Trust in the face of uncertainty: a qualitative study of intersubjective trust and risk.
Contemporary trust research regards trust as a way of dealing with uncertainty and risk.
Predominantly, it suggests that trust reduces uncertainty by means of risk assessment and
rational calculation. However, phenomenological research proposes that trust is an alternative
way of relating to uncertainty rather than a way to reduced uncertainty. This paper
investigates these propositions in an interview study on intersubjective trust. The study
focuses on the modes of uncertainty management employed in trust and risk, and particularly
on how knowledge, experience, familiarity, and decision-making are combined in the act of
trusting. The main finding is that trust and risk are better characterised as different ways of
perceiving the social and managing uncertainty, than as different elements of the same
decision process. The concept of ‘risk compartmentalisation’ is developed to describe the
different ways people work to contain risk and maintain trust by combining adaptation and
Morten Frederiksen, Dept. of Sociology and Social Work, Aalborg University,
Trust, Risk, Uncertainty, Niklas Luhmann, Interview study
Predominantly, sociologists have treated intersubjective trust as a kind of gamble involving
knowledge or experience. People, it is argued, gather evidence of trustworthiness to decide
whether trust is justified despite potential risks. Lascaux (2008), among others, has
convincingly argued that this approach fails to account for trust, but little empirical research
has investigated such propositions. However, phenomenological research on intersubjective
trust posits that trust does not require evidence, because when evidence is called for trust has
ended. This paper investigates these two propositions in a qualitative interview study on
intersubjective trust. It is argued from this study that trust is an alternative to risk as a way of
dealing with uncertainty.
Georg Simmel (1950:318) claims that: As a hypothesis regarding future behaviour, a
hypothesis certain enough to serve as a basis for practical conduct, [trust] is intermediate
between knowledge and ignorance about a man. Trust is neither certainty nor ignorance.
Rather, trust builds on knowledge which is sufficient for action but which is also insufficient
to be certain that trust is justified. Consequently, trust remains a hypothesis: a positive
expectation of the future that somehow overcomes uncertainty despite insufficient knowledge
(Möllering 2001). In what way trust overcomes uncertainty is, however, intensely debated.
Contemporary sociology on trust often claims that trust requires the potential gains from a
risky decision to outweigh the potential losses (Coleman 1982; Offe 1999; Sztompka 1999).
Furthermore, social intelligence and experience is argued to be important in judging
trustworthiness and consequently the level of risk (Hardin 1991; Yamagishi 2001). This
research regards trust as a bet or a gamble on uncertain outcomes subjectively justified by
experience and available information.
However, according to Simmel trust cannot be reduced to assessment or rational calculation
but also requires some further element that is ‘mediated neither by experiences nor by
hypothesis’ (Simmel 1950: 318). More radical claims can be found in the works of Løgstrup
(1997) and Grøn (2010), who argue that trust is a way of engaging the other that does not
require justification or evidence. These theories regard trust as a fundamental way of
engaging the social, rather than an expectation that requires justifications. This proposition
has not been explored empirically.
In this paper both claims about trust are explored in a qualitative interview study on
intersubjective trust. Two hypotheses are investigated. The first is that trust is a way of
dealing with uncertainty that requires choice, knowledge, and assessment of ‘objective’ risk.
The second is that trust and risk are different states of mind experiencing uncertainty in
different ways: risk requires knowledge and decisions, trust does not. The findings of this
explorative study provide some initial conceptual and empirical clarification of the
relationship between risk, trust, and uncertainty.
The paper first reviews sociological research on trust in relation to risk and uncertainty, and
sets out relevant hypotheses and research questions. Secondly, the interview study and
methodological considerations are presented. Thirdly, the analysis deals with three different
aspects of trust: familiarity, adaptation, and accepting vulnerability. Finally, the findings are
contrasted to other positions in contemporary trust sociology, and suggest new ways of
theorising and analysing trust. The study concerns only intersubjective trust – not trust in
institutions, experts, or systems. While the concept trust is often employed to describe both
kinds of relations, these are not identical kinds of trust. Intersubjective trust has to do with the
contingency associated with the agency of the other (Luhmann 1988, Seligman 1997).
Institutional trust and trust in systems have to do with performance and reliability (Luhmann
1988, Hardin 2002). Consequently, there is good reason to suspect that the experience of
trusting others is markedly different from the experience of trusting systems and instituions.
For that reason it seems advisable not to conflate these issue by including them in the same
Trust, risk, and uncertainty
In writings on trust, uncertainty is often used in a probabilistic sense, as uncertainty about the
respective probabilities of several known, possible outcomes (Blau 1964; Gambetta and
Hamill 2005; Lewicki et al. 2000; Misztal 2001; Offe 1999; Sztompka 1999). In this use of
uncertainty, risk describes the chance of incurring a loss or an adverse outcome (Coleman
1982; Elster 2007).
Luhmann, however, proposes a somewhat different rendering of uncertainty and risk.
According to Luhmann (1993:16) uncertainty is the subjective experience of contingency: a
multitude of potential futures exist, but only one of these will eventually manifest itself in the
present. Consequently, the future is experienced as uncertain since contingency may be
interpreted but cannot be reduced (Luhmann 1995). Luhmann (1993:104) conceptualises risk
as an integral part of any decision because the outcomes of decisions are evaluated at a point
in time when more information is available than at the time of making the decision.
Consequently, adverse but avoidable outcomes can be attributed to the decision and the
decision-maker. According to Luhmann, risk is a state of mind in which one tries to
anticipate this attribution, calculating or guessing the future consequences of present
Generally, researchers agree that trust somehow allows people to overcome uncertainty. The
predominant approach is that trust is a decision in the face of risk. Molm et al. (2000)
conceptualise risk as the probability of incurring a loss in exchange relations unfortified by
contracts or other institutions binding the relation. This probabilistic approach is found in
much ‘rational choice’ research on trust (Coleman 1982; Gambetta 1988). Trust is considered
a more-or-less reflexive decision based on the assessed level of risk. Other parts of the
literature focus on the character of the assessment, such as perceived trustworthiness,
benevolence, encapsulated interest, or the mental constructs of the trustee (Cook et al. 2005;
Yamagishi 2001, Lascaux 2008). The risk of an adverse outcome is assessed on the basis of
rational calculation of interests or probabilities (Coleman 1982; Hardin 1991), perceived
institutional and identity characteristics of the other (Bacharach and Gambetta 2001; Burke
and Stets 1999; Tanis and Postmes 2005) experience (Hardin 2002; Misztal 1996; Weber and
Carter 2003), roles and institutions (Barber 1983; Lewis and Weigert 1985) and the emotional
conditions of the assessment itself (Barbalet 2009; Giddens 1991; Lewis and Weigert 1985,
2012). The act of trusting is regarded as the decision to accept a specific level of risk and
vulnerability (Baier 1986; Mistzal 2011). The decision to trust is based on interpreting
interests, experience, reputations, role expectations, social identities, and the emotional
environment of the trust situation.
A second, phenomenological approach describes trust as an unspecific positive attitude
towards the other’s agency. Trust is a way of relating to others without our considering risk
and probability, because trust does not concern the specific intentions of the other, but rather
the shared character of human agency (Grøn 2010; Levinas 1979; Løgstrup 1997; Seligman
The basic character of trust is revealed in yet another way. In love and sympathy
there is no impulse to investigate the other person’s character. […] Of ourselves
we make no conscious effort, for the simple reason that there has been nothing
about the other to raise our suspicion. If, on the other hand, we are not in
sympathy with the other person, or if there is a certain tension between us and
the other because of something about the other regarding which we are
uncertain or against which we react with irritation, dissatisfaction, or antipathy,
then we begin to form a picture of the other’s character. We then begin to see in
him or her a variety of dispositions because we are on our guard. (Løgstrup
When trusting, one is delivered into the hands of the other, whereas without trust one must be
on guard. To be on guard is to consider potential outcomes, similar to Luhmann’s definition
of risk. Trust, in contrast, precludes this assessment of the other’s character and motives
because trust relates to the agency of the other, not to some specific action. Following this
line of reasoning the trust-risk relation becomes one of alternatives, rather than one of trust in
the face of risk. You either guard against the potential unwanted actions of the other, or you
trust, setting aside reservations.
I propose two relationships between trust and risk: 1. Trust may be a type of decision on a
specific course of action in the face of risk. 2. Trust may be a perception of social relations
and the consequences of agency that is different from risk and decision-making. In order to
investigate the strength of each of these hypotheses, two research questions are pursued in the
1. What modes of uncertainty management are employed to establish a positive
2. How are knowledge, experience, familiarity, and decision making combined in
Data and method
This paper is based on an interview study conducted in Denmark in 2009 and 2010. The 19
people interviewed represent a high level of socio-economic and demographic variation.
They were recruited through a municipality school, a municipal social services department, a
shelter for the homeless and the acquaintances of my secondary network. While not
representative of the Danish population, the selection criteria where developed from a
quantitative study of the types of social, cultural and economic differences most closely
correlated to different levels of trust within Danish population (Author A, B) The socio-
economic selection critiaria aimed at getting participants from both the working class, the
group of marginalised and the upper middle class. The cultural selection criteria aimed at
getting participant with both elemenmtry, vocational, short tertiary and long tertiary
education. The participants’ age range from 23 to 72 years and is evenly distributed within
that range. 11 men and 8 women were interviewed. The socio-economic background of the
interviewed represents both the marginalised, the working class, lower and upper middle
class and upper class. However, only one interviewee can be said to belong to the upper class.
Additionally, both elementary, vocational, short, medium and long tertiary levels education
are present in the group of interviewees. The assumption here is that variations in social
position will also entail variations in the type and quality of social relations as well as in the
attitude towards and perceptions of these relations (Bourdieu 1999). The interviews were
conducted either in the home of the participant, at the participant’s place of work, or at my
department. The size of the sample fits the explorative nature of this study, allowing the
investigation of very different trust experiences, without aiming for representativity. While
full theoretical saturation on an issue such as trust is improbable, theoretical saturation was
achieved on the fundamental ‘compartment’ concept (Charmaz 2003; Glaser and Strauss
1967; Guest et al. 2006). The differentiation of compartment types is, however, an initial
classification and other types may be relevant. The key concepts of adaptive- and strategic
risk-compartmentalisation proposed in this paper are evident in most or all of the analysed
interviews and contradicted by none of them. Individual structures of strategic risk-
compartmentalisations vary, but these variations are outside the scope of this paper.
The interviews lasted from 90 to 120 minutes and were recorded and transcribed in full. The
interviews were deductively coded according to the two research questions. From this code
structure, a number of codes were developed to describe different ways of handling
uncertainty, information, experience, and decisions (Alvesson and Kärreman 2011; Fereday
and Muir-Cochrane 2006; Weston et al. 2001).
A qualitative investigation of trusting is methodologically challenging. Luhmann (1979)
states that someone who trusts never lacks justifications. Asking why trust seemed reasonable
in a specific situation forces the participant to respond with some sort of reason or
justification in order not to appear naïve or gullible. For that reason another approach was
chosen: Rather than justifications, the interview study investigated qualitative differences in
trusting. This allowed the interviewee to tell about different instances of trust, betrayed trust,
and qualitative differences in trust across different people, situations, or objects of trust. This
approach proved to put much less pressure on the participants to justify their trust.
Furthermore, the qualitative differences in trusting proved valuable in analysing the trust-risk
The interview procedure was inspired by Lamont’s (1992) work on moral and cultural
judgment in the context of relations. The adopted procedure focused on the relational
characteristics of personal networks: friends, family, children, spouses, acquaintances,
strangers, colleagues, business partners, employees, childhood friends, distant relatives, in-
laws and so forth. The participants were asked to describe and compare their trust in these
relations and encouraged to exemplify both trust and mistrust. The interview approach was
iterative, pursuing differences within and between relations. Any answer given on trust in a
specific relation or context was followed by questions on how this was similar to or differed
from other relations and contexts. This approach introduced a rigorous grammar of similarity
and difference into the interviews. The great advantage of this approach, compared to other
in-depth investigations of trust (Brownlie and Howson 2005; Weber and Carter 2003), is the
focus on differences in the quality of trusting, rather than on the binary question of trust
justification. This yielded a very rich interview material on the processes of trusting.
The analytical approach was inspired by Holstein and Gubrium’s (1995; 1997) concept of the
active interview. The combination of the relational categories employed in the interview and
the constructivist/ethno-methodological approach of the active interview results in a
decidedly semi-structured interview (Kvale 1997). This shifted the emphasis of the interview
from the usual probing of in-depth meaning (Charmaz 2003), to the probing of both in-depth
meaning and differences in meaning. Consequently, the patterns of meaning generated in the
interviews were also questioned and fragmented through constant differentiation and
contradiction. The purpose of this was to create (de-)fragmentation and defamiliarisation,
allowing new knowledge and questions to emerge (Alvesson and Kärreman 2011).
Analysis: Trust in the face of uncertainty
The main finding of this analysis is that trust and risk are separate ways of perceiving or
dealing with uncertainty in social relations. Trust, it would seem, is a way to stabilise
expectations and make risk considerations less relevant. This stabilisation combines
familiarity and adaptation. Drawing on experience, familiarity helps to categorise specific
social relations, situations, and objects as areas of either risk or trust. Adaptation actively
excludes the need to make decisions and probability assessments from a situation, thus
shifting it from risk to trust. Furthermore, decisions to accept vulnerability are found to occur
only within close social relations.
The key concept developed in this study is risk compartmentalisation. A risk
compartmentalisation defines a specific social relation, a specific context of interaction, or a
specific type of object (e.g. money) as uncertain in a way that requires risk: the monitoring of
possible future outcomes of present actions or decisions. The process of defining and
maintaining such areas, however, simultaneously defines an outside – a ‘non-risk’ side.
Outside of risk, familiarity provides ways of going on with social interaction without the
monitoring of future consequences. Returning to Løgstrup, a risk compartment is a place
where suspicion is allowed to reside while preventing it from undermining trust in general.
Without this containment, risk and the preoccupation with the other’s motives becomes
paralysing (Luhmann 1979; Løgstrup 1997:155). Strategic risk-compartmentalisations, based
on experience and familiarity, are positioned in advance, while adaptive risk-
compartmentalisations are active ways of dealing with uncertainty while interacting.
In the following I will first deal with familiarity, then adaptation, and finally the acceptance
Familiarity provides ways to habitually go on with social interaction, drawing on reflexive
and embodied experience (Luhmann 1979). Familiarity is, however, not a single,
undifferentiated mode of engaging the social. Rather, in regards to the issues of uncertainty
and agency, there seems to be a division within familiarity between risk and trust. This
division separates specific relations, situations, and objects into areas which familiarity
designates as either trust or risk. The resulting web of risk and trust areas is navigated with
great confidence by the interviewees. Trust in someone can be maintained in specific social
relations, e.g. a work colleague or friend, by excluding specific issues, objects, or situations
from that relation in advance: a strategic risk-compartmentalisation.
A recurrent example of a strategic risk compartmentalisation is the lending of money. Money
is routinely excluded from relations with friends, family, and acquaintances because lending
money may lead to regrets. Consequently, the trustor is forced to consider the future
consequences of lending money: to adopt the risk state of mind. The interviewee Carl gives
an example of this:
Interviewer: What about [friends] helping you with something? Is that something
you have experienced practically or economically?
Carl: Economically, no. I do not feel that you should be obliged economically to
help your friends, and that is probably because I have borrowed money from
friends myself and not been that good at paying them back. So that is why I do
not think that that is something you should see as a duty between friends. It’s sort
of like something quite apart.
Interviewer: Why is it something apart?
Carl: I do not want to mix friendships and money, so in that way. If people
wanted to borrow money from me I would not take it as something positive.
By excluding money, Carl bars risk is barred from the relation. He does not have to decide
whether a specific loan is a good idea or whether he is going to regret granting it. Deciding to
lend money to a friend involves risk and justifications, whereas considering money ‘sort of
something quite apart’ separates the trustful friendship from the risky business of loaning
money. Such strategic compartments create a social topography of relations, situations, and
objects excluded from trust and ‘surrounded’ by areas where trust is taken for granted.
Strategic compartmentalisations may involve general exclusions, as in the quote above, but
are usually deployed in more intricate ways. An example is the differentiation of domains for
revealing secrets about oneself. The interviewees explain how they discuss specific problems,
emotions, and experiences only with people where the information will not destabilise social
identities. They reveal secrets and personal information to others only if this does not require
them to consider possible regrets. Risk is strategically compartmentalised in the latter
relations, while trust is maintained in the first. In this way, secrets and personal information
‘belong’ in specific, familiar secrecy relations and consequently risk need not arise. Judith
explains which of her colleagues she reveals personal information to and which subjects are
Interviewer: Which of your colleagues would you tell private things to?
Judith: Of those I work with there is probably…..The one I talk with the most has
just quit. That leaves three of us who were stationed together in Ecuador. We
have been on weekends together and walked around in swimsuits, so we know
each other really well. And then there is another one I am getting really close to
and we talk about families, husbands and children. And there is also a third
woman who I think I have some really good talks with. And we talk about things
that are difficult with our children and husbands. Maybe not so much about
husbands. That is usually with the girlfriends.
She presents three levels of intimacy with different groups of colleagues: the first is very
close, when she does not have to specify subjects; the second is fairly close and includes
families, husbands and children as relevant subjects. Finally, there is the level of good chats
with one colleague, though only about children, husbands being a subject suited only to the
closer girlfriends. Without being prompted to do anything as elaborate as this, she easily
differentiates the subjects which are permissible in the respective relations from those which
are excluded. These categorisations seem to draw on a nuanced topography of strategic risk
Drawing on experience and familiarity, the interviewees exclude issues and relations which
may usually require assessment and decisions. What remains are issues and relations that are
usually irrelevant to decisions and potential regret. This does not eliminate contingency, nor
does it provide certainty of future events. However, by eliminating issues and relations which
familiarity designates as uncontrollable and uncertain, what remains is perceived as stabile
and trustworthy. In this quote Jack considers which of his close relatives he would allow to
borrow his car:
Interviewer: So if we imagined that they [his sister and her husband] needed to
borrow it [his car]?
Jack: I would say ‘no way’. Also to my daughter, but not to my son. And that is
because of trust and how she drives. She is antisocial and on the dole, and think
of how many times she would be fined and would damage the car. So that trust is
ruled out with that. I do not want my car damaged.
Interviewer: And that goes for your sister as well?
Jack: I do not know much about her motoring skills, but I would have more trust
in my sister than in my daughter. But as for my son, it is a hundred percent
Jack’s daughter, sister, and brother-in-law are here all assessed on a scale of motoring skill
and general trustworthiness. Nonetheless, neither of them is in fact trusted to borrow his car.
His son, however, is trusted completely to borrow the car and Jack offers no assessment or
justification of his trustworthiness throughout the entire interview. His son is simply trusted
and this requires no justification. In regard to his sister and daughter, however, trust is
supplanted by risk as he considers the possible consequences and regrets.
These strategic compartments provide general structures of interpreting and relating to the
social, but they cannot in any final way account for actual instances of trusting. To get a more
complete picture, trust in interaction must be addressed.
While strategic compartmentalisations may be sufficient foundations of trust in some
instances, often the trust-risk distinction is more blurred in interactions. The concept of
adaptive risk compartmentalisation describes how people actively work at creating and
maintaining the risk-trust differentiation. It is not a decision to define a situation as one of
either risk or trust, but rather an effort to keep expectations of the future stable.
As noted, money is usually compartmentalised outside of close social relations. Nonetheless,
several respondents use an adaptive compartmentalisation to handle situations in which the
lending of money is required: they decide that the loan is in fact a gift. Some keep this
decision to themselves while others tell the person borrowing the money that they can repay
it if and when they have the ability. An example of this is given by Leon, who has recently
gone bankrupt with his construction company. The quote shows how Leon transforms a loan
into a gift. This compartmentalisation bars risk outside Leon’s relation to his brother, thus
removing the need to reflect on whether the loan is a good idea and what the future
consequences of such a loan could be.
Leon: My brother, I loaned him 60,000 Euros, so part of my own debt is due to
him. […] Then you stand there thinking: ‘but I had to do it’. That is who I am
and if I had not done it, I would not be who I am. So I will just have to take the
beating that comes with that. Of course I will lend my brother money now if he
owes somebody money. […] And if he cannot repay me it, that is okay, because
we agreed that he did not have to repay it until he was back on his feet.
Leon maintains this exchange as one of trust by avoiding issues of decisions and regrets. By
redefining the situation surrounding this exchange of money, his trust is not conditional on
repayment a later time. The situation is shifted away from the risky expectations of
negotiated exchange through an adaptive compartmentalisation. This compartmentalisation is
not a decision Leon makes on the basis of knowledge or calculation, but rather a way he
makes the situation fit to a familiar way of interacting with his brother. To return to
Luhmann’s theory of risk, this compartmentalisation basically works by changing the
conditions of evaluation for the exchange such that, at a later stage, if the money is not
returned, the outcome cannot be attributed to a bad decision.
Other adaptive risk compartments require social imagery to fit situations to experience and
generate familiarity. Lisa is asked whom she would ask to watch over her luggage in an
airport, if anyone. She combines superficial knowledge and social imagery to establish trust
in imagined strangers:
Lisa: […] definitely someone not alone. That makes it easier to identify people in
some sort of context other than just singling out one person.
Interviewer: So the fact that they are together with other people, what does that
Lisa: Then I would make up a story about what kind of people they are. For
instance, a family father or the mother of some children I would assume you
could trust. And that they would show me trust, which is of course taken from
thin air, but you know that they are people who have responsibilities and you can
identify a little with them.
Strangers are usually an area of risk, but redefining a stranger as someone with whom Lisa
identifies, allows her to adaptively compartmentalise risk. Lisa has no way of stabilising
expectations from the situation alone. However, by creating an image of social relations
characterising the stranger, she sets this person apart from the risky category of strangers. She
produces a narrative about the other, allowing her to insert trust and familiarity to bolster her
self-assurance and stabilise expectations.
The sheer familiarity of the interaction itself may also lead to adaptive compartmentalisation.
In this quote Karen tells how, without any prior knowledge of the other, or any institutional
guarantees, she freely entrusted her children to a stranger:–
Karen: For instance, my friend. I first met her when our children were playing,
and she came over and asked if she could take Louisa and Allan over to the
station to buy ice cream and [I thought] that was actually okay! I did not know
her – [I knew] her children, but not her. And then she came back and we stood
there talking for an hour about everything, without really knowing each other.
That was completely insane – I have never experienced anything like it.
Karen reflexively recollects what is otherwise tacit and embodied while expressing her
surprise with her own actions. The shared situation of being parents with their children on a
playground and the familiar ritual of going for ice cream stabilise identities within a familiar
process. This allows Karen to adaptively compartmentalise risk and stabilise her expectations
of the future. Afterwards she is aware that she has in fact left her children with a stranger,
which seems ‘insane’ to her. Being in a state of trust rather than risk in this situation,
however, potential regrets or negative expectations never entered her mind.
Risk compartmentalisations may, nonetheless, also fail. Yvonne gives an example of this. Her
adaptive risk compartmentalisation crumbles when her positive expectation falters. What she
initially perceives as a familiar and unthreatening situation of need and helpfulness turns into
one of potentially serious regrets.
Yvonne: [...] I was standing by a road where we had run out of petrol and we
stopped all these people. And of course nobody had a spare can. You do not have
that anymore. And then a man stopped and said he would drive me to the petrol
station, where I could buy a can and he would drive me back. Well, I said: that is
incredibly nice of you, and my friend stayed with the car. And it was not until I
sat in the car that I thought, oh my god, here you are, goddamn it, well done, and
what if something happens. […]But it was not until I was sitting there [I
thought]: I do not really dare to let you drive me, so would you please pull over.
But it was also really cool because he was so sweet.
To the observer, the available knowledge and potential for regrets remains constant before
and after Yvonne gets in the car. However, familiarity falters as she starts to consider the
justifications and potential consequences of her choice. When she starts to evaluate the
potential outcome of her decision, trust gives way to risk. Under normal circumstances
Yvonne does not hitchhike, but an adaptive compartmentalisation sets this situation apart.
Needing help, meeting a helpful stranger, and her friend’s reassuring presence works to
separate the ‘risk situation’ of hitching a ride with an unknown man from the ‘trust situation’
of accepting help from a kind and helpful man in a situation of need. Had trust been a
decision in the face of risk, Yvonne would not have been surprised by the potential
consequences of her decision. She might not have accepted the ride or would have asked her
friend to accompany her. These are alternatives she does not even consider – until familiarity
falters, trust collapses, and risk de-compartmentalises.
Decisions are generally absent from the interviewees’ descriptions of trust. When they do talk
about risk and the decision to trust or not, it is part of a decision not to trust. However, an
exception exists in very close social relations when interviewees sometimes decide to trust
despite knowing that they will probably be let down. Carol expects to be let down, but
decides to trust nonetheless:
Interviewer: So you say that you basically trust your brothers. But there are
issues where you do not trust them.
Carol: Yes lots, but it is nothing really. One I cannot trust because he does not
pay me back when he borrows money. The other does not honour the agreements
we make. These are things that I know or I am prepared that they will happen. So
really they are not abusing my trust even though they let me down in these less
Carol does not compartmentalise risk but she suspends the potential for regrets by accepting
the disappointment in advance. Others describe similar approaches to ex-wives, childhood
friends and family. This decision to accept vulnerability only occurs in very close relations
that are important to the identities and emotions of the interviewed. A more example of this is
found in another interview. Arnold’s childhood friend is very important to him. Nonetheless,
first his wife and years later his girlfriend and wife-to-be, left him for this friend. Despite this,
Arnold claims to trust the friend in matters concerning women and otherwise. In the
following quote he describes the painful process of letting go of regrets while in the process
of losing his wife to his friend.
Interviewer: But nonetheless, he is still your friend?
Arnold: Yes, because I have known him since ’63 […]
Interviewer: Do you still trust him?
Arnold: Yes, very much so […]
Interviewer: You have to spell that out for me. He leaves with your girlfriend,
even though you are very close, but you still trust him? What about future
girlfriends – would you be worried about them?
Interviewer: Why is that, is it because you know him?
Arnold: Yes, it is because…Because we have talked for hours on the phone.
Interviewer: About this?
Arnold: Yes, and also while it was happening. We talked for hours about her.
About him suddenly paying her visits. That was… I did not sleep for bloody three
days and I am not one to be jealous.
He painfully works on maintaining his friendship and making the past irrelevant to the future.
By doing this he avoids the strategic compartment of never mixing this friendship with his
love life, or perhaps of abandoning the friendship altogether. The friendship is far too
important to Arnold and too much a part of his biography for him to let these incidents end it.
There is certainly decision-making involved in this type of accepted vulnerability. Both
Arnold and Carol know that they have been let down and that this may happen again. This,
however, makes it a question of risk rather than trust. If trusting, they would expect positive
future outcomes despite any past experience to the contrary. In accepting vulnerability they
are deciding that it is acceptable or meaningful to be disappointed in some regards because
their relationship to the other is fundamentally much more important. Carol’s trust in her
brothers has a more permanent character which makes the trust issue in any specific instance
Interviewer: And then there is family. Are there also different groups within your
family where you trust some more in different respects?
Carol: No, I have three brothers and that is my close family. And then there are
circles, with my partner’s family a bit further away. I do not know if this is really
about trust, because I trust them all completely.
Interviewer: Why are you unsure if it is about trust?
Carol: I do not know what it is about, but I have really thought about it. Maybe it
is about trust, because I know that, no matter what happens, I have my brothers.
I do not know about their wives and partners; things may happen to split them
Carol: This is because no matter what happens, they will still be there. So I still
trust them at the core because I know they are there. And they have abused my
trust with a series of stupidities, but that does not affect the basic trust relation
that I have with them.
Her brothers are permanent and reliable parts of her life, and her accepting vulnerability in
any specific regard changes little in her relation to them. That is not to say she trusts them in
regard to money or agreements, but the fact that she trusts them fundamentally makes her
decide to accept vulnerability in specific instances. Her general trust in her brothers, however,
is not a decision but something given.
In the above analysis I have addressed two explorative research questions: 1. What modes of
uncertainty management are employed to establish a positive expectation of the future? 2.
How are knowledge, experience, familiarity, and decision-making combined in trusting?
The tentative answer to the first question is that managing uncertainty involves the separation
of situations and relations familiarly uncertain from situations and relations experienced as
stable in comparison. This separation, which I have called compartmentalisation, is a
distinction drawn between two types of familiarity: one of risk and one of trust. People seem
to maintain elaborate structures of familiarity in which this distinction is ingrained. It is not,
however, a distinction based on knowledge – knowing the difference between areas where
trust is advisable and areas where risk is required. Rather, this difference is in itself part of
familiarity and experience. However, the distinction is not a rigid projection of past
experience onto the present, but is subject to an interpretation or constitution in which people
may shift the situation in one direction or the other. The process of shifting the situation into
one of either risk or trust is not a decision based on available knowledge, it would seem.
Rather, it is characterised by either avoiding the need to choose and to assess the situation or
by embracing the need for a knowledge based choice.
The tentative answer to the second question is implied in the answer to the first. Experience
and familiarity are present in both risk and trust, whereas knowledge-based decisions belong
only within the state risk. The possible futures, as it were, of the trust side are familiarly
stabile, whereas the possible futures of the risk side are familiarly unstable. The above
analysis suggests that, when the very issue of a knowledge-based decision becomes relevant,
the door shuts on trust. If one has concerns about trust – whether it may lead to regret – one
has in fact already entered a state of risk.
My study suggests that perceiving trust as a ‘gamble’ confuses objective observation and
subjective experience. The objective claim that trust always involves the possibility of
misplaced trust fits the gamble description. The subjective experience of trusting is, however,
not that of taking a chance. Consequently, the trust concepts developed by authors such as
Molm, Sztompka, Coleman, and Cook may be useful in modelling trust probabilities, but
they do not provide an adequate description or understanding of trusting.
This answer, however, leads to another equally fundamental question: if trust is not a
knowledge-based decision, is it only built from experience or is there some further element
similar to knowledge and decision, as in risk? The answer may be that trust relies on
experience and familiarity in a way that does not require experience to be translated into
decision-grade knowledge. Rather, this familiarity is similar to concepts such as ‘lifeworld’
(Husserl 1989), or Schütz’s notion of the ‘natural attitude’ (Schütz and Luckmann 1973).
Some research on trust proposes that this type of phenomenological attitude is the foundation
of trust (Grøn 2010; Levinas 1979; Möllering 2005), and this is supported by the findings of
my study. However, my study contradicts Möllering’s notion of trust as a way of overcoming
uncertainty and the risk of potential adverse outcomes. The idea here is rather that trust is a
way of experiencing uncertainty as unthreatening: trust is a positive expectation of the future.
This means that contingency is expected to stay within certain bounds set by life-world
experience. Luhmann (1979; 1993) argue that agents deal with the contingency of socio-
temporality in two ways. The first is a mode of continuity in which the future is expected to
be identical to the past, which is characteristic of the way we relate to the natural
environment and institutions. The second is the mode of events in which a potential event
exists in the future, is manifested in the present and becomes experience in the past. This is
characteristic of the situations requiring risk and trust decisions according to Luhmann. My
study indicates that neither of these temporality types can account for trust. The first cannot
account for the issues of double contingency experienced when dealing with other agents.
The second turns experience into knowledge, and trust into decision, which this study
suggests is incorrect. A third mode is hinted at here, but cannot be clarified by this study: A
type of socio-temporality relevant to issues of trust may be that of ‘familiar uncertainty’:
People expect others to respect the trust placed in them and assume that the actions and
decisions of these others will not result in adverse outcomes.
Experience underpins the strategic risk compartments setting out the limits of social relations
governed by familiarity and beyond which risk and decisions are usually applied. However,
any situation is basically indeterminable and consequently some sort of interpretation is
required. Adaptive risk compartmentalisation describes the tension between embodied
familiarity and decision-making in the face of uncertainty. A successful adaptive
compartmentalisation confines uncertainty to what is familiar and acceptable. Without this
compartmentalisation, the level of uncertainty would require a state of risk in order for agents
to make the best possible decision.
It is this type of positive scope of possible futures Hardin (2006) hints at through the concept
of ‘encapsulated interest’. He claims that the trustor assesses whether the trustee has
incentives to take the interests of the trustor into account. If so, the trustee is deemed
trustworthy and trust is justified. Similarly, Lascaux (2008) argues that trust is a mental
construct of the other and the others social obligations necessitated by the absence of
certainty. This study supports the strong link between uncertainty and trust rather than risk
and trust. However, the results indicate that trust is already lost once you need to consider the
trustworthiness of the other. Encapsulated interest only works if it is taken for granted on the
basis of familiarity. Once uncertainty is reflexively experienced as a manifest lack of
knowledge necessitating trust, trust is no longer attainable. Trust exist where uncertainty and
contingency has not become an issue of reflexively dealt with.
An interesting notion is Möllering’s (2005) idea of familiarization. Building on Luhmann’s
idea of trust-building as an incremental process, Möllering claims that trust is the result of a
process of trial and error: People actively experiment with trust to see if it is justified and
from this process of familiarisation learn to trust each other. My study, however, indicates
that the experiment is one of risk rather than trust. The process of familiarisation is a process
of experimenting with or investigating a specific social relation, fitting it into life-world
categories of familiarity. The development of a relationship from one of being strangers to
then acquaintances and finally friends involves the transition from risk to trust. This is
achieved by building experience with the other that gradually allows the
compartmentalisation of risk through familiarity.
Finally, a third finding is that when people explicitly choose to trust in very close social
relations, they choose to accept vulnerability. Vulnerability, rather than trust, seems a way of
handling an adverse outcome in advance in order to uphold social identities, close
relationships, and more general trust relations. Even in highly familiar relations inseparable
from one’s social biography, risk may be relevant and the need may arise to consider adverse
outcomes of choices and actions. However, as Garfinkel’s (1967) breaching experiments
have famously shown, dealing with such issues in an objectifying manner may have
devastating effects on social relations. Accepting vulnerability is a version of risk in which
the adverse effects of avoiding a future event is greater than the adverse effects of the event
itself. This version of the risk mode is sometimes confused with trust. Weber and Carter’s
(2003) analysis of trust within close social relations fails to make this distinction and,
consequently, it becomes difficult to separate trust from the active effort to maintain
relationships despite risk or betrayal. This issue could prove an important area of further
investigation in understanding how trust and risk deal with uncertainty, as well as in
developing a more nuanced understanding of the emotional conditions relevant to trust and
The general conclusion of this paper is that trust is better described as an alternative to risk
than as a way of dealing with risk. Consequently, the usually objectifying approach to trust
within current research may lead to an incomplete understanding of trust as well as to a too-
encompassing notion of trust. Contingency and the risk of adverse outcomes are always
present in the eyes of the researcher, but not so in the eyes of the trustor. Consequently, it will
be important to investigate in two respects when and why risk is absent, rather than how risk
is handled. First, it may prove fruitful to investigate the life-worlds of trust: the strategic
compartmentalisations. These may be both subject to cultural conventions and reflective of
specific biographical and socio-economic experiences. Secondly, research on the process of
trusting may move forward by investigating how people use adaptive compartments – under
which circumstances they are possible and under which circumstances they collapse. Finally,
the concept of risk compartmentalisation may prove helpful in investigating Luhmann’s and
Giddens’ theories of trust as a type complexity reduction.
Alvesson, M. and D. Kärreman. 2011. Qualitative Research and Theory Development.
Bacharach, M. and D. Gambetta. 2001. ‘Trust in signs.’ in Trust in society, ed. K.S. Cook.
New York: The Russel Sage Foundation.
Baier, A.. 1986. ‘Trust and Antitrust.’ Ethics 96(2):231-260.
Barbalet, J. M. 2009. ‘A Characterization of Trust, and Its Consequences.’ Theory and
Barber, B.. 1983. The Logic and Limits of Trust. New Jersey: Rutgers.
Blau, P.. 1964. Exchange and Power in Social Life. New York: Wiley.
Bourdieu, P.. 1999. ‘Understanding.’ in The Weight of The World: Social Suffering in
Contemporary Society, ed. P. Bourdieu et al. Cambridge: Polity.
Brownlie, J. and A. Howson. 2005. “‘Leaps of Faith’ and MMR: An Empirical Study of
Trust." Sociology 39(2):221-239.
Burke, P. J. and J. E. Stets. 1999. "Trust and Commitment through Self-Verification." Social
Psychology Quarterly 62(4):347-366.
Charmaz, K.. 2003. ‘Qualitative Interviewing and Grounded Theory Analysis.’ in Inside
Interviewing: New lenses, new concerns, ed. J.A. Holstein and J.F. Gubrium. London:
Coleman, J. S. 1982. ‘Systems of trust. A rough theoretical framework.’ Angewandte
Cook, K. S., R. Hardin and M. Levi. 2005. Cooperation without Trust? New York: Russell
Elster, J.. 2007. Explaining Social Behavior: More Nuts and Bolts for the Social Sciences.
Cambridge University Press.
Fereday, J. and E. Muir-Cochrane. 2006. ‘Demonstrating rigor using thematic analysis: a
hybrid approach of inductive and deductive coding and theme development.’
International Journal of Qualitative Methods 5(1).
Gambetta, D.. 1988. ‘Can we Trust Trust.’ in Trust, ed. D. Gambetta. Oxford: Basil
Gambetta, D. and H. Hamill. 2005. Streetwise. New York: Russell Sage Foundation.
Garfinkel, H. 1967. Studies in Ethnomethodology. Englewood Cliffs (N.J.): Prentice-Hall.
Giddens, A.. 1991. Modernity and Selfidentity. Cambridge: Polity Press.
Glaser, B. G. and A. L. Strauss. 1967. The Discovery of Grounded Theory: Strategies for
Qualitative Research. Chicago: Aldine Publishing Company.
Grøn, A.. 2010. ‘Trust, Sociality and Selfhood.’ in Trust, Sociality and Selfhood, ed. A. Grøn
and C. Welz. Tübingen: Mohr Siebeck.
Guest, G., A. Bunce and L. Johnson. 2006. ‘How Many Interviews Are Enough? An
experiment with data saturation and variability.’ Field Methods 18(1):59-82.
Hardin, R. 1991. ‘Trusting persons, trusting institutions.’ in Strategy and choice, ed. R.
Zeckhauser. Cambridge: MIT Press.
Hardin, R. 2006. Trust. Cambridge: Blackwell Publishers.
Hardin, R. 2002. Trust and Trustworthiness. New York: Russell Sage Foundation.
Holstein, J. A. and J. F. Gubrium. 1997. The New Language of Qualitative Method. Oxford:
Oxford University Press.
Holstein, J. A. and J. F. Gubrium. 1995. The Active Interview. Thousand Oaks: Sage.
Husserl, E.. 1989. The Crisis of the European Sciences and Transcendental Phenomenology.
Illinois: Northwestern University Press.
Kvale, S.. 1997. InterView: En Introduktion Til Det Kvalitative Forskningsinterview.
København: Hans Reitzel.
Lamont, M.. 1992. Money, Morals, and Manners. Chicago: The University of Chicago Press.
Lascaux, A.. 2008 ‘Trust and Uncertainty: a critical re-assessment’ International Review of
Levinas, E.. 1979. Totality and Infinity: An Essay on Exteriority. Dordrecht: Kluwer
Lewicki, R. J., C. Wiethoff, M. Deutsch and P. T. Coleman. 2000. ‘Trust, Trust
Development, and Trust Repair.’ in The Handbook of Conflict Resoloution:Theory and
Practice ed. M. Deutsch, P. Coleman, and E. Marcus. San Francisco: Jossey-Bass.
Lewis, J. D. and A. Weigert. 2012. ‘The social dynamics of trust: Theoretical and empirical
research, 1985-2012’ Social Forces 91(1):25-31.
Lewis, J. D. and A. Weigert. 1985. ‘Trust as a social reality.’ Social Forces 63(4):967-985.
Luhmann, N.. 1995. Social Systems. Stanford, Calif: Stanford University Press.
Luhmann, N.. 1993. Risk: a Sociological Theory. Berlin: Walter de Gruyter.
Luhmann, N.. 1979. Trust and Power: Two Works by Niklas Luhmann. New York: John
Wiley & Sons.
Løgstrup, K. E. 1997. The Ethical Demand. Notre Dame, Indiana: University of Notre Dame
Mistzal, B.. 2011. ‘Trust: Acceptance of, Precaution Against and Cause of Vulnerability.’
Comparative Sociology 10(2):358-379.
Misztal, B. 2001. ‘Normality and Trust in Goffman's Theory of Interaction Order.’
Sociological Theory: 9(3):312-324.
Misztal, B.. 1996. Trust in Modern Societies. Cambridge: Polity Press.
Möllering, G.. 2005. ‘Rational, institutional, and active trust: just do it!?’ in Trust under
pressure: Empirical investigations of trust and trust building in uncertain circumstances,
ed. K. Bijlsma-Frankema and R.K. Woolthuis. Cheltenham: Edward Elgar Publishing.
Möllering, G.. 2001. ‘The Nature of Trust: From Georg Simmel to a Theory of Expectation,
Interpretation and Suspension.’ Sociology 35(2):403-420.
Molm, L. D., Takahashi N. and Peterson G.. 2000. ‘Risk and trust in social exchange: An
experimental test of a classical proposition.’ The American Journal of Sociology
Offe, C.. 1999. ‘How Can We Trust Our Fellow Citizens?’ in Democracy and Trust, edited
by M.E. Warren. Cambridge: Cambridge University Press.
Schütz, A. and T. Luckmann. 1973. The Structures of the Life-World. Illinois: Northwestern
Seligman, A. B. 1997. The Problem of Trust. Princeton, N.J: Princeton University Press.
Simmel, G.. 1950. ‘The Secret and The Secret Society.’ in The Sociology of Georg Simmel,
edited by K.H. Wolff. London: The Free Press of Glencoe.
Sztompka, P.. 1999. Trust. Cambridge, UK: Cambridge University Press.
Tanis, M. and T. Postmes. 2005. ‘A social identity approach to trust: Interpersonal
perception, group membership and trusting behaviour.’ European Journal of Social
Weber, L. R. and A. I. Carter. 2003. The Social Construction of Trust. New York: Kluwer
Academic / Plenum Publishers.
Weston, C., T. Gandell, J. Beauchamp, L. McAlpine, C. Wiseman and C. Beauchamp. 2001.
‘Analyzing Interview Data: The Development and Evolution of a Coding System.’
Qualitative Sociology 24(3):381-400.
Yamagishi, T.. 2001. ‘Trust as a Form of Social Intelligence.’ in Trust in Society, ed. K.S.
Cook. New York: The Russel Sage Foundation.