Concepts of science and trust

1.2.1. Science and sciences

The main reason why definitions are important is that they help us to avoid debating at cross-purposes. The key words in the title of this chapter are obviously science and trust. What do they mean?

How have science and its role changed in contemporary society? The answer will to some extent depend on how 'science' is understood. Discussing the views on science by John Ziman (2000), Raymond Spier argues (2002, p. 237) that though Ziman holds that: '... science is too diverse, too protean, to be captured in full by a definition' (Ziman 2000, p. 12), he nevertheless characterises science in a number of ways: it is 'a social institution'; 'Science is one of a number of somewhat similar institutions, such as organized religion, law, the humanities and the fine arts'; 'The peculiarity of science is that knowledge as such is deemed to be its principle product and purpose' (all Ziman 2000, p. 4); 'Science generates knowledge' (Ziman 2000, p. 5); 'By its very nature, science is a complex system. It cannot be understood without an explanation of the way that its various elements interact' (Ziman 2000, p. 7); '...science has to be thought of primarily as a heterogeneous actor network ...'; it is '...a sequence of refutable conjectures, or a bundle of research traditions, or a problem-solving computational algorithm' (all Ziman 2000, p. 8).

32 In the present context, science will be used in a wide sense, like the German 'Wissenschaft'. If science is not limited to natural sciences, there is obviously room for many ideas about what characterizes different kinds of research. This has been widely discussed in the history and philosophy of science from Dilthey to Habermas and von Wright. The idea of falsifiability – vigorously launched by Karl Popper – is an important ideal in many of them, but as to the organization, financing and role in society there are many important differences between the natural, social, behavioral sciences as well as the humanities, which cannot be discussed here.

However, in his response to Spier's reflections, Ziman (2002, p. 254) makes clear that he is not limiting 'real science' to the natural sciences: 'In spite of Spier's assertions to the contrary, when I referred in the book to the 'human sciences' or the 'social sciences' I thought it was absolutely clear that I regarded these as being as much parts of 'real science' as physics, chemistry and biology.'

1.2.2. Different definitions of trust

The literature suggests a number of definitions of the concept of trust. For example, it is defined in terms of good will, public good, certain beliefs, expectations, reliance, favorable actions, or combinations these concepts, each of which obviously can be understood in many ways.

There are several texts with references to the traditional definitions of 'trust' – without quoting these definitions, probably on the assumption that we all know which they are and that they anyway are equivalent. It is far from clear that these assumptions are well founded, so here are some examples of definitions, which seem not to be equivalent:

'Trust has been defined as the individual's belief that the subject of trust will behave in a favourable manner to the individual, or at least in a manner that will not be harmful to the individual' (Lapidot et al. 2007, p. 17, referring to Gabbetta 1988). Lapidot et al. (2007, p. 17) continue: 'Mayer et al. (1995) have extended this definition [by Gambetta] and added the willingness of the individual to be vulnerable to the actions of the other based on the expectation that the other will perform a particular action important to the trustor'.

Robinson (1996, p. 576) defines trust as a person's 'expectations, assumptions, or beliefs about the likelihood that another's future actions will be beneficial, favourable, or at least not detrimental to one’s interests'.

More complex characterisations of trust also exist in the literature, for instance by Hardin (1999).

He argues that trust requires (1) expectations about the likely behaviour of the person(s) to be trusted, (2) beliefs about their motivations, (3) beliefs about their capacity to act according to their motivations, and (4) beliefs that those trusted are seeking to act in the interest and for the good of those who trust.

The general idea is thus that trust is an attitude of the kind described by Robinson (1996), based on expectations, which in their turn are based on past performances. But if the expectations are not met, this may give rise to frustration, disappointment and actions. In practice, this may suggest using one of two main strategies, 'exit' or 'voice', described in other chapters of this report.

In the context of trust/mistrust in science, the exit strategy may mean: losing interest in reading about scientific results, less space in the media being devoted to scientific achievements, dropping out of doctoral programs, leaving research projects, making a living in other ways than by research, and so forth. If there are alternatives, it is possible to exit one option and select another.

But if there are no alternatives, this does not work. The exit strategy has to be replaced by a voice strategy: to protest and, for instance, demand stricter regulation.

33 1.2.3. Trust defined in this report

The definitions proposed or used in the contributions to this report have somewhat different scope.

They are intended for different areas of application. Some of them are also rather vague, others more precise and specific. In different contexts different definitions may be fruitful. But definitions have consequences, and it is therefore essential to be clear about the definitions chosen and their implications.

Boda uses the following definition: 'Trust is defined as an expectation that the given institution will produce positive outcomes'. This is a rather open definition in terms of what kind of positive outcome and for whom. There can be contexts in which rather open definitions are fruitful or where a more general concept is preferred.

For example, in her contribution, Morales is focusing on the more general concept of political confidence as she writes that it incorporates 'an expectation of competent role performance and of fiduciary obligation and responsibility'.

Pitlik focuses on individuals rather than on institutions: 'The main focus in the context of the present chapter is on trust as the general belief that people do not cheat, shirk, or act otherwise opportunistically in social interactions'. This definition specifies some of the means to undermine social relationships. Other definitions focus on the goal, i.e. to facilitate social relationships.

For instance, Hofmann characterizes trust as follows: 'Trust is commonly understood as a facilitator of social relationships'. This definition has a specific scope or area of application where it probably will work well, whereas in the context of trust in governments or the public sector other definitions may be suggested and argued for.

Wright notes in his contribution: '…for the purposes of this paper trust is defined as the extent to which the citizen has confidence that others (such as companies, intelligence agencies and government departments) will act in the public interest as opposed to narrow corporate greed or political interests.'

This definition differs from the one proposed by Robinson mentioned earlier in at least one important respect. Actions that will be in the best interest of the trusting person need obviously not be in the interest of the public good.

1.2.4. More than one concept of trust?

Thus, it seems clear that we need to be open to the possibility that there are several concepts of trust with unclear demarcation lines. Mechanic (1996, p. 173) distinguishes between two levels of trust: interpersonal and social. He argues that trust in persons is an intimate form, 'deriving from earlier experiences with family and other caretakers', while social trust 'in contrast, is more cognitive and abstract, and typically is based on inferences about shared interests and common norms and values'. However, he provides no explicit definition of these concepts.

To quote from an earlier text of mine (Hermerén 2013b, p. 63): 'In the philosophical discussions of trust it has been emphasized by several writers that accounts of trust that do not distinguish it from reliance are too broad, whereas (good) will-based accounts appear too narrow and take as paradigm case interpersonal relations rather than trust in governments and institutions,…'.

Discussing public trust in health care, Calnan and Sandford (2004, p. 96) point out that public assessment of trust tends to address the trust of care at the micro level. In the view of these authors, policy makers concerned with the erosion of public trust need 'to target aspects associated with patient centered care and professional expertise'. They also found that levels of distrust 'particularly with how the service is run and financed, are high'.

These suggestions can also be translated into discussions of eroding public trust in science. There are different levels which need to be looked into, not just cases of distrust of individual scientists

34 due to, for instance, allegations of fraud and/or misconduct. Aspects of the system of funding, reviewing and rewarding scientific research also need to be considered.

Thus, mere reliance is not enough for trust. It may be necessary but not sufficient. Trust may also include, as already mentioned, a disposition of those who trust to react negatively or with resentment if their expectations are not met. This is relevant to Enders' discussion (2013) which focuses on governmental reactions to alleged distrust of higher education institutions. Words like 'alleged' or 'perceived' incidentally raises epistemic issues about the grounds for trust or mistrust in science (Hermerén 2013a, p. 7). What is to be regarded as good reasons for trust and mistrust in this context?

Trust is usually analysed as a three place relation by philosophers, forward-looking 'A trusts B to do X', or backward-looking 'A trusts B to have done X'. Here A and B can stand for persons, institutions, organisations etc. The point of introducing the third variable is to be more specific about trust. If clarity is desired, it is essential to avoid talking about trust in general. Questions in surveys need to be specific; trust seldom extends to all domains of interaction. Trust can be graded, and it is tested repeatedly in different situations.

In their thorough study, McKnight and Chervany (1996) begin by identifying the lack of consensus about the meaning of trust and suggest that this is an obstacle for research on trust. They propose a classification system for types of trust, and definitions of six related trust types that can be combined into a model of trust. Going through articles and books on trust, they identify types of attributes to which trust refers in these articles and books (such as competence, expertness, predictability, goodness, etc). They also compare sets of definitions of trust in dictionaries. All this is useful and indeed essential for any in-depth conceptual analysis of trust. For the aim of this publication it is sufficient to be aware of the existence of different definitions which often serve different purposes.

What exactly is being trusted? In each context we may ask: are we focusing on trust – or erosion of trust – in systems or organisations or in persons working in these organizations? Several different levels can be distinguished, as I proposed in an earlier text (Hermerén 2013b, p. 64):

1. Research/higher education generally;

2. Research/higher educational institutions, such as universities;

3. A particular profession or activity as such, for instance social science;

4. Individual researchers.

The relations between these levels are not quite clear, in particular the extent to which information about trust or mistrust in relation to one level can be generalized to other levels. The difficulties are obvious in drawing conclusions from surveys focusing on trust in 1 or 2 to trust in 4 – and conversely. This seems to apply to policy-making and/or funding institutions as well.

Obviously, there might be lack of trust in the reasons why scientists chose certain problems, in the methodology they use, in the conclusions they arrive at, and in the way their results are applied and commercialized. It can also involve funding of science, education at different levels in science, or the will and capability to take responsibility for adverse effects of applications of scientific discoveries. If indeed trust is eroding in some aspect of scientific research or the institutions carrying out or funding it, we may want to ask ourselves: are there any signals indicating this?