Posted on Thursday 26th July 2012
by Dani Nedal
Department of Political Science and International Studies and the Institute for Conflict, Cooperation and Security (ICCS) at the University of Birmingham.
Trust and International Relations: tales of a fledgling research agenda
On 23 June 2012, Dr. Naomi Head (University of Glasgow and Honorary Research Fellow in the ICCS) convened a panel at the British International Studies Association meeting in Edinburgh on “The Politics of Transformation: Trust, Empathy, and Dialogue”. Joining Dr. Head were Professor Nicholas Wheeler (Director of ICCS at the University of Birmingham), Laura Considine (PhD candidate at Aberystwyth University), Dr. Vincent Keating (lecturer at Durham University), and Dr. Jan Ruzicka (lecturer at Aberystwyth University). Professor Karin Fierke, from St. Andrews University acted as the discussant. The panel developed out of work on trust in international politics that was first begun by the participants at Aberystwyth University and each has collaborated with Wheeler in producing outputs in this area under the auspices of his project on ‘The Challenges to Trust-Building in Nuclear Worlds. The papers presented were an illustration of the divergences on fundamental issues of epistemology as well as definition and measurement of the key concept of trust within the broad research agenda.
Keating and Ruzicka’s paper, entitled ‘Confidence and Trust in Relations among States’, set out to offer an operational definition of trust and a means to identify it. They define trust as a cognitive mechanism by which one sets aside risk and uncertainty, substituting it for the assumption that others will abide by agreements or acceptable social behavior so as to ‘reduce complexity to manageable proportions’. They contrast this with confidence, which they understand to encompass expectations that are calculated based on perceptions of the other’s interests. Many in the trust literature would call this ‘strategic trust’, as opposed to ‘moralistic trust’ (which Keating and Ruzicka call simply ‘trust’).
Having defined trust as something that eliminates the feeling of uncertainty/vulnerability (they use the terms almost synonymously, which is problematic in itself), they go on to argue that trust can be said to exist when states (who are assumed to be risk-averse as a rule) do not adopt policies aimed at reducing actual vulnerability, i.e. hedging strategies. Trust and hedging are seen thus as competing ways by which states deal with uncertainty in the international arena, and the absence of one is seen as an indicator of the other.
They contrast this means of identifying trust with three other practices: 1) conflation of trust and cooperation; 2) trust as the voluntary acceptance of vulnerability; and 3) discourse analysis. They reject the first method because cooperation is possible without trust as they have defined the term, requiring only what they have called confidence (and they correctly identify some level of confidence as necessary, but not sufficient for cooperation). The second method is seen as inappropriate because they not only define away the possibility of vulnerability co-existing with trust (as discussed above), but they also reject the notion that trusting relationships can be entered into knowingly. For them, trust is ‘habitual’, not a conscious decision. Their objection to the third way of identifying trust, the use of interviews and discourse analysis, is more methodological than conceptual. They see this approach as unreliable because it can be misled by the strategic use of the language of trust, when behavior might tell a different story. While their claim that official discourse should not be taken at face value is certainly correct, this is hardly a real debate. Scholars that work with documents, discourse analysis, surveys and interviews – not just on trust, but all other issues as well – are painfully aware of the challenges to producing strong inferences from these sources and have devised ways to avoid falling into this trap. These are not faultless methods by any means, but nor should they be readily dismissed, namely because inferring motives, ideas or other mental states (such as trust) from behavior alone can be an equally challenging task.
Ultimately, Keating and Ruzicka’s approach to identifying trusting relationships is lacking because it offers no positive indicators and instead suggests that trust can be said to operate when states don't hedge and all other possible explanations for the absence of hedging are also not there. The problems with this are twofold. First, this would depend on a properly specified theory of when states should balance or hedge against other states to establish a baseline to compare observed behavior with. IR theory has several candidates for this task, and they fail to clearly choose one. Even Realism, which appears to be their preferred target, actually encompasses several competing hypotheses on why, when and how states hedge (or balance, bandwagon, pass the buck, or adopt a myriad of other possible strategies) against others. The second problem is that not only can we find a very long list of competing explanations and predictions of hedging, we can also find an equally long list of progressive explanations (that is, explanations that don’t contradict these theories’ hard core assumptions and claims) for its absence in specific cases. This means that the method of elimination, although valid, is an extremely complicated one. This is usually where positive indicators and hypothesized causal mechanisms (‘process tracing tests’) come in to allow the researcher to adjudicate in favor of one theoretical explanation or another. Since Keating and Ruzicka offer no such thing, eliminating rival explanations does nothing for supporting their own.
Considine’s paper, ‘Trust as a rhetorical tool: An alternative approach to trust and IR’ stands in stark contrast to Keating and Ruzicka. She is interested precisely in the ways in which ‘trust’ is used not only by diplomatic language, but also by researchers. Her paper was a staunch critique of most social scientific research on trust, rejecting the common practice of treating it as a ‘thing that exists’ in and of itself. Instead, she argues that trust is an ineffable construct that cannot by its very nature have a universal definition and application, but will depend on what questions one asks. She claims that any attempt to treat trust as an abstract phenomenon and isolate it from how it is felt by people is futile. She identifies three major problems with these attempts to define a concept that ‘cannot be pinned down’; and the lack of clarity that ensues.
The three common mistakes she identifies in attempts to define trust are: 1) the use of false analogies; 2) definition by negative description; and 3) substituting typologies for definition. In the first category, she includes those who seek to conceptualize trust in IR by analogy with trust in interpersonal relations, (e.g. hiring a babysitter or buying a car), thus oversimplifying a much more complex phenomenon. The second error is that of trying to describe trust by stating what it’s not: familiarity, confidence, fully rational or fully irrational, faith, expectation, etc. The third category is reserved for those who, in failing to provide a description that captures the phenomenon of trust because of its inherent fluidity, attempt to generate typologies: strategic, moralistic, generalized, particularized, abstract, functional, and so on.
Considine’s preferred approach, she told the panel audience, would draw on the thinking of Ludwig Wittgenstein, C. Wright Mills, Kenneth Burke and Quentin Skinner, focusing on the rhetorical uses of trust. Unfortunately, however, she did not elaborate on what that might actually look like. Ironically enough, that leaves us with yet another ‘negative definition’ of trust.
Moreover, she is not suggesting, it seems, that trust is any more ineffable or elusive than any other concept or phenomenon studied by social scientists, which makes this really a case of epistemological and methodological commitments taking precedence over empirical questions. This is not the space to debate the merits and pitfalls of the linguistic turn in social sciences, but Considine’s own claim that our understanding of trust will depend on what questions we ask of it could be expanded to allow for a more epistemologically and methodologically eclectic (or agnostic) position. That is, while research questions about how people use the word trust might be better served by Considine’s proposed approach, this should not preclude us from attaching a precise definition to the term and then focusing on studying the phenomenon that we are describing. Social scientists’ definition of a term need not conform to common usage(s) as long as we are clear about what we mean by it and that we have proper indicators that allow us to identify the phenomenon (see discussion above). To give another example, the fact that most people use words like ‘cooperation’ or ‘anarchy’ rather loosely hasn’t stopped social scientists from agreeing to very narrow definitions of these terms and proceeding to study their causes and effects.
The paper presented by Head and Wheeler did not address questions of definition, epistemology or method directly, but did rely on implicit positions on these issues, which are worth discussing briefly. Their paper, titled ‘Conflicting Narratives of (In)Security: 2003-2006 Nuclear Negotiations Between Iran the E3’, seeks to show how mounting tensions concerning the Iranian nuclear program are driven not only by material factors and strategic considerations, but also crucially by what they term, following Jonathan Mercer, emotional beliefs. One of their key claims is that these mind-sets, ideological fundamentalism (the belief that the other is inherently bad) and peaceful/defensive self-images (the idea that one is benign and that this is unequivocally obvious to others) can drive international conflict even in the absence of fundamentally incompatible interests. (For a brief reflection on the more substantial aspects of their paper and empathy, see here).
I would argue, without trying to read too much into their paper (although they are sure to disagree with my reading of it anyway), that they unwittingly accept the epistemological/methodological position outlined above. That is, they are interested in how people think, feel and use language, but also propose more or less objective definitions of terms like trust, empathy and emotions that serve as conceptual tools apart from their use in common language. Although they fall into this position unintentionally, it helps them avoid the pitfalls flagged by Keating and Ruzicka and Considine. By taking language seriously, but not relying exclusively on it, they can use interviews with diplomats and officials to construct competing narratives about the negotiations that go beyond individual cognitive maps and account for their intersubjective nature, while at the same time manage to stand back from what people say and look objectively at what they are doing and how these narratives stack up to the ‘real world’ based on multiple evidence sources.
These types of issues lie at the core of any research agenda and their persistence front and center is a sign of a still underdeveloped area. Debating them openly and consistently, especially at this stage of the project, can be productive as it offers clarity and transparency to fundamental commitments that are often left unquestioned. Being able to achieve some form of minimal consensus on these issues, however, will determine the cohesion of the research agenda and whether it can advance to focus on the study of its object of analysis.