It is significant, we believe, that whenever the concept of trust is debated, one is led almost inevitably to some of the most fundamental problems and conundrums of sociology. We will try to show this below. We are thankful to our interlocutors — Elena Esposito (2024) and Guido Möllering (2024) — since their thoughtful critiques have helped make the stakes of this discussion clearer.

First, a brief mea culpa. We are novices to the field of trust research, and as a result, we missed some important contributions. Most importantly, we missed the fact that Guido Möllering (2013) has made a recommendation similar to our own, namely that trust researchers should “speak of trusting, not trust,” thereby emphasizing that “the object of study is not just a measurable outcome […] but the particular ways such outcomes are produced and used” (p. 286). This was clearly an oversight, but we do not think that it means that our approach is easily “reconcilable with the literatures [we] criticize,” as Möllering says (2024, p. 200).

At the same time, we also do not think that we have “slashed and burned” our way through this literature. This characterization fails to take account of what we mean by “critical reconstruction”. Möllering (2024) says that “instead of trying to debunk trust […] research,” we should have recognized it as “a rich resource for making sense of what the skillful practice of trusting entails” (p. 203). This is almost word-for-word what we say about our approach:

We list these issues not simply as methodological problems affecting the measurement of trust, but as potential evidence about the nature of trusting as practical action. As such, they offer a rich source of insights to inform the reconstruction of the sociological theory of trust (Eyal et al. 2024, p. 173).

In other words, existing and previous research on trust is still useful when we recognize its limitations and the conditions from which knowledge about trust is produced. And to borrow from Möllering’s description of his assessment of rational choice accounts of trust, we too think that current approaches in conceptualizing trust are “incomplete”.

So what do we mean by “critical reconstruction”, and why is it neither “debunking of”, nor easily “reconcilable with” previous trust research? We believe that we have taken previous trust research seriously, and examined it carefully on its own terms. Hence, this is not a debunking. We do not impose on it an external theoretical frame, terms not its own, against which it is measured and found wanting. At the same time, by carefully following the logic of trust research, we think we have shown that one is led to certain aporias, conundrums, and contradictions. These are not easily “reconcilable”. They require some work of reconstruction, precisely because, as indicated earlier, they are not unique to this field but go to the core of the sociological enterprise.

This goes directly to Möllering’s strongest disagreement with our approach. He says that “Eyal et al. make no effort in their paper to consider what trusting is about” (p. 202). We think it is quite the contrary. Namely, we make every effort to do so by carefully considering the most authoritative accounts that tell us what trusting is about. We followed the prescriptions of accounts of trusting in attempts to analyze and interpret our data, only to come up short. Our rejection of the idea of “trust itself” — the idea that there is an “essence or nature of trust(ing)”, which has to do with “the problem trust solves” — is not an assumption we began with, nor does it reflect lack of trying. It is a finding we arrived at through this process of critical reconstruction by paying careful attention to the inconsistencies that kept cropping up, even among the most careful of researchers and the most profound theorists.

What is trusting about, according to trust research? Both Esposito and Möllering make a similar argument about the essence of trust. Esposito quotes Luhmann to the effect that trust is a strategy “to deal with the complexity of situations and decisions in the most effective way, even when one does not have all the information” (p. 195). Möllering says that trusting is “a way of dealing with irreducible uncertainty and vulnerability in social relations” (p. 202). This seems to us to be roughly the same argument: When one does not have all the information, one is uncertain; one cannot have all the information, due to complexity. So both arguments hark back to Luhmann. But Luhmann (2017, pp. 7–8, 23, 79) has told us that besides trust there are many other strategies and practices that also perform the same function of reducing complexity and dealing with uncertainty — symbolism, familiarity, confidence, mistrust, hope, and rational prediction. If so, the reason why trust, and not let’s say, mistrust or rational prediction, may prevail in a particular situation does not have to do with reducing complexity or dealing with uncertainty, since these other strategies could have done so too.

It gets even murkier. Not only are the different strategies interchangeable in principle, they are also hard to distinguish in practice. Immediately after distinguishing trust from confidence, Luhmann (1988) qualified the distinction, saying that a researcher could not sort cases “according to whether they are based respectively on confidence or on trust,” because “a relation of confidence may turn into one of trust if it becomes possible […] to avoid that relation […] [and] trust can revert to mere confidence” if it is routinized or if one feels that their action may have no real influence (p. 98). Ultimately, “the relative emphasis on […] confidence or trust […] remains a matter of definition, and […] one can choose to see the relation — the decision to see the doctor — as either unavoidable confidence in the medical system or a matter of risky choice” (p. 101). You can see what we mean by an aporia. Immediately after defining trust by differentiating it from a term that serves as its limit on one side, trust opens up to reveal that the distinction is in fact internal to it. Essentially, Luhmann says that the very same action — going to the doctor — could be described either as a form of trusting or as a form of thoughtless, routinized confidence, depending on how one “choose[s] to see the relation,” depending on where the sociologist puts “the relative emphasis.” It should be obvious that the same goes for the other limits against which trust is defined, namely mistrust and rational prediction. The boundaries of the concept turn out to be a Möbius strip that leads one from outside to inside in the course of a single paragraph.

And what of Möllering’s definition of the essence of trusting as “suspension,” for which the “leap of faith” is a metaphor? Esposito, not us, offers a strong rebuttal. Trust is not hope, she says, relying on Luhmann. Both hope and trust respond to complexity and uncertainty, but the “hopeful simply wishes for a good outcome,” while in trusting “a decision must be made […] that is, one reflects future contingency in the present decision. Whereas hope disregards contingency, trust reflects contingency” (p. 195). It is of course possible that Möllering is right and Esposito/Luhmann are wrong, or vice-versa, but if we take them together as alternative permutations of the conceptual structure of trust research then we are faced with a similar aporia and a similar journey along a Möbius strip — the essence of the thing (suspension) is revealed to be its outer limit (hope or blind faith).

As we argue in the essay, this indeterminacy of trust — the way it keeps flickering, each time appearing in a different state — stems from trying to resolve in theory what for individuals is a practical problem. Namely, how to convince themselves and others that their trusting is responsible and not blind faith; or by the same token, that their skepticism is not debilitating mistrust. And in short, how to give their trusting an accountable form. When Luhmann (1988) says that “one can choose to see the relation — the decision to see the doctor — as either unavoidable confidence in the medical system or a matter of risky choice” (p. 101), what is left unsaid is who is this “one”? Clearly, Luhmann means the sociologist, but why shouldn’t we give the actors themselves this leeway? Isn’t it the case that how they go about this visit, how they frame and present it to others, what story they tell themselves and others, is the most important factor determining whether this is trust or confidence? However, Möllering says that the matter of accountability and responsibility is “a derived issue, not the core issue” (p. 202), insisting that the core issue is how to deal with irreducible uncertainty.

The obvious question we put to Möllering is: from whose point of view is the state of affairs considered uncertain? From whose point of view is the uncertainty treated as “irreducible”? If we take the point of view of the actors as our guide, it is obvious that to account for one’s trust as responsible also entails constructing uncertainty as reducible and minimal. Only when we view things from the scholar’s point of view, does irreducible uncertainty appear as a metaphysical constant with which actors always have to contend. Because, as we argued in the essay, the scholastic point of view freezes time in its tracks and demonstrates that whatever seems to be certain is, in fact, uncertain, provided you have enough time to go down the rabbit hole. Does going to the doctor involve “irreducible uncertainty”? It depends on whether the framing of a routine visit holds. The conventions of the visit at one and the same time entangle the individual in a network of expertise, and shove this network into the background, constructing familiarity (Luhmann, 2017, pp. 22–24). The frame of familiarity focuses attention on the patient–physician exchange as an exercise of informed, responsible trust. Uncertainty is constructed as minimal. But if the focus changes, if one looks around the corner, so to speak, the uncanny complexity of the risk calculations involved stares one in the face. Uncertainty suddenly appears irreducible, and one’s prior trust appears as unmerited blind faith. Just listen to this father, quoted by Brownlie & Howson (2005), worrying that the MMR vaccine might cause his child to become autistic:

Although it might be a very, very small percentage risk, it’s your child and if it gets that, you have to deal with that for the rest of your life, I mean would you ever forgive yourself? To feel that you were responsible and that you could have prevented that? (p. 226)

When uncertainty is constructed as irreducible, by looking outside the frame of a “routine well-visit”, what we are left with is “existential dread” (Giddens, 1990, p. 100), not knowing what would constitute trusting responsibly.

We began by saying that debating trust inevitably leads one to some of the most fundamental problems and conundrums of sociology. We hope that the reader can see by now what we meant. While we have learned an enormous amount from previous trust research — and sometimes reached similar formulations — our approach is not reconcilable with it. Instead of defining in advance what is trust and then comparing our definition with what actors do and say (and in the process, declaring some to be trusting and others not), we ask how actors make trusting happen as a matter of practice (in context, in vivo, embedded in concrete situations and in networks of expertise). We think that taking this ethnomethodological approach is imperative for trust research especially now because of what Giddens (1990) called “the double hermeneutic”, namely how “the discourse of sociology and the concepts, theories, and findings of the other social sciences, continually ‘circulate in and out’ of what it is they are about.” (pp. 40–45) One of us [Eyal] is reading now a book from 1980 — Street-Level Bureaucracy (Lipsky, 1980). One would have expected that in more than 200 pages dedicated to how police officers, teachers, health-care workers, etc. interact with the public, there would be some discussion of trust. Yet, the word does not appear in the index, nor is there any such discussion in the chapters. Dialing forward 44 years, one cannot open a newspaper without running into some complaint about loss of trust, growth of mistrust, and musings about what is trust, how to build it, etc. It is to Luhmann’s great credit that he took the word “trust” from everyday language and gave it a rigorous sociological conceptualization. But by now this conceptualization, worked and reworked by trust researchers, has circulated back into social practices — as those who study trust are called upon to help address the problem of mistrust — “thus constitutively altering their character.” (Giddens, 1990, p. 38). Sociologists and trust researchers should recognize that they are part of this constant reflexive movement, not its masters or detached observers.