This chapter looks at some of the basic questions concerning the development of scientific knowledge. These are fundamental questions if we wish to carry out a research project. We need to inform ourselves of rules and standards concerning the field of practice that we enter into. The disciplinary field that discusses the conditions of scientific knowledge and the rules of scientific practice is the theory of science. Before talking more about design, I will therefore make a short excursion into some basic questions of the theory of science: What are the characteristics of scientific knowledge? Does this knowledge have a different status than our everyday common knowledge? What is the role of cognitive frames in the development of scientific knowledge, and what is the role of theories and scientific concepts? Are theories to be understood as products of data and observations, or do our theoretical lenses shape the way we see the world? Is there such a thing as objective knowledge about society and the world?

The Characteristics of Scientific Knowledge and Practice

Carrying out a research project means entering the scientific field of practice. The scientific field, like the other fields of society, has its norms and rules for what counts as good and acceptable approaches and practices. In the field of science, these rules and requirements are about how to proceed to develop new or support existing knowledge. To perform quality research that adds relevant knowledge, we have to acquire knowledge in a way that is acceptable in the community of researchers.

What is the hallmark of scientific knowledge? Most experts today agree that the division between scientific knowledge and other knowledge is less fundamental than we are apt to think. There are many types of knowledge and ways of possessing it. For one thing, we can have skills knowledge, for instance being able to ski or cycle. It can be difficult to explain what exactly it is that we do to maintain balance and make a forward thrust when we perform such activities. However, we do know when we succeed in doing the required procedure. Another word for skills knowledge is procedural knowledge.

Another type of knowledge is the knowledge of our own experience, like when we hurt ourselves or are tired. Such bodily knowledge is in a sense subjective. We experience it ourselves, in a direct way. Others cannot have a direct access to our knowledge in the same way as we do. Still, we sense this knowledge as true. We do not doubt that we have pain or feel tired. There are also other forms of experience-based or acquaintance knowledge. What we have learnt about other people’s way of reacting—for instance pupils in a classroom context, clients in a conversation, or patients in a care situation—can lead to an experience-based knowledge that enables us to reflect on and deal with the situation. Everyday knowledge of the kind mentioned here is something we perceive as relatively trustworthy, because we have seen proof of its validity in various situations. We do not usually doubt what we experience, or what we see is working in practice. That does not mean that we have no room for doubt. Sometimes, we can doubt our senses or experiences. However, doubt in itself entails that we accept the existence of a division between true perceptions and those that are not true.

Consequently, it is not so that only scientific knowledge is objective and reliable, while other kinds of knowledge are subjective and unreliable. However, there is another requirement to scientific knowledge: It must be testable. The form of a scientific statement must be such that it claims something is the case or not. We can test a statement with such a form, in order to find out whether we can disprove it. “At noon the sun is in the south” is an example of such an assumption. “Wipe your shoes before entering” is a statement of another kind. It is an imperative in the form of an order or a norm for what you should do or are required to do. For a researcher it is important to understand the difference between these two types of statements.

We can test the first assertion through observations. It is possible to make many observations that apparently verify that the sun is always in the south at midday. Every day, provided it is not too cloudy, observations from my sitting room window are in accordance with the statement. Many other persons can also make similar observations. Still, it is possible to make the opposite observation, at least if we move to the southern hemisphere. South of the equator the sun is not in the south at noon; it is in the north. Thus, we can make observations that disprove or falsify the assumption.

This little example illustrates an essential quality of scientific assertions: Their form must be such that we can test them. Differently said: It must be possible to falsify or disprove them through new observations and data (Popper, 1992/1959).

Three Types of Knowledge

The theory of science usuallyemploys three mainforms of knowledge: Acquaintance knowledge (or familiarity), procedural knowledge, and propositional knowledge (Truncellito, 2007). Propositional knowledge is the form of knowledge that states that something is the case or not. It is a proposition describing a fact or a state. The common view is that scientific knowledge is of this kind. Procedural knowledge has to do with competence or skills, such as being able to drive a car or stitch together a wound on a patient. Familiarity or acquaintance knowledge has to do with recognition based on experience, such as recognising a person in the street or knowing Paris.

An important aspect of research and science is to organise observation and data collection in a way that puts established assumptions and ideas to the test. Together, and through critical discussion, scientists have developed systematic methods directed at making the knowledge and insight we arrive at as reliable as possible. Research is a systematic activity in the sense that it applies systematic and recognised methods and approaches. That is why we need to learn about research design and the choice of method. Scientific method is about concretising the requirement that scientific knowledge must be testable.

Research takes place in academic communities. When doing research, we relate to peers who know how to discuss, evaluate, and criticise what we do. Criticism is an important part of academic knowledge building, and a way to bring knowledge development forwards. We perform this activity in small groups in every research context, as well as in broader research communities both nationally and internationally. Researchers within the same discipline or topic exchange knowledge and experiences across national borders. Institutionalisation of these scholarly communities takes place through conferences where researchers from different environments attend and through journals where researchers in a particular discipline or field publish their scientific articles. Before a scientific journal publishes an article, it is peer reviewed and found acceptable for publication. Thus, peer reviews constitute an authorisation of scientific work, and publication gives a scientific finding or argument status as valid knowledge.

Knut Erik Tranøy summarised these points in a precise way with his definition of scientific activity (Tranøy, 1986), stating that it is

systematic and socially organised search for, acquisition and production of, as well as management and communication of knowledge and insight [translator’s translation].

Simply formulated we can say that scientific knowledge has a form that makes it testable; it is provided through the use of methods found acceptable by the research community and has not yet been convincingly disproven by new findings or criticism from other researchers.

In this way, it is possible to compare the scientific practice field with other social fields that have their own rules for what the purpose of the activity is and how we ought to perform it. Just as there are rules and norms for playing football or chess, or for entering a programme of professional study, it is the rules of the game that constitute scientific practice. We can perform this practice at different levels. There is a difference between children playing football in the schoolyard and a professional football match in the Champions League. My leisure chess game with a colleague is also at another level than the games played by chess champion Magnus Carlsen. Still, the rules and goals of the activities constitute them as a game of chess, a football match, or a programme of professional study. This is also the case in the scientific field. We cannot all perform research at world elite level and compete for Nobel Prizes. However, as long as we follow the rules that constitute the scientific field, we are doing research, which also pertains to the activity of writing a master’s degree or PhD thesis.

Scientific Statements

Let us take a closer look at the two statements cited above. “At noon the sun is in the south” is a statement claiming that something is always the case. “It is now noon and I see the sun in the south” is a less ambitious statement, just mentioning what I see here and now. The difference between these two types of statements is also important in science. The former statement is an example of generalisation, also called a categorical statement. Scientific theories belong to this type. Their form excludes certain experiences, for instance that the sun is in the north at noon. The second statement is an example of an observation statement. We can disprove generalisations by means of observations that contradict the claims of a generalising statement.Footnote 1

The idea that a scientific statement must have a testable form and thus face the possibility for disproval or falsification dates from the theory of science philosopher Karl Popper (1992/1959). He put forward this idea as a reaction to perceptions maintaining that the goal of science is to arrive at absolutely reliable and irrefutable knowledge. According to Popper, all scientific knowledge is in principle uncertain, as we can refute it by new observations. Thus, an observation statement can never completely prove a theory. Further, a wide acceptance of a theory is no final criterion for its validity, as shown by several historical examples. As we know, it was at one time a generally accepted claim that the sun circled around the earth. Astronomical observations led to the development of another theory with wide acceptance today, although it is less straightforward to grasp when assessing it by everyday observations.

Rather than seeing a division between perfectly reliable knowledge and unreliable ideas, Popper’s thinking makes us understand scientific knowledge as more or less robust. When a categorical statement has withstood many attempts to refute it, we will consider it as robust knowledge. The great benefit of Popper’s approach is that it fosters critical thinking in the researcher community, and that its focus on testing may give us better and more reliable knowledge.

In addition to the division between observation statements and generalisations (categorical statements), there are a couple of other differences between statements that it is useful to know. One is the difference between definitions and empirical statements. If we turn the categorical statement “All swans are white” into “Swans are white”, we arrive at a definition. We must assess definitions by their usability, not by their degree of truthfulness. The definition “The horse is a four-legged animal” is not refuted by the observation of a horse born with three or five legs; instead, we conclude that this particular horse has a flaw, a deformity. It is not how a horse should be. We can understand definitions as guidelines telling us something about how things should be. In this sense, definitions are norms; they are not empirical statements. In short, we can say that empirical statements are “be” statements while normative statements are “should” statements.Footnote 2 However, as the examples above show, the differences between normative and descriptive statements are not carved in stone; although definitions have a normative side they also contain a statement that something is something (swans are white, horses are four-legged animals).

Empirical Statements, Norms, and Practice

The greater philosophical discussion about the relationship between norms and empirical statements is not a topic here. Most contributors will agree that the difference is relevant, although not entirely clear. They will also usually believe that we may not infer normative conclusions and practical advice from empirical knowledge alone. For normative advice based on scientific activity, we need three components:

  1. 1.

    What is valuable; a difference between good and bad

  2. 2.

    What we have reasons to do

  3. 3.

    What we should do

The point of the three-piece division above is that a statement saying what you should do (“remove your shoes before entering”) does not follow directly from an empirical observation (“your father has just mopped the floor”). That father has just mopped the floor can, of course, be understood as a reason to remove one’s shoes. However, in order to draw the conclusion about what you should do you also need a norm of the type “you should respect the work of the person who has just mopped the floor”, or component number one in the above list. When scientists take the role of practical advisors in social matters, such advice always has a normative component. That is the case even when this component is only implied.

At the same time, the difference between empirical and normative statements is not entirely clear. All our statements and arguments about the world have a certain element of evaluation. Looking at the empirical statement above, “your father has just mopped the floor”, we find an evaluative element in accepting somebody or something as “father” or “floor”. For most practical purposes, this is not of great interest. Quite likely the two involved in the dialogue above will not have any problems in agreeing on the empirical content of “father” or “floor”. Few people will also have problems with understanding the difference between the statement “Norway is a society with a great deal of equality between people” and the statement “Everybody should have the same wages no matter what work they perform”. The former statement is clearly empirical, the latter just as clearly normative. Being able to distinguish between the two is essential to understand what kind of questions we discuss at any time.

Although there is not always a crystal-clear division between science and normative issues and advice, we need to make clear what practice we are involved in at any given time. Science and research play an increasingly important role in society, and politicians, experts, and others often refer to research when justifying an action or a point of view. Science is without doubt an important supplier of conditions for social practice in a number of areas. Yet that does not mean that science is able to tell us the right thing to do in a given situation. Normative assessments and acquaintance knowledge about a given context will also play a role for practical action.

More on Testing of Scientific Statements

It is a basic rule in the scientific community that scientific statements must have a form that makes them testable. Yet the procedure of falsification testing does not necessarily solve all problems related to the solidity of scientific knowledge. Firstly, it is always possible to save a categorical statement by introducing additional premises. In the above example of the sun, we could add as a premise that the observer stays on the northern hemisphere: “Seen from an observation post on the northern hemisphere, the sun is in the south at noon.” This saves the theory to the extent that there will no longer be observation statements to refute it. Yet the additional premise reduces the scope of the theory, which then becomes less interesting. As we may make other observations on the southern hemisphere, a researcher will be more interested in a general theory that can explain the variation in the sun’s position relative to a given point on earth.

Secondly, we can always ask what is more reliable, the theory or the observation. Our measuring instruments, for instance a compass, may have been defective, or our senses have deceived us. If we trust the generalisation “all swans are white”, and then observe a black swan, we may for instance think that somebody wants to play a hoax on us and have spray-painted the swan black. Or, in the case with the sun, we may think that a local magnetic field has played havoc with our compass, making us confused about the directions.

A third problem concerning the requirement of testing and falsification as formulated by Popper is that it presumes a clear distinction between observation statements and generalisations. What then if our observation statements are not neutral, but in some way or other dependent on our position?

Consequently, there are a number of different pitfalls and problems when we try to develop and test scientific knowledge. First, we will look more closely at the relationship between observation statements and categorical statements seen from an epistemological perspective. In scientific knowledge, we usually distinguish between data on the one hand and theories, concepts, and models on the other. We produce scientific data, also called empirical data, through some kind of observation of the world. In that sense, we can talk about data production as a way of establishing observation statements.

Scientific method is about how to collect and analyse data in a reliable way. This is an important topic for textbooks on methodology, qualitative as well as quantitative ones. We need to be familiar with research methods in order to collect reliable data and make judgements about what conclusions we can infer from the given data. However, methods of data collection and analysis are not the main topic in this book; there are method books for that. This book introduces what we have to do before beginning to collect data.

In what follows, I want to make the reader aware of the relationship between observations and generalisations. How do observation statements and categorical statements, or data and theories, relate to each other? If mutually dependent, how should we understand their relationship? What do we do when drawing conclusions about the world based on theories? What role do our observations play when we formulate, stick to, or change theories?

Our Social Construction of the World

How do we actually acquire knowledge about the world through observations and generalisations? Do the observations come first so that we as a next stage use the observations to create generalising statements? This has long been the prevailing belief in the theory of science. Establishing reliable knowledge was a question of depicting the world as objectively as possible. Said differently, our minds and perceptions were supposed to mirror the world. Then, is this the way we establish knowledge?

Let us look at the image in Fig. 2.1. We will quickly form an opinion of what the image is supposed to represent. However, we do not all see the same thing or form the same notion. Some will see a vase while others will see two faces turned to each other. When we have become aware of the ambiguity of the image, most of us will agree that it may represent one as well as the other of the alternatives.

Fig. 2.1
An image of an optical illusion. It has two faces facing each other and at the same time, an image of a goblet.

A representation—of what?

Full size image

The point illustrated by this image is that what we see depends on our focus. If our focus is on the centre of the image, we see a vase. If we focus on the sides, we see the profiles.

The other point made evident by this exercise is that we focus and interpret observations through cognitive frames and mechanisms that make us recognise patterns that we are familiar with. Strictly speaking, this image is neither a vase nor two profiles. We could also describe it as two coloured fields on a white paper. What happens when we interpret these coloured fields and the space between them as a vase or two faces cannot be understood as a mirroring of the world. It is rather a re-presentation or reconstruction where we actively interpret our visual impressions through acquired cognitive patterns and mechanisms. In fact, it is a double representation. The image represents the vase and the faces while our interpretations and concepts represent the image as either “vase” or “faces”.

A theory of knowledge maintaining that our knowledge re-presents the world rather than mirroring it is what we call constructivism (Berger & Luckmann, 1967; Searle, 1995). Constructivismindicates that our observations depend on our pre-understandings. Pre-understanding in its turn is a product of our personal experiences and the time and society in which we live. Today’s theory of knowledge, however, sees language as the most important element of pre-understanding. Language gives us tools and frames to understand the world (Rorty, 1967; Wittgenstein, 1953). Concepts like “vase” or “face” are meaningful; they constitute cognitive tools we can use to categorise sense impressions. Our cognitive tools also contribute to understanding and classifying scientific data or observations. Creating meaning and order entails that we place our observations in mental “drawers” or categories. As such, our observations are not neutral; they are “charged” with theories and pre-understanding.

Another way of saying it is that we try to achieve a congruence between what we observe and experience, and the conceptions and cognitive frames we work with. If we cannot establish this congruence, the consequences are often that we disregard or fail to see observations that are not in accordance with our cognitive frames; what we observe depends on what frequencies we tune in to.

We develop cognitive frames and conceptions largely through language. It is through language we name and categorise our experiences. We can also say that we “grasp” the world through linguistic concepts. That language is a decisive element in our pre-understanding has another important consequence: neither is our knowledge an objective mirroring of the world, nor is it completely arbitrary or subjective. There is no such thing as a private language. Language is a meta-individual communication tool that exists prior to the individual language user. We express ourselves through language, thus giving others a basis to understand us. As such, language is intersubjective: it is between subjects. Through language, we establish many common perceptions of what is true or right.

Still, some new issues arise if we accept the view of our knowledge as linguistically and socially constructed. How are we to decide whether a theory or an argument is better than another? How do we develop better knowledge? How can we verify theories by help of observation statements if the observations depend on our pre-understanding or our theoretical constructions? These are some of the big and controversial issues in the theory of science.

One point is that we should always understand the question of congruence between observation statements and categorical statements as a matter of coherence between different types of statements rather than a correspondence between reality “out there” and our theories. Language always stands between the world and ourselves. Thus, what we attempt to achieve in our scientific work is a coherence between statements on different levels.

Another point is that even though our observations are “theoretically charged”, that is, dependent on our pre-understanding and our beliefs, they are not necessarily charged with the beliefs and theories that are to be tested in scientific work. Even simple observations depend on an entire system of former beliefs that we cannot test in a research project. There are always things that we take for granted. What we can and should do in scientific work is to clarify our assumptions as much as possible and be explicit about what is to be tested.

I use an example to illustrate the point in the paragraph above, looking again at Fig. 2.1. Even a simple observation like the one we did of the drawing in the figure depends on a set of former impressions that exceed our ability to distinguish between the concepts of vase or face; for instance, that we know what a drawing is and accept the drawing as a way of representing three-dimensional physical objects. We also know what a vase is by knowing something about its normal use, and what a face is by knowing what a human being is. All this is an example of backgrounds and perceptions that we do not question or doubt when judging what the drawing is supposed to represent.

We may also imagine that we do not see the contours in the figure as a drawing, but as a contour behind a curtain with backlight. We may then have imagined that if the black fields would start to move, we must be looking at faces rather than a vase. Let us say we observe for a minute or five before making up our minds about what we have seen. What we have actually done is constructing a testing situation. We do that by specifying what observations would be able to support or disprove one or the other of the two impressions we have. Yet there is a lot we take for granted and do not doubt in this situation, for instance that faces belong to people who can move, while vases cannot move on their own.

This is also the case in scientific work. A team of researchers will always be working with a number of previous beliefs and impressions that are not problematised in a given research project. Theory of science philosophers have given various names to those frames we take for granted in scientific work. They have been called paradigms (Kuhn, 1996), research programmes (Lakatos, 1978), and backgrounds (Searle, 1995), to mention some well-known concepts in the discussion of the role played by pre-understanding in the development of scientific, and for that matter also other knowledge. Their point is that when we produce and interpret data and observations, we do so in the light of such general frames and theories accepted within the frames.

Science and Knowledge Development

Even though we interpret and develop knowledge in the light of pre-understanding, our frames do not completely imprison us. We can develop our knowledge by specifying expectations of what observations will be in accordance with the frames and what observations will not (Gilje, 1987). Let us imagine a researcher who considers research a creative activity rather than a craft, also believing that there are reasons to think that we perform creative activities better in flat, network-organised research teams than in top-managed hierarchies. Based on this perspective the researcher would claim that network-organised research teams are more creative than hierarchical research organisations. This is a categorical statement, similar to the examples presented earlier in this chapter. For such a statement to be tested, the researcher must collect data on research environments and research activities. In order to direct the data collection according to the research purpose, the researcher must also specify what kind of observations should characterise creative research. Is it the number of published articles, the number of patents, the surrounding world’s perceptions of a research environment, the participants’ perceptions, or is it the one, history-making discovery that the researcher should take as proof? Only then can she use the observations for an evaluation of the background statement.

It is still too early to discuss the different ways we can design the research. I deal with that in later chapters. Here, I aim to make the reader aware of what may happen if the researcher arrives at a conclusion in conflict with the general beliefs in his or her community of researchers. The problem is, in fact, about how we accept something as valid knowledge. Let us say that the community of researchers believes networks are the most creative, while our researcher believes her findings indicate the opposite. Reactions among the colleagues could then vary. For one thing, they may doubt her findings, criticise her method or the conduction of the research. Secondly, they may say that this is interesting but in no way reliable; we need more investigations in order to conclude. Thirdly, they may claim that we still have no sufficiently good theories about what makes research creative and that this is a field where more work is needed. Finally, it is conceivable that someone will launch a new theory, explaining why research is more innovative when coordinated through a hierarchical organisation. As shown, the discussion in the community of researchers may take different directions. A community of researchers will have to live with the fact that there are conflicting findings in a number of research fields, often also conflicting theories (Kuhn, 1996). Our knowledge is not organised and certain; there are many uncertainties and anomalies (Gilje, 1987).

At the end of the day, testing and development of knowledge has to do with how good our reasons are for the one or the other belief. We cannot defend as true those beliefs that we have no good reasons or arguments to keep. The discussion and argumentation in the community of researchers and in other contexts will deal with the plausibility of an empirical statement or observation statement, what theories or generalisations it is coherent with, and what theories are most in accordance with the observations we believe to be valid.

Epistemological Positions

We can roughly talk about three sets of positions in the science of knowledge. One is the correspondence theory that sees knowledge as a question of correspondence between the world and our representations of it. Early theory of knowledge believed this question of correspondence to be a question of how thoughts and beliefs mirrored what exists in the world around. This simple mirroring theory proved to be unusable through the attempts of the logical positivists to perfect it, as well as the proofs of the language philosophy that we represent the world through language rather than mirroring it. This was the beginning of a coherence theory of knowledge, where knowledge is a matter of coherence between statements at different levels rather than a correspondence between statements and reality. In such a coherence-theoretical context, we can reformulate the question of true knowledge as a question about the relationship between justification and truth. The third position of the theory of knowledge is the pragmatic or practice theory of knowledge , seeing truth as a question of what functions in practice. This position opens up for an understanding of knowledge that goes beyond statement knowledge, and thus for a discussion of the relationship between statement knowledge and the other forms.

According to GunnarSkirbekk (2004a, 2004b), when we leave the simplified conception of knowledge as a mirroring of the world around us, the question of truth again arises, now as a question of how we justify our statements about the world. What are our reasons to maintain that something is true or not? In principle, we can view every justification or reason as interim and incomplete. It depends on time, place, and person. A justification we see as valid at one point in time may be lost in another time.

What is truth? This is a difficult question. Some science philosophers regard truth as that which can be justified under ideal conditions (Apel, 2003; Habermas, 1999). Others are sceptical to such a view of an ideal truth beyond time and space. The most radical theorists will maintain that we do not even need concepts of truth beyond that which can be justified (Rorty, 1999, 2000). Others again will emphasise the critical element—truth is something we can approach through critical approaches, testing of argument, and development of new and better descriptions of phenomena and connections. Moreover, we cannot delimit the question of truth to the propositional knowledge of science. Action-related knowledge is also a “window to the world” and a point of departure we can use to develop statement knowledge (Skirbekk, 2003, 2004a; Wellmer, 2003). We have a lot of everyday knowledge developed through our own experiences and practices, which we find no reason to doubt.

The Contested Truth

The realisation that our knowledge is socially constructed has led to a debate of whether it is possible to know if there exists an objective world that we can grasp with our knowledge. If we depend on language and our cognitive constructions to experience the world, how can we know that the external world is real? Some have drawn such drastic consequences from social constructivism that they argue there is no such thing as reliable knowledge about the world; everything is our subjective experience. The so-called postmodernists are the most prominent among those who have maintained this view. From this view derives that we ourselves, as the producers of words and texts, are also the producers of reality.

A dilemma concerning this view is that it cannot be true if what it claims is true. If we maintain that there are only subjective experiences and no objective knowledge about the world, this assertion cannot be true either. In academic terms, we call such a theory self-referentially inconsistent. By presenting itself as knowledge of the non-existence of objective knowledge, we cannot take it seriously in terms of knowledge. The logical problem is of the same kind as this more humorous one: If a person claims that all Londoners lie, and this person is from London, the statement is self-referentially inconsistent because if it is true that all Londoners lie it follows that it is a lie when a Londoner claims that all Londoners lie.

Another way of approaching this issue is to state that the doubt whether we can possess objective knowledge about the world assumes that there is such a thing as objective knowledge. In doubting that objective knowledge about the world exists, we assume a division between such knowledge and knowledge that is not objective (Malnes, 2012). We do not need to believe that we possess objective knowledge about the world. We may very well think that our own knowledge is incomplete and insufficient. What we defend is the logical assumption that something like objective knowledge can exist. We may very well regard completely objective knowledge as an ideal it is difficult or even impossible to achieve in practice. However, we still hold on to the ideal. From this perspective Popper’s belief in testing and falsification as a scientific approach makes sense. We cannot regard our knowledge to be final and reliable; it is always possible that new observations can contradict our experiences and so disprove our beliefs. Yet through systematic testing we can at least develop our knowledge to make it become better and more reliable.

In our practical world of action, we take a lot for granted. When I wake up in the morning, I look out of the window and check the weather to decide whether I should bring an umbrella to keep dry when walking to work. I take it for granted that I obtain knowledge about the weather by this simple observation. If in doubt, I can check the weather forecast or bring the umbrella anyway. However, I never doubt there is something out there that I can observe. In our everyday actions, we take at any time many things for granted. Otto von Neurath talks metaphorically about knowledge development as similar to rebuilding a boat while on the water (Neurath, 1932). We cannot change all the parts at the same time because then the boat will sink. Transferred to the matter of knowledge: If we doubt everything at the same time, our boat will sink in the sense that we are no longer able to orient ourselves in the world. Still, we can focus on one weak point at a time and do something about it. This is where systematic scientific approach comes in. Using scientific methods requires us to clarify what it is that we take for granted, what we are testing, and how we are doing it.

Another approach to this discussion is to differentiate between theories about how we gain knowledge about the world, and theories about what the world is like. This is equivalent to the philosophical division between epistemology and ontology. Epistemology is the study of knowledge and knowledge formation, while ontology is the study of what the world is like and what is in the world. In the sections above, I have dealt with epistemology or the philosophy of knowledge. The main point has been that we construct our knowledge linguistically and socially, but these constructions are not arbitrary. It is possible to assess whether an argument or proposition is better or more probable than another. The theory of social and linguistic constructivism is thus an epistemological theory, having to do with how we achieve knowledge about the world and what are the conditions for knowledge production. It is not an ontological theory, that is, a theory of what is or is not in the world (Ferraris, 2014; Hacking, 1999).

From an Understanding of Knowledge to Social Science Practice

Scientific knowledge is knowledge of statements and propositions and should have a testable form. We test general statements by comparing them with observations. Yet such an approach is not without pitfalls, because our observations depend on our frames of understanding. Language as well as our cognitive frames is between the world and us. Thus, our knowledge is socially constructed. Yet social constructivism has to do with the way we gain knowledge; it is not a proposition that reality too is an arbitrary construction. How to develop valid and useful knowledge under these conditions is a general challenge as well as a challenge for science. In the following, I will describe how social scientists proceed in practice when they attempt to develop new knowledge about a social phenomenon.