Abstract
This article takes as its point of departure Stefano Guzzini’s recent call for ‘ontological theorizing’ as a reflexive engagement with central concepts. In an attempt to advance this agenda, the article presents an accessible overview of different approaches to concept analysis to stake out the field for a discussion of what ontological theorising might entail. The article advances the notion of concepts as ‘basic’ and lays out the parameters through which they obtain meaning, followed by a discussion of three approaches, which tackle the multifaceted nature of basic concepts within and across different contexts. These approaches are labelled ‘historical’, ‘scientific’ and ‘political(critical)’ and presented through the work of Reinhart Koselleck, Giovanni Sartori and Michel Foucault, respectively. The article notes that concept analysis, as discussed here, stands in tension with modern forms of theory building yet is a creative source for theorising that accepts the unstable, political and context-bound nature of ontology.
In a recent contribution to a discussion about the state of International Relations (IR) theory, Stefano Guzzini calls for what he termed ‘ontological theorizing’, namely a reflexive engagement with central concepts. 1 Guzzini argues that such an engagement is necessary, indeed inevitable, for two reasons. First, concepts give the field of IR its ontology. They provide analysts with an understanding of what is ‘out there’ and in doing so help to grasp relevant phenomena by naming and giving meaning to its features. Thus, Guzzini writes ‘one could also call it “constitutive” [theorizing] since it is mainly about theorizing the central phenomena that constitute the field of inquiry’. 2 Second, concepts are needed to construct theories: they not only provide the ontological building blocks of a theory, often in the form of basic assumptions, but also the components out of which theorists generate their arguments. As such, concepts are ‘co-constitutive of theories; they are the words in which…our theorizing is done’. 3 In other words, concepts give us the language both for formulating the phenomena we seek to explain/understand and the frameworks we build to explain/understand them. Given this, calling on theorists to pay attention to key concepts and reflect on their use makes sense, especially in the wake of the so-called Third Debate, which, among other things, encouraged scholars to question core assumptions on which theoretical arguments rest.
However, answering such a call is not an easy task and requires addressing two related but different sets of questions. First, what exactly does paying attention to concepts entail? How can, or should we reflexively engage with key concepts? Second, what consequences does this have for the way we think about and ‘do’ theory? In other words, how does a reflexive engagement with key concepts affect the practice of theorising? While these questions might be seen as two sides of the same coin, it is useful to treat them as two steps in the intellectual endeavour of ontological theorising because it allows becoming familiar with specialised debates underpinning each question and, hence, helps to appreciate the complexity of the task. Guzzini’s article offers pointers mainly in response to the first set of questions (although its ambition is clearly to address the second). He suggests that ontological theorising is, basically, a form of concept analysis. Borrowing a metaphor from Anna Leander, he notes that it involves writing an ‘unfinished dictionary [about the ontology] of the international’ based on ‘a much wider understanding of “conceptual analysis” than usually offered’. Yet his article does not go beyond noting that this understanding involves recognising the relevance of history and philosophy for ‘thinking the empirics’ by tracing a concept’s ‘historical anchorage’ and intellectual history. 4 While Guzzini’s formidable work on power offers specific examples in this regard, 5 it is useful to take a step back and ask what a ‘wider understanding’ of concept analysis might entail from a broader perspective. This is the aim of the present article. Picking up Guzzini’s thread, it addresses the first set of the above questions by presenting an overview of different approaches to concept analysis as the first leg of ontological theorising. That said, the second set of questions cannot be ignored, and so this article will also offer some thoughts on the link between concept analysis and theorising.
Of course, the study of concepts has long occupied a prominent place among historians and philosophers, a number of whom are cited in this article. The first generation of IR scholars was attentive to the complex nature of key concepts, although this sensibility was later lost in the quest for parsimony and the reduction of concepts to variables. 6 Over the last two decades, attention to concepts has returned to the field of IR as a result of two related developments. First, in the wake of the linguistic turn, parts of the field developed a more critical attitude towards the categories and terminologies we use and the mentalities behind them. Disillusioned with mainstream paradigms as analytical devices to comprehend the world, an increasing number of scholars now organise their research around prominent concepts such as security, power, democracy, sovereignty, or hierarchy and trace their historical evolution, multiple interpretations, manifestations, and political performances. Second, there is a sense that we are living through a period of social and geopolitical transformation, entering a world with late-modern features accompanied by challenges to structures of Western dominance which have shaped the IR discipline since its inception, changing our language along the way. 7 While these changes are experienced differently, they generate dissatisfaction with established concepts, more precisely a feeling that their meaning is outdated and inadequate, prompting modifications and inventions as witnessed in the emergence of relatively new terms (such as globalisation, regionalisation, climate change) or the recovery of long-neglected concepts (such as empire, capitalism, race).
This article complements specialised engagements with particular concepts by offering a general and systematic discussion of heuristic frameworks, or approaches, which tackle the multifaceted nature of basic concepts within and across historical, theoretical, and political contexts. It provides an accessible overview of three prominent approaches called here ‘historical’, ‘scientific’, and ‘political(critical)’, which are discussed through the work of Reinhart Koselleck, Giovanni Sartori, and Michel Foucault, respectively. To my knowledge, no such comparison exists in the IR literature and together they provide a comprehensive picture of what concept analysis might entail. It is also worth noting that, whereas Guzzini directs attention to Sartori, the following discussion relies heavily on Koselleck. Rarely discussed in IR, 8 Koselleck is a leading figure of concept history and best-known for co-editing what is perhaps the most comprehensive ‘unfinished’ dictionary of concepts in existence, the Geschichtliche Grundbegriffe. 9 The article thus not only uses Koselleck as representative for the historical approach but also draws on his work, especially his notion of a ‘basic concept’, in preparing the ground for thinking about concept analysis more generally. With this in mind, the discussion proceeds in three steps: the following section introduces the notion of basic concepts, clusters and the issue of context, the main section outlines the three approaches, and the conclusion returns to the question of ontological theorising.
Approaching Concepts
The article’s underlying premise is that the very question of what a concept ‘is’ does not have a single answer. Concepts never just ‘are’, they are human creations and the form they take is not only internally complex but also varies significantly. And any effort to grasp this form is already an effort in conceptualisation. Loosely following Adi Ophir, 10 the very attempt of defining a concept involves discussing its appearance, at least if one approaches this task with a sufficient degree of reflexivity, and thus leads into concept analysis. So while this section will advance a particular kind of concept – the basic concept – and discuss the contextual configuration it is embedded in, it considers the ‘definitional effort’ and the task of concept analysis not two discrete steps but intertwined moves, which reveal the concept’s plural, contingent, and open form.
Concepts as Basic
As a general point of departure let us say that a concept is an abstract frame that helps generating knowledge about the world by organising, naming, and giving meaning to its features. In Max Weber’s words, concepts are our way of overcoming ‘the extensively and intensively infinite multiplicity’ of empirical reality. 11 They help us to grasp the world epistemologically and give us an ontology we can relate to. Indeed, as William Connolly notes, naming something offers more than a description, it also gives it a character. 12 Concepts enable us to communicate and research the world, also catering to the modern belief that once we can name something we not only know but also can control it.
That said, as an abstract heuristic device, a concept is not considered an accurate representation of reality/the world – regardless of the fundamental question of whether such representation is possible – but a mental image which meaningfully organises this reality/world, perceived through sensory experiences, in the mind. Concepts thus are often seen as taking the form of what Weber called an ‘ideal type’. The ‘ideal’ stands here for the abstract and general nature of the concept’s content that takes form by highlighting certain elements of a worldly phenomenon (or class of phenomena) in a stylised fashion, which do not exist in such stark and pure forms in empirical reality. 13 While some concepts like ‘chair’ or ‘table’ can be seen as manifested in one concrete material object, this is not the case for most concepts prominent in IR, such as ‘the state’, ‘anarchy’, ‘war’ or ‘globalisation’, which designate broad and complex phenomena whose material manifestations are plural, shifting, and incomplete – if they are accepted to exist in the first place.
This leads us to two accounts of how and where concepts are formed and can be said to exist. The first treats a concept as a cognitive property of the mind. It assumes that the concept has a pure abstract form that can be learned and becomes imprinted in our cognitive system, perhaps even leaves a physiological trace in the brain. Once this has been achieved, the concept is ‘perceived from the mind’s eye’. 14 The second account accepts that concepts have a cognitive function but sees them as part of a linguistic structure. That is, in contrast to the first account, where language is merely the medium through which the mental representation is acquired, it holds that concepts are created through and exist primarily in language. Rather than individual cognitive achievements, concepts are seen as social, inter-subjective constructs that gain meaning in what Wittgenstein called language games. 15
The second account has risen to prominence through the ‘linguistic turn’ in the humanities and the social sciences and has been deeply influential in concept analyses. In asking what language game(s) concept analysis is dealing with, we then may differentiate between two levels: (i) academic research, where concepts are used as analytical categories, as tools for abstract logical reasoning and heuristic devices guiding empirical study, with no (necessary) direct correspondence in socio-political discourse; (ii) everyday life, where concepts guide thought and action of individuals and collectives across all sectors of society, sometimes captured as ‘indigenous categories’, 16 or ‘categories of practice’. 17 This intersects with an important question regarding the focus of concept analysis: does it explore how concepts (as categories of analysis) are used as building blocks of academic theories, or does it analyse how concepts (as categories of practice) operate in society more widely? This question implies a separation between two scholarly endeavours: on the one hand, academics with a specialised theoretical language that use concepts as analytical instruments with little or no connection with the frames used in broader society; indeed, not even making an assumption about whether and how concepts matter in everyday life. On the other hand, an academic approach with a theory about concepts as socio-political forces, yet which is less concerned with abstract explanations and more with carving out how concepts operate in everyday life. This distinction makes sense both for theorists who prefer to work things out in logic and those who follow a more empiricist approach. And yet, when analysing concepts used for studying socio-political phenomena the distinction becomes problematic due to the simple fact that scholars of world politics are part of the world they are studying and their analytical categories are intertwined with what Connolly calls the ‘terms of political discourse’. 18
The overlap between academic and socio-political conceptual language and, hence, ontologies, is visible across the theoretical spectrum. As Guzzini reminds, classical realists adopted basic concepts such as the ‘balance of power’ from practical knowledge, that is, from a ‘sense of history and the experience of politics’. 19 In their view, practical knowledge served as both source and recipient for scholarly knowledge and, hence, was considered the only valid point of reference for IR’s theoretical vocabulary. While classical realists struggled with formulating a conception of theory that allowed maintaining that practical connection, even realists like Waltz who claim to be more ‘scientific’ rely on the same concepts and, thus, can only pretend to operate on a separate level. Liberal theorising grapples with the overlap in its signature thesis on ‘democratic peace’, where the academic discussion of this thesis clearly intersects with its ‘political biography’. 20 And disagreements about the validity of the ‘empirical fact’ that two democracies have never fought a war against each other partly rest on whether categorising political systems as ‘democracy’ and their interaction as ‘peace’ follows abstract indicators chosen by the analyst or the perceptions of political actors at the time. 21 At the other end of the spectrum, constructivists often theorise through the analysis of political discourse. A classic example is scholarship on securitisation, which problematises the treatment of security as an objective analytical category to explore how the concept is (successfully) employed by political actors and with what consequences. Yet their analyses still rely on a conceptualisation of security speech having a particular logic inscribed in its ‘grammar’ and scholars are debating the theoretical-political nature of their approach. 22 Critical theorists, feminist and postcolonial scholars have perhaps most clearly exposed the connection between concepts as categories of analysis and as categories of practice, warning that scholarship often becomes part of a political discourse. The problem has been highlighted, for instance, in research surrounding the concept of ‘identity’, and while some have urged using other concepts in its place this proves difficult even for those who try. 23
These examples suffice to illustrate that IR scholars are well aware of the fact that their key concepts operate in both academic and socio-political discourse. In the terminology used in this article, they work with what Koselleck calls ‘basic concepts’ [Grundbegriffe]. For Koselleck, a concept is basic if it plays a central role in our socio-political language, and so if we consider it a concept ‘we cannot do without’. In his words, ‘basic concepts combine manifold experiences and expectations in such a way that they become indispensable to any formulation of the most urgent issues of a given time’. 24 This quote points to two interrelated core features of basic concepts. First, they grasp, or refer to, fundamental features of a socio-political system. More specifically, they are leading terms [Leitbegriffe] of our vocabulary trying to categorise and grasp (links between) fundamental structures, processes, and events; they are keywords [Schlüsselwörter] and slogans [Schlagwörter] used by major social, economic, and political organisations and movements, and scholarly attempts to describe them; and they are core terms found in major theories and ideologies. 25 As such, basic concepts do not merely exist as specialised terms within academic circles but permeate public/popular discourse. They are, as Foucault would put it, ‘fundamental codes of a culture’. 26 Their ability to guide thought and action is due to, second, their temporal structure containing a stock of experiences and an aspirational outlook. That is, Koselleck argues, the ability of a basic concept to grasp key features of social relations is tied to the ‘experiential content’ (Erfahrungehalt) it has accumulated and to the ‘innovative expectations’ it raises. This feature of concepts as both backward and forward-looking conveys not only that concepts have a temporal dimension but also that they may allude to movement and contain a promise of progress, which hints at their normative content. 27
I contend that most, if not all prominent concepts used by IR theorists fall into the category of basic concepts. 28 Some research projects capitalise on this, yet for many theorists the link between academic and socio-political discourse is uncomfortable as it is not clear how to deal with the connection and its possibly circular dynamic. The easy way out is to either ignore the link or conflate the two discourses, for instance by equating theory with ideology. Yet leaving those two (quite popular) moves aside, thinking about the connection is challenging also because it reveals the complexity and openness of a concept. After all, to say that a concept is basic does not imply consensus on its meaning, if anything the opposite is the case. Basic concepts tend to be contested precisely because they are basic and open to interpretation, which prompts different actors to try to claim a monopoly on its meaning. 29 With contestation occurring on the theoretical as well as the socio-political level, the challenging scope of the task becomes apparent.
Grounded in a commitment to reflexivity, ontological theorising cannot get around addressing a concept’s ‘basicness’. More precisely, concept analysis – the first leg of ontological theorising – as a systematic study of the formation and performance of core concepts involves exploring the overlap between academic and socio-political discourse, as well as strategies for dealing with it. Before outlining different avenues along which this may take place, it is necessary to take a closer look at the configuration through which a concept’s meaning is created by reading it as a cluster and embedded in context(s).
Concepts as Clusters and in Context(s)
Concepts tend to be attached to words, although not exclusively and not necessarily always to the same word. As such, a concept is more than a word. Whereas the meaning of a word points to one particular thing, a concept catches and bundles multiple elements, aspects and experiences and relates them to each other. Similar to Jacques Lacan’s ‘button tie’ [point de capiton] through which an otherwise empty signifier (temporarily) fixes and knots together meanings, a concept is an idea that connects a number of ‘things’ in an ‘ideal type’ configuration. Connolly thus speaks of ‘cluster concepts’. 30
To make sense of this configuration, or cluster, it is common to draw on structural linguistics and view a concept situated in a ‘semantic field’ formed out of links to other concepts, creating a particular web of concepts [Begriffsnetz]. 31 In this web, three kinds of links are prevalent: to supporting concepts, which are integral to the meaning of our concept (sovereignty for the state); to cognate concepts with similar meanings, or whose meanings correspond with each other and bear what Wittgenstein called a family resemblance (football and basketball are both games, to use his famous example); and to contrasting concepts that are opposite in meaning, sometimes even taking the form of counter-concepts (as in reactionary-revolutionary), which relate to and (in)form each other through a dialectic. These links forming a conceptual web do not need to be grounded in logic but can be habitual, sentimental or normative, and thus seemingly arbitrary in character. And they become particularly interesting when presented as (supporting) a causal relationship, as in the case of the aforementioned democratic peace thesis. Here ‘democracy’ is linked to ‘republics’ to ground the thesis in the philosophical authority of Kant, it contains the concept of ‘freedom’ and is tied to ‘peace’ arguably because both freedom and peace are considered desirable/good, and the thesis relies significantly on the concept of ‘non-democracy’ and, especially, ‘war’ to stabilise its meaning. 32 Yet not only are these supporting concepts embedded in semantic fields of their own, the echoes of which influence how we read the basic concept, or the particular causal argument at hand. As Connolly argues, it also is problematic to place concepts within a cluster in a causal relationship as this converts a conceptual connection into a purely analytical one. 33
A basic concept is not an umbrella term. Its meaning is not arrived at by simply ‘adding up’ the constituting elements, not least because it is the concept which enables us to organise and make the connections between those ‘elements’. In the words of Koselleck, a concept ‘is not simply indicative of the relations which it covers; it is also a factor within them’.
34
In his programmatic introduction to the Geschichtliche Grundbegriffe, Koselleck uses the example of the state to illustrate this crucial point:
‘Think of all the elements that enter into the word ‘state’ so as to turn it into a concept: domination, territorial sovereignty, citizenship, legislation, adjudication, administration, taxation, and military force, to just name the common ones [and one could add government, police, and borders]. All of them, with their own complex subject matter and terminology, are incorporated into the word state, which then becomes elevated to the status of a concept. Concepts are thus concentrations of many semantic contents’.
35
This internal richness raises the issue of definition. On the one hand, it seems difficult to define something that contains many elements and meanings. And, indeed, if we understand the purpose of a definition to be fixing meaning, concepts cannot be defined; as Nietzsche famously put it ‘only that which has no history can be defined’. 36 On the other hand, it is through the definition that we move beyond the word to express the concept’s constitutive elements and indicate how they relate, that is, how the cluster is formed. Such a definition does not fix meaning. Staying with the above example, consider Weber’s well-known definition of the state as ‘that community which holds the monopoly over the legitimate use of physical force within a particular area [Gebiet]’. 37 While this is a seemingly specific definition that appears to capture important elements, it is still vague and requires of us to clarify the meaning of terms such as ‘community’ or ‘legitimacy’. Of course, Weber’s definition is still partial and misses elements that others might consider integral to the state. So the selective nature of a definition constrains meaning, yet even a seemingly clear definition will be ambiguous, or sufficiently vague, and allows for multiple readings. Thus, as Koselleck notes, concepts cannot be given definite meaning, they can only be interpreted. 38
To understand why and how certain elements and relations are chosen and why a concept is interpreted in a particular way, we need to pay attention to ‘context’ – the frame, environment or field within which a concept is embedded. More precisely, we need to pay attention to the relationship between concept and context. That is, accounting for concepts in context not only involves asking about how the former is placed in the latter and how this imbues the concept with meaning; it also involves having an eye on how the concept is situated in a particular context and shapes (our understanding of) the latter. Thus, we might say concept analysis involves tracing how a basic concept is formed and how it performs within a context. But how to conceive of ‘context’ and a concept’s location within it? Even the specialised literature tends to give only a vague answer and often prefers not to define ‘context’ at all. 39 For the present purpose, it is clear that we need to move beyond semantic fields. To capture the formation and performance of a basic concept, we need to account for the ways it is used and understood across different sectors of society, carve out its expression in extra-linguistic forms, and understand its historical depth, which often provides the concept with interpretative authority, but also is a major source of its complexity. Given this, it seems sensible to speak not about context (in the singular) but about different contextual layers, or dimensions. For analytical purposes, at least, ‘context’ can be broken down into four such dimensions: political, temporal, material, and theoretical.
The socio-political dimension asks for attention to the formation and performance of a concept in (international) society and within a political system: to trace how a concept is used and its meaning manifested by political actors, its diffusion throughout society/the system and the different understandings and usages seen in different parts of that society/system. It also directs attention to how a concept shapes society and how its meaning becomes a subject of political contestation. The temporal dimension involves studying the historicity of a concept and how its meaning or content is formed and evolves over time. It views concepts as embedded within a particular historical moment and/or particular structures stretching over time into the future. Conversely, analysts may ask how a concept shapes our understanding of time, namely how it directs our temporal orientation and privileges certain readings of past and future. The material dimension asks us to look at the material space(s) and bodies in which the concept is used and manifests itself, including what happens to meaning when the concept travels from one material context to another. Equally, it asks us to be sensitive about how concepts organise and shape (our awareness of) material spaces and bodies. Finally, the theoretical dimension directs attention to how concepts are situated in a broader ideational framework, or narrative. It requires exploring the role a concept plays in a particular theoretical ontology and argumentative logic and how it acquires meaning through this role. Koselleck hints at ‘theories included in [basic] concepts’. 40 Keeping in mind the reciprocal relationship between concept and context, we also need to ask what a concept does for/to a theory, including how and why theorists build their theories around, or through, particular concepts. In other words, we must understand how a theoretical narrative weaves concepts together into a particular web and is itself constituted through this web.
While questions of theorising direct attention to the last of these dimensions, all four hang together and a comprehensive analysis would explore their connections. What is more, it makes sense to conceive of contexts as dynamic rather than static and, given their interrelated nature, the same goes for concepts. Consequently, exploring a concept’s clustered nature within context poses a formidable analytical challenge. It is not the ambition of this article to provide such an analysis, but to present three approaches that emphasise different dimensions and trace the formation and performance of basic concepts in different ways.
Three Approaches to Concept Analysis
This section outlines three approaches to concept analysis through the work of Koselleck, Sartori, and Foucault, respectively. 41 Their differences rest to an important degree on their respective motivations for engaging basic concepts in the first place, that is, on what they see as the purpose of concept analysis. The first approach, represented by Koselleck, is motivated by the goal to treat a concept historically by tracing its evolution over time and, thereby, excavate diverse readings. The second approach, represented by Sartori, aims at specifying and grounding the concept empirically in travels across space, motivated by the attempt to maintain the idea of a universal theory anchored in empirical data. And the third approach, represented by Foucault, seeks to critically expose the political formation and performance of a concept and the discourse it is part of. The discussion starts with the historical approach for the same reason it is highlighted by Guzzini: while perhaps the least theoretical, it is the approach that most comprehensively alerts us to the complexity and openness of basic concepts and their grounding in practical knowledge. It thus can be seen as a baseline for the other two approaches, which I suggest mark two poles for ways of dealing with a concept’s ‘basicness’: at one end the ambition (and struggle) to contain it through scientific method, at the other end a commitment to embrace it through genealogy as a form of political critique.
The Historical Approach
As the name suggests, the historical approach foregrounds the temporal context to provide a better understanding of a concept’s place in and evolution throughout ‘history’. Starting from the view that concepts are ‘in motion’ it traces how a particular concept is understood and employed differently throughout history, how it evolved, and how we arrive at the meaning(s) we employ today. As such, the motivation is not simply historical curiosity and to illuminate how concepts were used in the past. It also generates a better understanding of how concepts ‘push us to think along certain lines, thus enabling us to conceive of how to act on alternative and less constraining definitions of our situation’. 42 Historians of ideas, philosophers of history, as well as political and literary theorists have developed comprehensive and rich studies in this regard, which can be subsumed under the label of ‘history of concepts’ or ‘concept history’ [Begriffsgeschichte]. Koselleck is a leading figure in this field and much of the preceding discussion has already drawn on his insights. 43 Here I outline some key aspects of his approach and tensions within it.
Analysing the history of a basic concept goes far beyond etymology, the tracing of a word’s evolution from its alleged ‘roots’ or ‘origins’ to current usage. It deals with the evolution of conceptual language and thought in the context of historical experience and is especially interested in recording and explaining changes in a concept’s meaning. Crudely put, one could say that this approach tries to understand, and tell a story about, the ‘life’ of a basic concept by looking at four phenomena: (1) Concept invention (emergence): how a new concept establishes itself in a particular historical context. (2) Concept fixation (reification): how a particular meaning becomes hegemonic and gains ‘common sense’ status. (3) Concept transformation (modification): how a term takes on a new meaning, or meanings. And, finally, (4) concept disappearance: how a concept ceases to be used and drops from our vocabulary. Concept history not only describes such dynamics but also seeks to explain why concepts emerge, are modified, or disappear.
The approach advanced by Koselleck is especially interested in exploring how conceptual changes correlated with the discontinuity of political, social, and economic structures, and how and why certain experiences and structural changes were understood in particular ways. 44 Thus, it analyses convergence and divergence between ‘real’ history and how ‘history’ was framed by contemporaries, with a particular focus on how divergences generated new concepts or changed the meaning of existing ones. Loosely following Willibald Steinmetz, conceptual change occurs for three reasons: first, the loss of plausibility of existing concepts which cannot grasp new or unexpected events and experiences; second, an irritation prompted by adopting words from another language or switching between socio-linguistic contexts; and third, the strategic usefulness of a new concept through its ability to absorb diverse perspectives and its intuitive resonance (its metaphorical quality, for instance). 45 Moreover, the approach explores the configuration of the societies and historical periods in which concepts emerge or are transformed. Thus, it treats (the analysis of) conceptual change as a window into understanding societal transformations and, as such, historical change. In line with Koselleck’s note that concepts are not only indicators of but also factors in change, which implies that concepts do something, the historical approach also considers the concept’s representational performance. That is, it asks how a concept influences decisions, actions and social relations during particular historical moments and, ultimately, shapes the course of history.
Analysing all this requires sensitivity to how concepts and socio-political structures implicate each other, in other words, how concepts become basic. It also grapples with fundamental questions regarding the possibilities and limits of historical representation, often expressed as tensions. One tension concerns its ambiguity regarding the relationship between language and historical reality: despite emphasising the primacy of language, concept history does not take the radical position that there is nothing outside language. If that was so, it could not investigate the relationship between concepts and socio-political structures, including material context, in a way that lends the latter some sort of causal force. While this epistemological dualism may strike one as problematic, Hans Ulrich Gumbrecht suggests that remaining in a state of ‘indecision’ was a conscious move that enabled contributors of the Geschichtliche Grundbegriffe to use this ambiguity productively. 46 Another issue is that a focus on linguistic and socio-political structures neglects agency. After all, concepts do not act (on their own); looking at a ‘concept in action’ requires looking at who is using it. So to understand their historical evolution we must also ask what historical actors do with concepts, why and how they assign and manipulate meaning. Tracing the use of concepts among a variety of agents in a given society, beyond a few elite figures, is not an easy task and requires a fine-grained analysis that is difficult to do when taking a macro-historical perspective. 47
This leads to the methodological challenge of tracing patterns of political language surrounding a particular concept both in a specific place and time (synchronic analysis) and across space and time (diachronic analysis). For Koselleck the two modes of analysis are inseparable, and for him it is the alternation between the two that reveals that a concept can never be reduced to a single meaning. 48 The decision about how to balance and combine diachronic and synchronic analysis is partly informed by the kind of change one wants to look at: whether one seeks to trace ‘concepts in motion’ from a macro-historical perspective, which analyses the evolution of a concept across a long period of time, or to reconstruct ‘concepts in action’ in a micro-diachronic analysis, which delves into the complexity of how a concept performs and changes in a temporally and spatially confined setting.
Relating basic concepts that were developed and used in different historical periods is often presented as a technical challenge. However, the deeper issue is not technical, as illustrated in J.G.A. Pocock’s scepticism about tracing the evolution of a concept through time. In Pocock’s view, analysing a particular historical context should take priority, thereby placing ‘premium upon the synchronic’. 49 This is in line with a mode of historical analysis attributed to the Cambridge School, which highlights the importance of thinking carefully about how to conceive of and capture historical context and what might be adequate sources for reconstructing a concept’s meaning. 50 Yet Pocock, who was influenced by Thomas Kuhn’s notion of paradigm, 51 goes further in suggesting that (the meaning of) a concept is always bound to the discursive and, especially, ideational structures governing a society at a particular point in time and only really exists within them. Thus, the issue is not only the methodological complexity of tracing a distinct conceptual cluster in a multi-layered and dynamic context. Rather, Pocock doubts the very notion of a concept having a history whose meaning can (or needs to) be understood across political contexts and, hence, time. 52 This scepticism appears to be grounded in the fact that identifying a basic concept is not simply an empirical undertaking but also a theoretical act. The reluctance to grant this role and, hence, the ability to identify a basic concept across time to the analyst reveals a discomfort among some historians not simply with macro-history but with theory.
The Scientific Approach
The scientific approach is driven by the modern ambition to demystify the world and sees concepts as methodological tools for measuring, explaining, and predicting that world. Labelling this approach ‘scientific’ does not imply that the historical approach is ‘unscientific’, if science is broadly understood as sophisticated, systematic, and methodologically guided analysis for the purpose of producing knowledge. Yet whereas the historical approach aims at understanding the use and change of concepts in/over time, the scientific approach outlined here engages in what Herbert Blumer calls ‘precision endeavors’, concerned with improving our conceptual toolkit. Guided by the conviction that only ‘clear concepts’ bring scholarly knowledge ‘into close and self-correcting relations’ with empirical reality, 53 the scientist seeks to sharpen their tools by refining and revising concepts. However, the aim to develop better, in the sense of both more accurate and more useful concepts that ‘work’ in different geographical and cultural locations faces a tension. By treating concepts as analytical categories embedded in universally valid theories that can travel across space, the approach maintains the separateness of the theoretical dimension and, thus, seeks to distance its concepts from practical knowledge. At the same time, the approach holds that the meaning of concepts is substantiated and, by extension, theories are validated through empirical research in particular places. In trying to deal with this tension, namely to combine the universal and the particular, the (reflexive) scientist also grapples with a concept’s ‘basicness’.
An influential figure in Political Science representing this approach is Giovanni Sartori, who recognised that social scientists use concepts not only as elements of a theoretical system but also, indeed primarily, for collecting empirical data and, hence, treat them as ‘data containers’. 54 Yet his training in political philosophy also made Sartori sensitive to the fact that concepts have a history and, hence, come laden with meaning. This led him to argue that researchers need to think carefully about how to employ – apply – concepts empirically, especially in comparative research. While accepting that the meaning content of a concept is built up in the process of ‘measuring’, that is, gathering data in particular places, Sartori worried that concepts are used across geographical and cultural contexts without understanding a concept’s history and the field it is made to work in. 55 He warned that failure to recognise this richness – and, effectively, the ‘basicness’ of a concept – produces ‘conceptual confusion’. 56 To tackle such confusion, Sartori argued, analysts must address ‘meaning before measurement’ and work on reconciling universal and particular meanings that would allow concepts to function as ‘empirical universals’. 57
Essentially, this approach takes a careful look at how a concept ‘travels’ across space and what happens to it in the process. It asks researchers to pay attention to (i) the concept’s ‘extension’, the process in which a concept widens and shifts its boundaries to include more and/or different elements, thus becoming broader and more complex in meaning, and (ii) the concept’s ‘intension’, the process which zooms ‘inwards’ and highlights its core elements, thus specifying its meaning. Scholars offering strategies to navigate this dynamic often do so in an overly technical discussion about logic, semantics and methods. 58 Here it suffices to say that central to this is the process by which the concept, in its clustered form, travels from theoretical reasoning to empirical investigation and back. That is, the approach seeks to understand how a concept is altered in the process of application, or operationalisation, when it moves from the abstract level, where it carries a general meaning, to the concrete empirical level, where it functions as a ‘data container’. To use Sartori’s well-known metaphor, it traces how a concept is modified when the researcher climbs down ‘the ladder of abstraction’. 59 Sartori tries to capture the modification through the notions of ‘high level’ and ‘low level’ concepts, yet the more basic point is that this ‘climbing’ exercise, including the upward move of placing empirical findings within a theoretical argument, is a creative task of translation, adaptation, and perhaps even transformation of meaning. 60
That said, driven by the aim to maintain universally valid theories the scientific approach seeks to manage and minimise ‘distortions’ by devising techniques and rules that allow concepts to travel across space without disappearing into the particular. Not surprisingly, Sartori’s solution is to call for a ‘minimal definition’ that fixates a concept’s ‘basic structure’ and allows capturing the ‘core attributes’ of a phenomenon across contexts. 61 There are two interrelated problems with this. First, the scientist’s ambition to construct ever more accurate concepts sees vagueness and ambiguity as ‘defects’ and a ‘basic deficiency’. 62 Yet this goes against the point made earlier that concepts are inherently and necessarily vague and ambiguous, and it ignores that this is what often makes them useful to scholars and practitioners. After all, theories have a long shelf life not least because their key concepts remain vague. Similarly, in diplomatic negotiations compromise and agreements tend to rely on sufficiently vague concepts that can be interpreted differently by each side.
Second, the notion of universally valid ‘minimal definitions’ is confronted with scholarly disputes about the meaning of basic concepts. As W.B. Gallie famously pointed out, scholars using the ‘same’ concept in different ways often are unable to agree which meaning is ‘better’ or ‘more accurate’, rendering the concept ‘essentially contested’. Gallie noted that ‘endless disputes’ about the proper use of basic concepts such as ‘democracy’ or ‘justice’ are not the result of sloppy appropriations or misreadings but, rather, are ‘perfectly genuine: although not resolvable by argument of any kind, [they] are nevertheless sustained by perfectly respectable arguments and evidence’. 63 For the scientist this is puzzling, especially as a dispute suggests that there must be agreement that both sides are actually talking about the same concept. Gallie argues that this is the case when both derive their reading from ‘a process of imitation and adaptation from an exemplar’, that is, when both sides use the same source from which their respective reading of the concept flows. 64 Yet because the ‘exemplar’ – like all basic concepts – is internally complex and open/ambiguous, theorists are able to focus on different elements, interpret the same configuration differently and develope different adaptations. Thus, essential contestedness arises out of a combination of the complex internal logic/configuration of a concept and its evolution in different, perhaps incompatible, contexts.
For the issue at hand, this suggests that in many cases the ‘same’ concept cannot be transferred between theoretical frameworks as its meaning is embedded in different conceptual webs, logics, and incompatible epistemological stances, and is informed by different practical knowledges and normative commitments. Consequently, the shared exemplar does not have the authority to resolve the dispute as both sides can claim to ‘own’ it, and neither is able to persuade the other that its claim to ownership is ill-founded. Although this limits the ambition to build ever more accurate concepts, the scientific approach still invites investigations into how different readings are developed out of and projected back into a historic ‘exemplar’ and to establish a sensible typology of different readings without favouring one over another. In doing so, it points to the crucial question of whether (a) the different types are compatible and can be integrated into a single research design, that is whether they can be combined, or are claimed to be combinable, into one master concept that cuts across different theories (what might be called the ‘Sartori’ position), or whether (b) the types are essentially contested and, thus, incompatible and cannot be used to link different theories together, let alone be combined under an overarching theoretical framework (the ‘Gallie’ position).
The Political (Critical) Approach
The third approach starts from the dictum that knowledge is power. It acknowledges that basic concepts fulfil a cognitive function and are central parts of knowledge production in modernity. However, in contrast to the scientist it views this stabilising function and the attempt to distinguish between analytical and practical categories with suspicion. Influenced by a post-modern stance, this approach highlights that the order created by concepts is artificial, blocking out the complexity of the world and upholding certain power structures that benefit some and disadvantage others. Thus, the approach has a dual aim: first, to explore how concepts form and become reified through their use across society, and to highlight that concept formation and performances are implicated with structures of power. In doing so, it overlaps with concept history and it makes the scientific approach part its object of analysis. 65 Second, the approach seeks to shine light on blind spots and silences in the dominant understanding/use of a basic concept by revealing and reactivating ‘subjugated knowledges’. 66 It does so to disrupt and challenge reified meanings and underlying power structures to open the door for alternative conceptions and, ultimately, social change. As such, this approach takes a critical stance and explicitly understands concept analysis as an engagement with politics, indeed as political in motivation.
To understand reification processes, the political(critical) approach investigates how particular kinds of knowledge are produced, that is, how concepts are used and perform in society and with what effect. It highlights that, by giving meaning to ‘things’, concepts do not just make these things intelligible, they actually make things, that is, in the words of Michel Foucault they ‘systematically form the objects of which they speak’. 67 As such, this approach focuses on the productive power of basic concepts as not merely guiding but constituting thinking and action, identities, or subjectivities, on both elite and the subaltern levels, as well as altering material realities. One important aspect of this analysis is to explore how basic concepts and the epistemes, that is, the concept webs, theories and broader discourses surrounding them become institutionalised. Investigating the process of institutionalisation involves tracing how the ideational content of the concept becomes entwined with material reality. In other words, it studies a concept’s material manifestation, including the aesthetic expression of a concept in architectural designs or ‘hard’ forms of infrastructure. Yet it also can take more intangible forms if ‘institution’ is understood not only to encompass formal structures with a concrete physical presence, but also informal practices which are not tied to a particular place but are diffused throughout society in a seemingly uncoordinated manner. Analysing how basic concepts exercise power thus largely depends on how we think productive power works: does it unfold directly and overtly, or in subtle and hidden ways? Is the power of concepts orchestrated by political actors, or do they take on a life of their own without being controlled by anyone? Are concepts embedded in a dominant ideology imposed from the top (as in Gramscian hegemony), or do they emerge out of and are reproduced by everyday practices (as in Foucauldian governmentality)? Analysts need to consider all these angles and most likely a combination of the two, as exemplified in Foucault’s analysis of ‘crime’ and ‘madness’. 68 In any case, the political approach faces a conundrum: the attempt to analyse the political performance of a basic concept – the production and effect of its meaning – relies on an understanding of another basic concept, namely power.
As noted, the critical/political approach does not stop at exploring the reification and reproduction of basic concepts by unmasking power configurations. It also seeks to weaken these configurations and act as a critical voice against the temptation to find a singular meaning and accept a ‘common sense’. It does so by highlighting techniques of resistance, subversion, and contestation of meaning and, thus, showing that meaning is (or should be) a matter of politics. Indeed, as the name suggests, the political approach actively participates in and seeks to advance contestation and regards concept analysis as a mode of resistance. 69 One prominent way to accomplish this is genealogy, which overlaps with concept history in exploring the evolution of a concept within socio-political structures. Yet genealogy is distinct from concept history in that its aim is not tracing change, but showing how a concept is assembled into something taken for granted while, simultaneously, showing that any notion of a concept having a unified or coherent meaning is an illusion. It brings to the fore the historical contingency of stabilised knowledge through the dual move of critiquing (deconstructing) the notion of a coherent historical narrative about a concept and bringing to the foreground (reconstructing) the marginalised and forgotten aspects. As Connolly puts it, genealogy is the ‘reversal of the project of interpretation’. 70
The genealogical approach rejects the existence of a single source, or ‘origin’, to which a given concept can be traced. Instead, Foucault argues, following Nietzsche, if the genealogist ‘listens to history he finds…the secret that they have no essence or that their essence was fabricated in a piecemeal fashion from alien forms. What is found at the historical beginning of things is not the inviolable identity of their origin…it is disparity’. 71 Rather than trying to identify an unbroken continuity in the historical life of a concept, this approach asks the analyst to ‘identify the accidents, the minute deviations…the errors, the false appraisals, and the faulty calculations’ that underpin the life of a concept. 72 In the words of Jens Bartelson, who employed the approach for a study of sovereignty, this involves writing ‘a history of the unthought parts of our political understanding [of sovereignty]…a history of the epistemic discontinuities, conceptual reversals and changing metaphors that…animate the discourse of sovereignty’. 73 Such a study of discontinuity and contingency – of how things have become accidently attached to the concept and inform its shape – aims at revealing aspects of a concept’s life that have been glossed over by conventional accounts and become invisible. It seeks to carve out meanings/uses that are ignored because they do not fit established discourses, are considered irrelevant or invalid and, hence, are marginalised, squeezed out of sight. The resulting picture is not a concept with a unifying shape but fragments of meaning.
While for Foucault these subjugated meanings are expressed in local, everyday practices permeating society, they may also be found in the contradictions and tensions of dominant discourses, or in elite voices that have become forgotten. Either way, genealogy asks for attention to detail and explores how meaning is tied to ‘the most uncompromising places, in what we feel is without history – in sentiments, love, conscience, instincts’. 74 The commitment to carve out aspects that do not fit modern presentations of a concept and are politically inconvenient thus requires a sensitive mind and willingness to go against the grain. And it puts reflexivity to the forefront. While other approaches can silently fall back into claiming to pursue an objective analysis, the political(critical) approach faces the tension between the ambition to destabilise taken for granted meaning without proposing to present more comprehensive knowledge about a concept and, hence, a ‘better’ understanding of its meaning(s). This is a balancing act, often mitigated through the style of analysis and its presentation, namely by avoiding conventional logics and rigid systematicism in favour of a playful, improvised and disruptive mode which consciously avoids offering a simple and coherent account.
Conclusion: Implications for Theorising
Responding to Guzzini’s call for a wider understanding of concept analysis as a mode of ‘ontological theorizing’, this article offered a systematic overview of what this might entail. The article started out by suggesting that the key concepts employed by IR scholars are ‘basic’, that is, permeating both scholarly and socio-political discourse and central to both. This was followed by a discussion of how unpacking such concepts requires seeing them as clusters, or webs, embedded in different contextual dimensions. The article then sketched three prominent approaches to concept analysis drawing on the work of Koselleck, Sartori, and Foucault. 75 These approaches do not present polished roadmaps but are dynamic projects guided by different views about why and how to engage concepts, including how to deal with the ‘basicness’ of a concept: for the historical approach, this quality is the very reason for tracing a concept through time and exploring its role in societal change, although the approach remains vulnerable to the question of whether ‘basicness’ is determined primarily by the historical context or by the analyst’s own situatedness. While the scientific approach tries to keep the theoretical sphere separate and to block out the basic quality of a concept, it encounters the overlap between scholarly and practical knowledge in the process of operationalisation in different cultural spaces. Finally, the political(critical) approach takes the issue of ‘basicness’ head-on by highlighting the intertwined nature of theoretical and socio-political discourses and the power dynamics underpinning concept formation and performance.
Despite having distinct aims and features, the approaches also offer points of connection. Most obviously, all three demonstrate, each in their own way, the open and plural nature of basic concepts: by highlighting historical change and diversity; by climbing the ‘ladder of abstraction’ and establishing typologies; and by showing their contingent form and recovering silenced readings. Thus, these approaches present us not with a concept that exists in singular form but, instead, with conceptions. 76 The form these conceptions take varies, of course: a dynamic concept that evolves throughout history; a concept that takes on different empirical forms yet still maintains an abstract unity; and a concept that is artificially reified yet at closer inspection reveals nothing but fragments. Still, despite their differences in emphasis and angle, all three approaches demand an analytical effort that oscillates between deconstruction and reconstruction, and it is this pendulum-like movement that shows the complexity, contingency and openness of basic concepts and that gives their analysis its unsettling and creative character.
Here, then, we return to the issue of theorising. To recall, the article began with the premise that concept analysis is the first leg of ontological theorising, not a mode of theorising in and of itself. Of course, the work of Koselleck, Sartori, and Foucault is infused with philosophical debates and all three are recognised as important theorists who think little of a strict separation between empirical research and theorising. Tracing the evolution of concepts, their use and performance may strike us as a largely ‘empirical’ undertaking, but it has an important meta-theoretical underbelly. Not only does identifying a (basic) concept, its clustered nature and contextual dimensions involve a significant theoretical effort. A reflexive engagement with basic concepts also touches on fundamental questions surrounding the meaning and possibility of doing historical/scientific/political analysis and concerning suitable methods for undertaking the required, or desired, deconstructive and reconstructive moves.
Recognising all this, we still need to ask more directly how concept analysis advances our understanding of theory and influences the way we think about and ‘do’ theory, that is, how it affects the practice of theorising. So let me conclude with a few concrete thoughts which bring together some of the points from the discussion. Most obviously, all three approaches make us better theorists because they train us to carefully reflect on core assumptions and building blocks of theoretical frameworks. Tracing the ways a basic concept is embedded in a particular theory, how it relates to other concepts and how these relations form a conceptual web on which a theoretical argument rests, contributes to our understanding of how that theory is built. And having a better picture of a theory’s internal composition makes it easier to see its logical strengths and limitations. One implication of the historical approach is that it invites, if not requires, historicising this picture. Koselleck’s angle asks for tracing the historical evolution of theories via their core concepts, following the logic that if the building blocks change, the theoretical house takes on a new form as well. Slightly different, the Cambridge School angle would hold that, through their concepts, theories are bound to a particular historical environment, that is, they belong to and only make sense in a particular historical context. Either way, the approach has implications for how we understand a theoretical ‘tradition’, 77 and certainly makes it difficult, if not impossible, to uphold the notion of theory as a stable construct over time. This may be disconcerting to traditionalists, yet seeing different meanings behind a shared term and understanding their historical trajectories enables both more targeted conversations and recognising the limits of such conversations. Here, the question of whether typologies can be connected through a minimal definition or are essentially contested, as debated in the scientific approach, has crucial implications for scholarly communication and collaboration: it either means that a concept can serve as a bridge for scholars working with different theoretical frameworks, enabling fruitful exchange, as has been suggested for the concepts of power 78 and practice, 79 or it can deceive theorists into believing such exchange is taking place when in fact they are talking past each other. 80
More importantly, awareness of multiple meanings and, indeed, of the instability of conceptual meaning raises the question why we employ a concept in a particular way and what the consequences are, and it opens the door to formulating an alternative. Here the creative role and responsibility of the theorist comes to the foreground. If anything, concept analysis highlights that theorists are not passive consumers and distributors of thought, but active minds who (can) make a choice about which reading of a concept (not) to adopt. The scientific approach alerts us that this choice is not made once and for all at the outset of a research project, but made continuously throughout. Against the sterile view of theories being ‘tested’ against empirical evidence, it directs attention to the process in which theories are transformed when coming in contact with the real world, i.e. when their key concepts are substantiated empirically. This reflexive move is further exploited by the historical and, obviously, the political(critical) approaches, which function as an antidote to theoretical reification and (should) guard against ‘the imperialism of categories’, to borrow Susanne Hoeber Rudolph’s phrase, by embedding a critical element within theorising. 81 Genealogy, especially, can be used to reveal the narrow configuration of a theory that is complicit in reifying conceptual meaning and to destabilise this theory by fragmenting its basic concepts. By blurring the line between theoretical discussions and ‘conceptual politics’, 82 it asks us to explore and critique a theory’s political performance. And by exposing the idea of a stable concept/theory as not only a convenient illusion but a political project, it also reminds us that a concept analysis which moves us beyond this idea is also a political move.
This leaves us with the question of whether and how concept analysis, including the critical variant, can support the process of theory building which, in my view, should be a central aspect of ontological theorising. Can the fragmentation of theory, the exposure of its contingent nature and the reconstruction of lost meanings be used as a productive stimulant for constructing (another) theory? Can a constant (re)writing, expanding, and updating of our conceptual dictionary still provide the sort of building blocks that theorists are looking for? The crude answer, I suggest, is ‘no’, but ‘yes’. If concepts are to perform the modern function of bringing order and reducing complexity so that theorists can reduce them to discrete variables, then concept analysis undermines this function. Once we go about unpacking concepts and show their complexity and openness, their contingent and context-bound meaning, to the modern mind these conceptions look more like a can of worms than building blocks. Rather than constituting theories, concept analysis seems more effective in destabilising and unraveling them. That said, it would be wrong to conclude that concept analysis makes theorising impossible. It just requires departing from the notion of theory as a timeless abstraction that can be tested empirically and continuously improved into a universal form of knowledge. Instead, it prompts us to conceive of theorising as an unstable and unfinished process that builds on conceptions, and to see sensitivity to the contextual and contingent nature of concepts not as a problem but as a resource for creative theorising. Acknowledging that our theories and their core concepts are unstable and our reading of them context-dependent may be difficult for some IR scholars, not least because this limits their explanatory reach. Yet such an acknowledgment does not discredit theorising as a reflexive process of exploring and constructing abstract arguments through basic concepts.
The bottom line is that the deconstructive and reconstructive moves intrinsic to concept analyses free space for thinking differently and devising alternative meanings and, thereby, enable theory building. They encourage scholars to play with definitions and change the meaning of a concept central to a particular theory, to revive hidden concepts that are integral to a theory yet had been silenced, or to introduce new concepts into our theoretical discussions. Such creative endeavours have always taken place in IR and they keep the field alive. They might lead to the formulation of a new theory, but they may also reveal that the most enduring IR theories are open and dynamic configurations with unstable concepts in the first place. In the end, when contemplating their approach to and choice of concept(s), theorists should keep in mind Connolly’s point that welcoming conceptual pluralism is not incompatible with the commitment to appropriate and advance a particular reading that suits our situation and agenda. 83 After all, once we recognise that basic concepts are not simply analytical tools but forces that organise social life, we may well take the value position that one reading is better than another.
Footnotes
Acknowledgements
For helpful comments on previous versions of this article, I would like to thank Jens Bartelson, Philippe Bourbeau, Christian Bueger, Stefano Guzzini, Annette Freyberg-Inan, Ned Lebow, Aaron McKeil, Evgeny Roshchin, audiences at a seminar in Cambridge (2014) and ISA conferences in New Orleans (2015) and Atlanta (2016), as well as the Millennium editors and editorial board. I also benefitted from thoughtful critiques by anonymous reviewers. This article is accompanied by an edited volume (Concepts in World Politics, Sage, 2016) and I am grateful to all participants in this project for insightful discussions on various occasions.
Funding
Research for this article benefited from a Venture Workshop Grant by the International Studies Association (ISA).
1.
Stefano Guzzini, ‘The Ends of International Relations Theory: Stages of Reflexivity and Modes of Theorizing’, European Journal of International Relations 19, no. 3 (2013): 521–41.
2.
Ibid., 534.
3.
Ibid., 535.
4.
Guzzini, ‘The Ends of International Relations Theory’, 534–37.
5.
Stefano Guzzini, ‘Structural Power: The Limits of Neorealist Power Analysis’, International Organization 47, no. 3 (1993): 443–78; Stefano Guzzini, Power, Realism and Constructivism (London: Routledge, 2013); Stefano Guzzini, ‘Power’, in Concepts in World Politics, ed. Felix Berenskoetter (London: Sage Publications, 2016).
6.
For an exception see Hedley Bull, The Anarchical Society (Basingstoke: Palgrave, 2002 [1977]).
7.
One aspect of this is the critique of Eurocentrism and the corresponding emergence of/turn to/search for ‘non-Western’ voices, see Arlene B. Tickner and David L. Blaney, eds., Thinking International Relations Differently (London: Routledge, 2012); John M. Hobson, The Eurocentric Conception of World Politics (Cambridge: Cambridge University Press, 2012).
8.
For an exception, see H. Jordheim and Iver B. Neumann, ‘Empire, Imperialism and Conceptual History’, Journal of International Relations and Development 14, no. 2 (2011): 153–85.
9.
The full title (translated) is Basic Concepts in History: A Historical Dictionary of Political and Social Language in Germany. See Reinhart Koselleck, Werner Conze, Otto Brunner, eds., Geschichtliche Grundbegriffe: Historisches Lexikon zur politisch-sozialen Sprache in Deutschland (Stuttgart: Klett-Cotta, 2004 [1972]). For an interesting contemporary attempt to build a ‘critical lexicon’ of political concepts, see 
. Last accessed May 13, 2016.
11.
Cited in Thomas Burger, Max Weber’s Theory of Concept Formation (Durham: Duke University Press, 1987), 77.
12.
William Connolly, The Terms of Political Discourse, 3rd Edn (Princeton: Princeton University Press, 1993), 23.
13.
As Koselleck reminds, there is a difference between concept and metaphor, although metaphors can become concepts. See Reinhart Koselleck, ‘A Response to Comments on Geschichtliche Grundbegriffe’, in The Meaning of Historical Terms and Concepts, eds. Hartmut Lehmann and Melvin Richter (Washington DC: German Historical Institute, 1996), 60.
14.
Ophir, ‘Concept’, 1.
15.
Ludwig Wittgenstein, Philosophische Untersuchungen/Philosophical Investigations, 4th Edn, trans. G.E.M. Anscobe, P.M.S. Hacker and J. Schulte (Hoboken: Wiley Blackwell, 2010 [1953]).
16.
Frederick Cooper, Colonialism in Question: Theory, Knowledge, History (Berkeley: University of California Press, 2005), 7.
17.
Rogers Brubaker and Frederick Cooper, ‘Beyond Identity’, Theory and Society 29, no. 1 (2000): 1.
18.
Connolly, The Terms of Political Discourse.
19.
Guzzini, ‘The Ends of International Relations Theory’, 528.
20.
Piki Ish-Shalom, Democratic Peace: A Political Biography (Ann Arbor: University of Michigan Press, 2013).
21.
Ido Oren, ‘From Democracy to Demon: Changing American Perceptions of Germany during World War I’, International Security 20, no. 2 (1995): 147–84; Stephen Chan, ‘In Search of Democratic Peace: Problems and Promise’, Mershon International Studies Review 41, no. 1 (1997): 59–91.
22.
Barry Buzan, Ole Waever and Jaap De Wilde, Security: A New Framework for Analysis (Boulder, CO: Lynne Rienner, 1998); Ole Waever, ‘Politics, Security, Theory’, Security Dialogue 42, nos. 4–5 (2011): 465–80.
23.
See Brubaker and Cooper, ‘Beyond Identity’; Charlotte Epstein, ‘Who Speaks? Discourse, the Subject and the Study of Identity in International Politics’, European Journal of International Relations 17, no. 2 (2010): 327–50; Richard Ned Lebow, The Politics and Ethics of Identity (Cambridge: Cambridge University Press, 2012). See also Hopf’s (not quite successful) attempt to refrain from ‘pre-theorizing’ when analysing identity discourses, Ted Hopf, Social Construction of International Politics (Ithaca: Cornell University Press, 2002).
24.
Koselleck, ‘A Response to Comments’, 64.
25.
Reinhart Koselleck, ‘Introduction and Prefaces to Geschichtliche Grundbegriffe’, trans. Michaela Richter, Contributions to the History of Concepts 6, no. 1 (2011): 8.
26.
Michel Foucault, The Order of Things (New York: Vintage Books, 1970 [1966]).
27.
This ties in with Connolly’s critique of the distinction between ‘normative’ and ‘descriptive’ concepts and his point that key concepts always help us to grasp the world in both ways. See Connolly, The Terms of Political Discourse, 22–35.
28.
For a selection, see Berenskoetter, Concepts in World Politics.
29.
Koselleck, ‘Response to Comments’, 65.
30.
Connolly, The Terms of Political Discourse, 14.
31.
Koselleck, ‘Introduction and Prefaces’.
32.
Michael Doyle, ‘Kant, Liberal Legacies, and Foreign Affairs’, Philosophy and Public Affairs 12, nos. 3–4 (1983): 205–35, 323–53; Stephen Chan, ‘In Search of Democratic Peace: Problems and Promise’, Mershon International Studies Review 41, no. 1 (1997): 59–91.
33.
Connolly, The Terms of Political Discourse, 17–20.
34.
Reinhart Koselleck, Futures Past: On the Semantics of Historical Time (New York: Columbia University Press, 2004 [1979]), 86.
35.
Koselleck, ‘Introduction and Prefaces’, 20.
36.
Ibid., 20.
37.
Max Weber, Politics as a Vocation.
38.
Koselleck, ‘Introduction and Prefaces’, 20.
39.
Anthropologists, for whom sensitivity to context is of central importance, concede that ‘it does not seem possible…to give a single, precise, technical definition of context (…) the term means quite different things within alternative research paradigms, and indeed even within particular traditions seems to be defined more by situated practice’, C. Goodwin and A. Duranti, eds., Rethinking Context (Cambridge: Cambridge University Press 1992), 2. See also Lawson’s discussion of historical context, George Lawson, ‘The Eternal Divide? History and International Relations’, European Journal of International Relations 18, no. 2 (2012): 213–16.
40.
Koselleck, ‘Introduction and Prefaces’, 16.
41.
These thinkers function as representatives and sources of inspiration; the aim here is not to provide a specialised discussion of their work and the debates surrounding it.
42.
Melvin Richter, ‘Appreciating a Contemporary Classic: The Geschichtliche Grundbegriffe and Future Scholarship’, in The Meaning of Historical Terms and Concepts, eds. Hartmut Lehmann and Melvin Richter (Washington DC: German Historical Institute, 1996), 10.
43.
On his work and influence see Lehmann and Richter, The Meaning of Historical Terms; Willibald Steinmetz, ‘Vierzig Jahre Begriffsgeschichte – The State of the Art’, in Sprache – Kognition – Kultur, eds. H. Kämper and L.M. Eichinger (Berlin and New York: Walter de Gruyter, 2008): 174–97; and articles in the journal Contribution to the History of Concepts.
44.
Reification is a distinct concern of the third approach discussed below.
45.
Steinmetz, ‘Vierzig Jahre Begriffsgeschichte’, 188–91.
46.
Hans Ulrich Gumbrecht, Dimensionen und Grenzen der Begriffsgeschichte (München: Wilhelm Fink Verlag, 2006), 5.
47.
See Steinmetz, ‘Vierzig Jahre Begriffsgeschichte’.
48.
Koselleck, ‘Introduction and Preface’, 25.
49.
J.G.A. Pocock, ‘Concepts and Discourses: A Difference in Culture?’, in Lehmann and Richter, The Meaning of Historical Terms, 49.
50.
The other representative is, of course, Quentin Skinner.
51.
See J.G.A. Pocock, Politics, Language, and Time (The University of Chicago Press, 1989); J.G. A. Pocock, Political Thought and Political History (Cambridge: Cambridge University Press, 2009).
52.
Pocock, ‘Concepts and Discourses’. It is worth noting that Skinner’s later work moved closer to Koselleck’s approach, see Quentin Skinner, ‘A Genealogy of the Modern State’, Proceedings of the British Academy 162 (2009): 325–70. As do some endorsements of the ‘Cambridge School’ approach in IR, see Christian Reus-Smit, ‘Reading History Through Constructivist Eyes’, Millennium – Journal of International Studies 37, no. 2 (2008): 395–414; Lawson, ‘The Eternal Divide?’. For a comparison of Koselleck and Skinner, see Kari Palonen, ‘The History of Concepts as a Style of Political Theorizing: Quentin Skinner’s and Reinhard Koselleck’s Subversion of Normative Political Theory’, European Journal of Political Theory 1, no. 1 (2012): 91–106.
53.
Herbert Blumer, ‘What Is Wrong With Social Theory?’, American Sociological Review 19, no. 1 (1954): 5–6.
54.
Giovanni Sartori, ‘Concept Misformation in Comparative Politics’, The American Political Science Review 64, no. 4 (1970): 1033–53. On Sartori’s work and influence more generally, see David Collier and John Gerring, Concepts and Methods in Social Science (London: Routledge, 2009).
55.
Sartori was especially critical of quantitative research that reduces concepts to static variables. See also Susanne Hoeber Rudolph, ‘The Imperialism of Categories: Situating Knowledge in a Globalizing World’, Perspective on Politics 3, no. 1 (2005): 5–14.
56.
A situation when the same concept ends up describing very different things. Or, conversely, where analysts disagree over whether seemingly different empirical phenomena can be made to fit under the same concept.
57.
Sartori, ‘Concept Misformation’, 1042.
58.
See Giovanni Sartori, ed, Social Science Concepts: A Systematic Analysis (London: Sage Publications 1984); Gary Goertz, Social Science Concepts (Princeton: Princeton University Press, 2004); Collier and Gerring, Concepts and Methods.
59.
Sartori, ‘Concept Misformation’, 1040–41.
60.
For a study that creatively widens the concept of democracy and shows sensitivity to a new context, see Lisa Wedeen, ‘Concepts and Commitments in the Study of Democracy’, in Problems and Methods in the Study of Politics, eds. Ian Shapiro, Rogers M. Smith and Tarek E. Masoud (Cambridge: Cambridge University Press, 2004): 274–306.
61.
Collier and Gerring, Concepts and Methods, 97–150.
62.
Sartori, Social Science Concepts; Blumer, ‘What is Wrong with Social Theory’, 5.
63.
W.B. Gallie, ‘Essentially Contested Concepts’, Proceedings of the Aristotelian Society 56 (1956): 169.
64.
Gallie, ‘Essentially Contested Concepts’, 176.
65.
Tellingly, Foucault at one point called his work ‘antiscientific’ in orientation. See Michel Foucault, Society Must be Defended (New York: Picador, 2003).
66.
Foucault, Society Must be Defended, 7.
67.
Michel Foucault, The Archeology of Knowledge (London: Tavistock Publications, 1974 [1969]), 49.
68.
Michael Foucault, Discipline and Punish (London: Penguin Books, 1991 [1975]); Michel Foucault, Madness and Civilization (New York: Vintage Books, 1988 [1964]).
69.
Foucault speaks of ‘an insurrection against the centralizing power-effects that are bound up with the institutionalization and workings of any scientific discourse organized in a society such as ours’, Foucault, Society Must be Defended, 9.
70.
Connolly, The Terms of Political Discourse, 232.
71.
Michel Foucault, ‘Nietzsche, Genealogy, History’, in Language, Counter-Memory, Practice: Selected Essays and Interviews, ed. D.F. Bouchard (Ithaca: Cornell University Press, 1977), 142.
72.
Foucault, ‘Nietzsche, Genealogy, History’, 146.
73.
Jens Bartelson, A Genealogy of Sovereignty (Cambridge: Cambridge University Press, 1995).
74.
Foucault, ‘Nietzsche, Genealogy, History’, 139.
75.
This spectrum is not exhaustive, of course. For a Bourdieusian take, which Guzzini also seems to favour, see Rebecca Adler-Nissen, ed., Bourdieu in International Relations (London: Routledge, 2013). For an approach inspired by Derrida, see Badredine Arfi, Rethinking International Relations Theory via Deconstruction (London: Routledge, 2012). For tracing a concept through actor-network theory, see Christian Bueger and Felix Bethke, ‘Actor-networking the ‘Failed State’ – An Enquiry Into the Life of Concepts’, Journal of International Relations and Development 17, no. 1 (2014): 30–60.
76.
Koselleck at one point uses the term ‘collective singular’, Koselleck, Futures Past, 34.
77.
Historians of ideas have long grappled with this issue, of course.
78.
Michael Barnett and Robert Duvall, ‘Power in International Politics’, International Organization 59, no. 1 (2005): 39–75.
79.
Emanuel Adler and Vincent Pouliot, ‘International Practices’, International Theory 3, no. 1 (2011): 1–36.
80.
See Felix Berenskoetter, ‘Thinking about Power’, in Power in World Politics, eds. Felix Berenskoetter and M.J. Williams (London: Routledge, 2007): 1–22; Guzzini, ‘Power’; Ringmar, ‘The Search for Dialogue as Hindrance to Understanding: Practices as Inter-Paradigmatic Research Program’, International Theory 6, no. 1 (2014): 1–27.
81.
Hoeber Rudolph, ‘The Imperialism of Categories’; Daniel J. Levine, Recovering International Relations: The Promise of Sustainable Critique (Oxford: Oxford University Press, 2012); Cooper, Colonialism in Question.
82.
See, for instance, Christopher Hobson and Milja Kurki, eds., The Conceptual Politics of Democracy Promotion (London: Routledge, 2012).
83.
Connolly, The Terms of Political Discourse, 227.