model of well-organised and funded teams is still quite marginal to social science. (Robert had little interest and less success in funding applications.) Publication is the ultimate social process in social science. If we believed in scientific realism, we could reorganise accordingly. Methodological diversity would be required in reviews, in conference discussions, letters of support for promotion, and in literature reviews. Specialised comfort zones would be replaced by debates seeking, not paradigmatic victory, but tentative, temporary, mid-range consensus on empirical phenomena.

Claims and disclaimers
Of course, this previous section is speculative and such organisational changes unlikely. Nevertheless, I think we should consider that, although scientific realism may be a reasonable and moderate philosophy of science, the consequences of taking it seriously are radical. Thus, I submit, did Robert Elgie move from a being a talented follower of a dominant paradigm of rational-choice institutionalism to writing a two-volume set with the potential to establish him as meta-paradigmatic leader. This contribution seems to offer a hope, however faint, of finally ending our unsatisfactory fragmentation, not with the triumph of a hegemonic normal science, or the splitting into separate paradigms, but with a new acceptance and harnessing of diversity that is both rigorous and moderate. I am not sure how much of this Robert would agree with, but I hope he would be pleased that I was so inspired by his work I could not restrain myself to merely summarising or analysing it. I wanted to run with it and that others will want to take your ideas further is the dream of any scholar.
Funding Open Access funding provided by the IReL Consortium.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visithttp:// creat iveco mmons. org/ licen ses/ by/4. 0/.