Concept Analysis as Empirical Method

Bas Levering

Bas Levering, PhD., Professor, Universiteit Utrecht, Utrecht, The Netherlands

Abstract:

Although there is not much support anymore for a clear cut distinction between object-language and meta-language, the view is generally accepted that the outcomes of concept analysis concern knowledge of the structures of language and not knowledge of reality. So it is rather risky to present concept analysis as an empirical method. Rules about language use, however, play an important part in what people do and how they behave and thus inform about specific parts of social reality. The so-called ‘thick concepts’ provide a special access to the way people interact. In this article, several examples of the analysis of ‘thick concepts’ are presented. In some cases, outcomes of concept analysis turn out to be strictly language and cultural bound; in other cases, conceptual knowledge appears to transcend those boundaries. The fact that empirical research is often nothing else than a form of language analysis calls for a tuning of conceptual-analytic and empirical research.

Editor’s Note: This article has been previously published in Dutch:

Levering, B. (1998). Conceptuele analyse als empirische methode. Pedagogisch tijdschrift, 23, 285-302.

Levering, B. (1999). Conceptuele analyse als empirische methode. In B. Levering & P. Smeyers (Eds.). Opvoeding en onderwijs leren zien. Een inleiding in interpretatief ondrezoek. Amsterdam: Boom, pp. 257-269.

Citation information:

Levering, B. (2002). Concept analysis as empirical method. International Journal of Qualitative Methods, 1 (1), Article 2. Retrieved DATE from http://www.ualberta.ca/~ijqm/



Many people have distanced themselves from the idea that there is a separation between language and reality. Linked with the idea of separate worlds—one with words and concepts, and another with matter and events—is the idea of a dualism between methods, an idea which should also be condemned. A separation between analytical methods, which provide information about language, and empirical methods, which inform about reality, is untenable. Formulated thus, my viewpoint can hardly be contradicted, but I want to risk taking it further. From the premise that analytical and empirical methods cannot be strictly separated, I conclude that analytical methods must be empirical methods up to a point. The opinion that the external world can be analyzed through an analysis of language, from an easy chair in an armoured house as it were, will surely seem provocative. Most of the qualitative research methods seem to derive their superiority from the fact that researchers close the door on their own "wild ideas" and enter the "real" external field of research. Intriguing thoughts crowd in upon the reader. "To what extent does language analysis cover the same area as the analysis of reality?" and, conversely, "Does empirical research indeed produce more than the knowledge of language?" These are the questions that will keep me occupied throughout the argumentation in this article, before coming to a conclusion.

Let me first give the reader an illustration of the way something that has long been considered an analysis of reality appears to be, on deeper investigation, an analysis of a word’s meaning.

It had always been a strange story to me how that powerful, unsinkable ship on her maiden voyage from Southampton to New York, on the 14th of April 1912, had collided with an iceberg. I could easily imagine that "unsinkable" did not really mean unsinkable. While we know that the number of lifeboats was insufficient by far, pride and honor clouded tales of the ship’s demise. There remained a veil of secrecy that impaired our insight into very last hours of the ship: the stories about the exceptional discipline on board; the fact that there was no panic, whatsoever; "women and children first, in all possible quiet."

Perhaps this was not the issue governing my life at the time, but it was absolutely marvelous when, in 1986, to commemorate the ship’s 75th anniversary, Michael Davie’s book The Titanic: The Full Story of a Tragedy appeared. The book’s data were collected by method of oral history, and although now highly popular, it was used there for the first time. After 75 years there were people still alive who had lived through the tragedy, making it possible to question survivors about their experiences. Michael Davie himself had been fascinated by the unimaginable sang froid that crew and passengers had shown during these horrible moments. What was the source of that remarkable courage which we have been told about in the stories? When Michael Davie puts that question to one of the survivors, he reports the following: I don’t think it was a matter of some special kind of courage, because none of the persons on board thought we would go down. If we had thought so, we would not have made jokes, such as: if you come back tomorrow, you will need a ticket.

On reading this, it became instantly clear to me. In my mind’s eye I saw the men hanging over the railing and heard their comments as they were waving goodbye to the people who had taken the few spaces in the lifeboats. Indeed, this had not been a case of being courageous, and actually I had always had an inkling of this, but the exceptional way in which the interviewee told his story verified my suspicion. He did not provide the interviewer with empirical information, so it seems, but with conceptual information. The situation, as reconstructed by the interviewee, was that because they thought the ship could not go down, this could not be a case of courage among the persons on board. For in order to be courageous, it is necessary to be at least slightly scared. To be more precise: ‘courage’ supposes ‘danger awareness.’ A person who hastens to help somebody else, in the knowledge that he or she is also in danger of falling through the ice, can rightfully be called courageous. But somebody who acts in the same way, while thinking that the ice can easily carry their weight—even though this is not true in reality—can hardly be called courageous. Likewise, the person who thinks it courageous of a child courageous to just run across a dangerous street without even looking does not know the meaning of courage.

Although we now have the Titanic’s situation clearly in our mind’s eye, we still do not know what to call it. We understand the almost serene peace and quiet shown by people who should have been in agony. We now understand why there was calm when there should have been panic. So the conclusion must be that, while it is now obvious that the people on board the Titanic were not courageous, it was not easy for the interviewee to precisely define their mental attitude. The mood seems to have been a slightly euphoric one (which is congruent with cracking jokes), and this assured that the gravity of the danger confronting them did not strike home. But this is not mentioned as such by the informer; we have to deduce it. The fact that it is so difficult to find the right word to describe the behavior of the people on board makes it understandable that the witnesses’ reports from those days have created a confusing myth. Somewhere between the report and its transference into words, intrepidity must have been confused with undauntedness.1

On reality and social reality

At the end of the nineteenth century, the German philosopher and historian Wilhelm Dilthey made a distinction between the natural sciences and the humanities. This distinction was based on the view that natural reality is fundamentally different from social reality. In the social sciences, methods used are based both on science and on human sciences, or rather both causal/analytical and interpretative methods, so various interpretations of social reality are used in conducting research in this field. A survey of those differences in the image of reality and the portrayal of mankind may provide a keen insight into one of the most fundamental methodological controversies in social sciences, the controversy between explanation and understanding.

The causal/analytical methods have as their starting point the belief that social reality is controlled by laws. Knowledge of these laws guarantees the predictability of human behavior, because we know its causes and that is what social science is all about. Knowledge of the individual human’s deliberations of whatever kind is totally superfluous for the explanation of the individual’s behavior. This behavior itself is seen as determined by external factors without intervention. If the frustration/aggression hypothesis were to be based on a law, each expression of frustration would lead to a form of aggression for each human individual.

The interpretative methods are based on the notion that in the social reality people’s intentions play an important part. If we really want to know what is happening, we will have to know people’s intentions. Based on our knowledge of intentions, we can understand human action. This seems, however, not enough to understand human behavior. If we know that a man wishes to catch the 16:15 bus, for example, we understand why he is making haste at a certain moment. But knowledge of intentions alone is not enough to predict people’s actions.

In the opposition between a causal/analytical approach and an interpretative approach to social reality, we see that explaining is posed opposite understanding and behavior is posed opposite action. In this rough opposition, many fundamental knowledge-theoretical and anthropological issues have been included. It is an opposition between the knowledge of the so-called "hard facts" and the soft interpretation of intentions. It is a matter of determinism as opposed to indeterminism. In this schema, the idea that human behavior can be explained and predicted is roughly based on the concept of external factors being caught in an accidental cohesion, and the idea that human actions can be understood, but not predicted, is based on the concept of freedom.2

To counterpoise behavior and action opposite determinism and indeterminism is indeed too crude. It is possible to distinguish various intermediate stages between behavior and action, which give the pallet of human behavior a more variegated appearance. This also assures that the "yes or no" situation between the causal/analytical and the interpretative approach may be overcome. Possible intermediate forms are: reflex behavior, instinctive behavior, impulsive behavior, customary behavior, routine behavior, rule-governed behavior, rule-governed action, and deliberate action.3 It is questionable, though, as to whether this continuum (i.e. the sequence of behavior—action) can indeed be placed on a straight line running from less to more freedom. All sorts of theories are at issue in the various intermediate forms between behavior and action, placing them in a position to the left or the right of, and under or above, the direct line of the continuum. This does not make the distinctions less fertile for my purpose.

By naming intermediate forms along the behavior-action continuum, we introduced, as it were, reality as a social reality. The two extremes—behavior and action—form far-reaching abstractions. Human behavior that is projected into a network of causal factors does not differ from the "behavior" of physical objects. But it is just as much alienated from social reality to make human actions completely dependent on the individual’s interpretation, ideally speaking. The intermediate forms between behavior and actions can only be understood if they are related to terms like tradition and customs, and these terms are specifically bound to social reality and have no meaning in physical reality. These concepts, to us, are indeed characteristic for reality as being social reality. In this context, the concept ‘rule’ is absolutely crucial—that is why I will consider this more closely.

There is a sharp contrast between rules and laws. As we have seen above, we can explain and predict human behavior based on the knowledge of laws. Once more, if the frustration/aggression hypothesis would imply a law, each frustration would result in aggression. Because we are familiar with the rules, we can understand human "behavior", but we cannot predict it. The everyday saying that the exception proves the rule tells the tale in this case. If we see someone step on the brakes while approaching a red traffic signal, we can understand such "behavior" 4 based on the relevant traffic rules. But based on the rules, we can not predict this behavior. People do not always stop at a red traffic light. People may break the rules. Based on the knowledge of laws we may give universal explanations; our understanding based on the knowledge of rules is context bound5.

There are two kinds of rules. The first kind are known as regulative rules and are the kind of rules that regulate existing practices. An easily appealing example is indeed the traffic rules. Traffic, be it automobile or pedestrian, is a common occurrence among people that has long been in existence, and traffic rules regulate the orderly procedure. Regulative rules are arbitrary. It is possible to conduct discussions about the advantages and disadvantages of left or right traffic combined with right of way for the right, but what is essential is that there are rules. The same holds, of course, for etiquette.

The second kind of rules are known as constitutive rules. Constitutive rules not only regulate existing behavior, but also create a possibility for new kinds of behavior. An easily understood example of constitutive rules are rules of play. Without the rules of chess, moving chessmen about on the board has no sense whatsoever. The game of chess is created by its very rules. It is true for all kinds of social phenomena that they derive their meaning from rules. Voting behavior, for example, may be researched causal/analytically, but we can only understand people going to polling places and pressing the voting computer’s knobs if we know what elections are, what democracy is, and so on, as defined by the rules of what constitutes a democracy.

Rules form the basis of social reality, not in an ontological sense, but in the sense that people can orientate themselves on rules in their actions, and social scientists understand the social reality by means of understanding the rules. What social reality stands for does not correspond with the rules, but knowledge of the rules gives us entry to the social reality. I could go on much longer about all kinds of aspects relating to rules, for instance about the fact that rules are context-bound, for rules may have a very locally determined meaning. Some social rules, though, seem to hold for all cultures and for all times. I should also address, for instance, the arbitrary character of rules, because it seems as if some rules may be easily, even whimsically, changed, while others seem to be eternal and immutable. About the matter of how rules come into existence, for example, some rules seem to have been created on purpose at a certain moment, while others seem to be the result of a process of development without any human intervention and implemented as unspoken norms. But I shall deal with these aspects later, within the context of the status of certain concepts. Enough for now about the difference between reality and social reality, which was necessary to be able to take the step towards the conceptual analysis which is my real concern.6

Another language, the same meaning

In the area of concept analysis, everyday language gives access to reality in the same way as, in the area of phenomenology, everyday experiences give access to reality. At first sight, language seems to have its limitations, because we know and have experiences which are impossible to put into words. In the case of those people on board the Titanic, for example, I have not been able to find words to characterize their doings. Sometimes a word can be found for a certain concept in one language, while such a word is lacking in another language. Mitfreude (rejoicing with others), for example, is a slightly archaic German word for a certain feeling, the equivalent of which word is lacking in Dutch. The Dutch word medelijden (‘mede’ = mit = with; ‘lijden’ = Leiden = suffer[ing]) and the German word Mitleid linguistically have the same roots; the words express the same recognizable emotion. From the fact that the Dutch language knows no equivalent of Mitfreude, we cannot deduce that the Dutch do not know the emotion. Of course Dutch people know the feeling of joy for someone else who passed a difficult exam. In other words, the existence of concepts which are language bound and language/context bound does not guarantee the availability of words covering these concepts.

Our understanding of concepts is acquired by our living with them. Think, for example, of the case of the incompatibility of colours, which is defined as the fact that in the same place, at the same moment, only one colour can be visible: one thing cannot simultaneously be both red and yellow, for example. In the first instance, the concept of incompatibly of colours seems to have little to do with the way of acquiring knowledge. The fact that at the same place only one colour can be present does not seem to be the result of a process of social definition, but a characteristic of things (in this case: colours) themselves. However, if we take a closer look at the way we acquire a concept of colours, we see that a child indeed acquires the concept of the incompatibility of colours in the same way as he or she acquires knowledge of the colours themselves. "This is red," we say to the child, "and that is yellow." The concept of the incompatibility of colours has been included in the fact that we define colours ostensively. Indicating a colour by demonstration, the expression "This is yellow" can only have a meaning if the colour cannot simultaneously be red, blue, green, or any colour other than yellow. When children begin to grasp the concept of colours and try to provoke us by mentioning a wrong colour on purpose, by saying: "This is red, is it not" (when the colour is yellow), we correct this by introducing the concept of incompatibility. Not, of course, by mentioning the word incompatibility, but by saying: "You know that is impossible, because it is yellow, and you know it!" The fervour with which adults will defend their choice in naming a colour, for example, illustrates how arbitrary the rule is which determines the colours’ incompatibility. The controversy "This is yellow," "No, this is red" can only make sense if the starting point is that "it can only have one colour."

It is possible that there are differences of language-context in discerning colours. Welsh, for example, knows only one word for blue and green: glas. Russian, though, has two words for blue: sinij and goloboj, meaning respectively ‘navy blue’ and ‘celestial blue.’ It becomes evident that the incompatibility of colours is not bound by language context, from the fact that one is obliged to choose between navy blue and celestial blue when naming something blue (Van Delft 1997). Incompatibility of colours seems to be a matter of a universal quality belonging to colours. In the case of a meaning, which is bound by its usage, it makes no sense to use the concept universality, however. An understanding of colour which does not suppose the colours’ incompatibility is not unthinkable either, but also in that case, the lack of this quality would be contained in the use of words denoting colour, and would be acquired by using it.

It is of course evident that much more is acquired by using a certain word than simply the meaning of this word or concept. Through saying certain things in a certain way, or by definitely not saying them like that, personalities are formed and emotional systems are established. That the British are more introvert than the Irish or the Dutch, in general, has to do with rules about what one is supposed, or not supposed, to say in conversation. Each culture knows its written, and specifically unwritten, rules concerning these customs. Sometimes there are attempts to regulate people’s behavior by issuing such rules. Also in The Netherlands, the word ‘manpower’ has been replaced, in certain circles, by ‘human power’, but in the United States of America, the regime of politically correct usage is much more dominant than in The Netherlands and Belgium and linguistic shifts such as this are more commonplace.

I want to demonstrate that we can acquire empirical knowledge of social rules by means of concept analysis. I am completely aware of the fact that this way of acquiring knowledge has its limitations. Certain forms of knowledge of social correctness indeed require a direct confrontation with people, like, for example, the market vendor discussing on Amsterdam local television that nowadays you cannot afford to say everything to anybody any longer: "Not long ago I had a lady in my market stall, who wanted to buy a handbag from me. She says: ‘Can this bag stand the rain?’ I tell her: ‘But of course, lady, but you cannot go swimming with it’ and off she is, mad at me!" (Abrahams, 1998) (Abrahams is a well-known Dutch columnist in a leading national paper, NRC-Handelsblad).

Not all words and concepts are equally suitable for giving us access to empirical social reality, in as far as concept analysis is capable of doing so. We prefer the so-called large concepts to give us access to social reality, because they indeed tell us much about the interaction between people. They are the opposite of general concepts, such as ‘educating,’ ‘teaching,’ ‘learning,’ and ‘training,’ and while these concepts equally refer to interactions between people, they actually only lead to theoretical discussions. That is why general concepts seem to have become so far detached from reality that they have actually lost their empirical content.7 Examples of large concepts are: ‘deceiving,’ ‘promising,’ ‘courage’ (yes, indeed), ‘forgetting,’ ‘shyness,’ and ‘jealousy.’ 8 Let me work out another example.

It is not everyday that, as a concept analyst, one is presented with a conceptual problem, but in June 1997 a gentleman from Nieuwegein (a city near Utrecht) wrote me the following: "In many languages, children are described as ‘sweet.’ Up till which age does that occur in The Netherlands? And please do not consider this a trivial problem." Words which are suitable for the concept analysis I am aiming at seem to belong to a category of older words. That is quite obvious. Older words have established themselves in language, in many ways, in the course of time. Their meaning has become more differentiated and clear by long usage. That holds for a word like ‘sweet’ in any case.

Sweet children are children that are no trouble for the people who take care of them. It seems as if only young children are ‘sweet.’ Being sweet is a way of being rather than a self-chosen form of behavior. A child is sweet, it can not act sweetly. Age limits are hard to define, but, in Dutch, it is strange to speak of sweet children when they are over five years old. One would say sweet is being good, but it is not the opposite of being naughty.

Concept analysis soon outgrows pure language analysis, if such a concept exists at all. The fact that we reserve the word sweet for young children lies embedded in the language. When we want to characterize child-like competence, our normal experience seems to be at stake from the very beginning. Experience teaches us that ‘sweet’ (meaning good) babies may change into most troublesome children in just a few months’ time. This experience gives body to the difference between being sweet and acting sweetly. The moral connotation ‘sweet is good’ lies contained in the phrasing, but is open to discussion.

The first thing that strikes me is the fact that a word like ‘sweet’ indeed gets its meaning in relation to an adult person. Since young children can be a lot of work and continually require attention, it’s a bit of luck to have a sweet child. If children are playing sweetly, they are completely self-contained and they do not need us at all. Also troublesome children—and this is something we know from experience and not from language usage—can be playing sweetly occasionally. One can also give a child something with which it plays sweetly for some time. Young children actually have no objection at all to that slightly manipulative behavior. Adults, though, experience a sop (literal translation from the Dutch is a ‘sweetkeeper’) for themselves as negative, because of its manipulative aspect—they are aware of being fobbed off9 in that case. It is quite common for an adult (in the author’s home country, Holland) who is given a task requiring concentrated attention for quite some time to use the expression: "That will keep me sweetly occupied for quite some time." What the speaker probably implies is that he or she accepts the task freely, but still has the feeling that he or she cannot refuse.

As soon as children can sit quietly of their own accord and are able to concentrate and make the conscious choice to be no bother to the adults, they can no longer be sweet. To be ‘sweet’ is not a choice, it happens to you. Whether or not being sweet is a good thing is indeed open to discussion. Likewise, it is a point of discussion as to whether or not we still appreciate those sweet children of olden times. It remains undecided whether or not the toddlers and under-fives now are the same since our nursery schools have merged with elementary schools (children now start at age 4 in Holland), and since the day nursery has been replaced by daycare. We do not like goody-goody children so much any more; where once this was seen as a virtue, nowadays we want our children to be assertive from the first moment.10 And I should add the following: I know very well that our wish for children to be good has its origin in the adult’s self-interest rather than the child’s interest. The line from a Santa Claus song: "Sweets for the sweet children, the rod if they have been naughty" is based on a false opposition. Sweet is indeed opposed to troublesome and a child capable of being naughty cannot be sweet any longer. Being sweet, practiced by an older child, is a pose at the very least.11 This is clear from one of the findings of a qualitative research project done in 1997 by cultural anthropologist Mieke de Waal, cited in the leading left-wing morning paper Volkskrant. De Waal studied children in groups five and six (9 and 10 year olds) of a non-denominational school in Geldermalsen (an average provincial town in Holland). She mentions, as an appealing attention-getter, the word-battle about bullying and teasing. "This yammering is so important, because, if it is a matter of bullying, the bully will have to change. But if it is a matter of teasing, the victim should not be so childish" (Hendriks, 1997). Apparently arguments can also be settled, in an informal setting, by establishing which concept applies.

Language usage provides no information on the issue of which action must be qualified as bullying and which as teasing. That is what the whole word-battle is about and the dispute can only be settled afterwards by making inquiries. I am only concerned with what has been presupposed in this word battle. The fact that there are forms of interaction which are morally acceptable, yes or no, and the fact that qualification in terms of bullying or teasing is considered to influence people’s doings are not the result of empirical research—observation and interviewing as done by de Waal. These were already contained in and accessible in the language itself before de Waal even started her research. For understanding the relevant differences between bullying and teasing, the concept analysis of thick concepts suffices.

Another language, another meaning

Not enough can be said about the language-context connection. A distinction as meaningful as the one between bullying and teasing is often to be found in exactly the same way in other languages, and it seems as if there is only one language worldwide. But in other cases, we not only seem to speak different languages, but even to live in completely different worlds. There are words which are untranslatable, because they derive their meaning from the landscape and the climate from which they were generated. Attempts to translate such words are doomed to fail in much the same way as enjoying the bottle of Pastis, which we brought back home with us because this drink tasted so wonderful in that sidewalk cafe‚ in the South of France. At home in our Dutch garden, it is as if the contents of the bottle were changed on the way home.

These are the aspects of the language-context that one can lose in the concept analysis of thick concepts. The outcome of language analysis always seems impoverished compared to the riches of language. The Dutch word gezellig is held to be untranslatable (closest possible translation is something like, but not exactly, cozy, pleasant, comfortable, snug). Rom Harré has stated that gezellig may look like the English word cozy, but at the same time it is also different. Both words have the combined meaning of, on the one hand, an internal feeling, and on the other hand, of describing a certain place, but it is impossible in Dutch to feel gezellig when one is alone (Harré‚ 1986, pp.11-12). The word gezellig has a positive connotation for most people, but for some it is utterly unpleasant to have to be ‘gezellig.’ Most Dutch people use the word gezellig in every other sentence.

In the context of words like gezellig, Harré speaks of quasi-emotions. Being bound to time and place makes this type of words outstandingly context bound.12 To compare certain concepts between several languages, the explanation of competent users of these languages is absolutely necessary. Also, Harré uses the information of users of Dutch for a further explanation of the concept ‘gezellig.’ And yet it seems that only a slight feeling for languages suffices for one to notice that the concept does not fit—that the meanings of seemingly the same word are not the same in various languages. Cees Schuyt mentions the reaction of the American sociologist, Alfred Schütz, to the question whether or not he had become happy in the end, twenty years after fleeing from his homeland. Schütz replied: "I am happy, aber glücklich bin ich nicht," and while we think we understand what Schütz meant to say, while we will never have the command of both languages as native speakers (Schuyt, 1995, p.24).

Just how incomprehensible a language may be for someone who misses the context, I was able to demonstrate in South Africa during the 14th International Human Science Research Conference in 1994 (a conference on qualitative research) by commenting on the poem "Jonge Sla" (Young Lettuce) by Rutger Kopland, a Dutch poet.13

I can bear anything,
beans withering, dying flowers,
dry-eyed I can watch them
lift a patch of potatoes,
indeed, then I am a tough one.

But young lettuce in September,
just planted, still limp
in their moist little beds, no.

This is an utterly Dutch poem. Who else would bother about something so trivial as new lettuce? Kopland pictured his young lettuce plants as new-born babies. The small beds appear in a double meaning which has still remained in the English translation. But the definite barrier for understanding this poem is, of course, contained in the word "September." In the Southern hemisphere, spring starts in September. The new lettuce in Kopland’s poem in the Dutch version becomes extra vulnerable against the background of the coming autumn’s approach. For my South African audience the poem was incomprehensible: the key words have a different meaning because of the total difference in context.

Concept analysis itself is not poetry, but, yes, it has its eye open for the poetic elements in language. Concept analysis is argumentative, it is not evocative. It tries to explicate the evocative elements in language. It tries to answer the questions of which feelings are evoked by those words and how it is possible that these words evoke those feelings. Such questions can only be answered by referring to the context. If rich language loses a part of itself, however small, it loses its complete capital in this way. In general, the more limited an area is where a certain language is spoken, the richer that language. If one does not know this specific area, one understands nothing at all.

Concept analysis resembles phenomenological analysis. This is not only because of the related access—everyday experience has been replaced by everyday language, as I said before— but specifically so because of the methodical similarities. In the area of phenomenology, one is hardly concerned with a true reality which would be found behind the phenomenological reality. There is no true reality beyond everyday experience. We will have to make do with the world as we find it in the everyday reality. We must not look for meanings in the relation between the phenomena and a world beyond, but in the relation between the phenomena mutually.

In order to be able to analyze everyday language, I shall have to sport an uninhibited attitude. That is to say, in approaching language, I shall have to give up definitions as I have got to know them. That is, and I want to stress this point once more, the most important reason why I prefer the analysis of thick concepts. General concepts, like, for example, ‘education’ seem to have been analyzed to death, and what has remained seems to be a survey of fruitless differences between fixed opinions. Had anyone ever ventured to undertake the analysis of the concept ‘sweet?’ The world of the thick concepts is still full of virgin territory.

In order to be able to analyze everyday language, it is necessary to conduct a systematic variation. In this systematic variation of a certain concept, I try to trace the concept’s idiosyncratic qualities of meaning. In concept analysis the criterion is: "It is absurd to say…." The discovery that one can not say something in a certain way shows one the way to the correct usage. We not only get to know certain concepts by using them, but we analyze a concept’s meaning to discover its correct usage. In this way we discover which characteristics belong to the concept, and which do not. In Dutch, though, it is really absurd to use the concept ‘sweet’ in relation to an adult.

The criterion "It makes sense to say" is disputable. In the discussion of concept analysis as a philosophical method, this criterion is evidently also a weak point. Ludwig Wittgenstein, in Filosofische Untersuchungen 246, states that it is absurd to say: "I know I am in pain." We say "I am in pain" and Wittgenstein concludes that ‘to be in pain’ is not a form of ‘knowing.’ In his plausible criticism of this identification of meaning and use of the word, John Searle says that we do not say that we "know that we are in pain," because that is too obvious. In Searle’s vision it is so evident that, when we are in pain, this is also a matter of knowing, so we can easily omit it (Searle, 1969, Ch. 6).14 The author Gerrit Krol had long been struck by the fact that we Dutch know the word ‘headwind’ and ‘sidewind’, but not the word ‘tailwind.’ Of course the phenomenon ‘tailwind’ exists, as does the phenomenon ‘headwind,’ but we actually do not feel a tailwind. We are only aware of how easily our cycling went at the moment the wind changes into a headwind. It seems as if when someting is not substantial, we have no substantive for it; but if this substantive is lacking, it naturally does not imply that the phenomenon does not exist.15

John Langshaw Austin’s dictum that language has the first word—and not the last—remains in full force for concept analysis. The fact that we simply do not say something in a certain manner is the start of the argument—not the conclusion. Moreover, the identification of meaning and use of thick concepts contains far fewer risks because of the perspective in which we place the concepts’ analysis than is the case with analysis of general philosophical or general pedagogical concepts. I am not so much concerned with the concepts’ meanings, but rather with their rules. And these rules, I have found, are food for thought. I shall have to determine my position with regard to these rules. The discovery that it is improper to place ‘sweet’ and ‘naughty’ opposite each other gives rise to the question whether or not it makes sense to admonish children to be sweet (in the sense of good). Because ‘sweet’ equals ‘not troublesome,’ we are confronted with the question whether or not the preference for ‘sweet children’ serves a pedagogic interest or simply the educator’s own interest. Etcetera.

Can concept analysis really be defended as an empirical method?

In the attack on the simplistic opposition of language and reality, the conclusion could only be that concept analysis not only provides us with information about language, but also with information about the (social) reality. In the analysis of language, experience itself has a part to play from the beginning. The description of a sweetly playing child as a young child completely absorbed in its own activity and, because of that, not asking for attention has, of course, been based on previous observation and not on a description of the rules of language usage. This is an observation that has never been systematically described in writing, but that can be recalled if necessary. In principle, any competent language user is capable of concept analysis, but a successful analysis requires a great deal of life experience, a substantial gift of observation and considerable imaginative powers. Also in the language itself, centuries of experience lie hidden. All in all, in language analysis several aspects of reality are analyzed, as well.

But this does not mean that empirical research in social sciences can be replaced by concept analysis. Apart from the starting point that knowledge of social reality really profits from multiple (and pluriform) access roads, the empirical content of a concept analysis remains limited. It is perhaps not necessary to go to the school playground to discover the difference between ‘bullying’ and ‘teasing,’ but the fact that children may be practically bullied to death won’t be discovered from one’s easy chair, that is for sure. And of course, nobody doubts the fact that, in order to get to know how children suffer under other children’s bullying behavior, it will be necessary to speak with them and observe them. Moreover, every researcher should be convinced of the need to look beyond qualitative research methods to be able to do adequate research into such a phenomenon.16

Endnotes

1. For the record: the sentence "I don’t think it was a matter of some special kind of courage, because none of the persons on board thought we could go down" seems to contain empirical information, but it contains nothing but conceptual information, information about the use of the term ‘courage.’ The sentence "If we had thought so, we would not have made jokes, such as: if you come back tomorrow, you will need a ticket" contains empirical information: (the cracking of jokes as) an indication for the contents of thoughts.

2. The issue is more complex than can be indicated here. ‘Not being able to predict,’ e.g., does not exclude determinism. The philosophical discussion on this issue has been enriched with a closer differentiation between hard and soft determinism, in which latter case the possiblity of explaining human behavior remains if the individual’s latest preference is known as a causal factor.

People of flesh and blood, for that matter, have no trouble whatsoever in switching regularly from hard determinism to soft indeterminism when describing their own behavior. The inclination towards an unambiguous vision of human behavior seems to be a preoccupation of philosophers and forms the basis for much discord about the organization of science and its research methods.

3. Such distinctions are found among many analytical philosophers.

4. This is the last time I will write "behavior" in quotations. Rule-governed behavior is, of course, not behavior in a causal-analytical sense. In the way that rule-governed action is not a completely free action, but an action which is oriented towards rules. That is the kind of more precise specifications I mean here.

5. Confusion of concepts is ubiquitous in this respect. Everyday language does not distinguish sharply between ‘behavior’ and ‘action’ as I started doing. The same holds for ‘patterns’ and ‘rules.’ Social patterns are not patterns, but rules. This is valid for traffic rules, for instance.

6. It is not only the concept ‘rule’ I should say more about. In philosophy one finds unfinished discussions about the content of many of the concepts I have mentioned above, such as ‘cause,’ ‘intention,’ and the differences between regulative and constutative rules, matters I can not go into here.

7. In this respect it is interesting to see what Paul H. Hirst and Richard S. Peters have to say on the analysis of general educational concepts in The logic of education in 1970. In relation to what I have said here, they provide us with a concise but nuanced introduction of the place of conceptual analysis in the philosophy of education (1-16). How different is Peters’ view on the fruitfulness of the analysis of general educational terms in 1976. D. Cooper recalls him saying at the celebration of the 10th anniversary of the PESGB "... that he could scarcely bear reading yet another article or dissertation with a title like ‘Two senses of teaching’ or ‘The concept of the curriculum’" (Cooper, 1986, p. 4).

8. Williams (1985) mentions the following characteristics of thick concepts. They form a unity of fact and norm, but the large character is found in the descriptive part (129). Thick concepts are action-guiding, they are not decisive for the action, but play a part there (140). There are concepts you can not use either as a man or if you are not local (143-145). Thick concepts occur less in modern societies (163). Ethical life is based on the use of thick concepts (200).

Of course, there is a connection between the analysis of thick concepts in conceptual analysis and the preference for thick description in cultural anthropology. This term is usually attributed to Clifford Geertz (1973), who, in his turn, refers to Gilbert Ryle.

9. If babies cannot fall asleep, they allow themselves to be soothed to sleep by means of a bottle of sugared water. But there are children who refuse to go to sleep and simultaneously refuse the sweetener.

10. Perhaps this is a case of an enormous local prejudice. I know I should continually be conscious of this consequence of the words’ meaning’s contextuality. All through the ages, foreign visitors have been annoyed by Dutch children’s brazen and improper behavior. In the eighteenth century the Frenchman Denis Diderot mentioned it and the Englishman Colin White and the American Lourie Boucke wrote about it in The Undutchables. An observation of the Netherlands: its culture and its inhabitants (1989).

11. Perhaps it would be a good thing to subject my suspicion that the Dutch language is a special case, where specific characterization of children’s existence is concerned, to more general critical research. In French, German and English, related equivalents of ‘sweet’ are indeed found (doux, süss and sweet), but they can also be used for adults, so they are not reserved for young children only, as in Dutch.

An important issue which crops up here is whether or not one can oblige children to be sweet. I think not. It is only possible to admonish older children not to be naughty, "Sois sage!" (= be good) makes sense.

The meaning of thick concepts may change with the passage of time. It is also possible that certain concepts do not fit the period any longer. It remains disputable whether or not the moral connotation, as mentioned above and the childlike competencies which are included in the concept sweet are acceptable to everyone. The next problem is then, however, whether ‘sweet’ has (acquired) another meaning or the word has become obsolete in this sense, because children are no longer sweet in its original meaning.

12. Being time and context bound is crucial and this is found explicitly in dialects or regional speech. The most recent Dutch research project in this field dealt with the use of animal names in certain expressions and it is obvious that the names of those animals crop up in expressions which are found in that particular region. The research findings have not yet been published.

13. I followed the example of Kees Fens (a Dutch writer and literary critic) who once took "Jonge Sla" with him on a trip to Indonesia and worked on the poem with a number of students of Dutch. Kopland himself later did the same thing on a trip to Eastern Europe.

14. From the example in On Certainty, 591, it becomes clear that Wittgenstein is aware of the problem stated by Searle. In this example, Wittgenstein demonstrates that "I know" in some cases is only added as an extra certainty.

15. Note how obvious it is for us, Dutch, implicitly to refer to cycling experiences, in this case. The example itself would have been less obvious in a less windy country.

Alright, the word ‘go-with-me-wind’ does not exist in the Dutch language; we know the word ‘tailwind.’ Tailwind (literal translation from the Dutch, would be: ‘back-wind’) does not occur before the 12th edition of the authoritative Van Dale dictionary, when the phenomenon was apparently noted (1992). At a certain moment in time, one started stating the amount of tailwind during certain sport records and since then the term has become fashionable.

16. In the NRC-Handelsblad of 16 May 1998, Thijs Goldschmidt (social scientist) pleaded for a long-term behavioral-biological research project into bullying behavior in the schoolyard, while doing away with the term and playing down the experience aspects. "Teacher-badgering" is very similar to the way young chimps try out the older chimpansees’ authority. Based on his findings, he criticizes the system of classes formed by pupils of the same age, all having to learn the same subject matter. All those young people of the same age in the same class, having to listen uncreatively, that is just asking for bullying.

J. Junger-Tas and J. van Kesteren stated in a quantative research project in 1998, among other things, that children who bully regularly often come from deprived backgrounds and experience little support from their teachers. They also stated that among bulliers, many children from different cultures are found — though it may be that cause and result intermingle here — and that remarkably little bullying occurs at homogenous Christian schools.

Summarizing, there are also processes controlling the landscape of bullying outside the participants’ conscious actions which we researchers, should of course, know as well.

References

Abrahams, F. (1998, May 14). Gedachtenpolitie. NRC-Handelsblad.

Austin, J. L. (1976). Philosophical papers. (2nd ed.). (J. O.Urmson & G. J. Warnock, Eds.). Oxford: Clarendon Press.

Beekman, A. J. (1982). Theoretische pedagogiek. Een existentieel fenomenologische visie. In B. Spiecker, B. Levering, & A. J. Beekman (Eds.), Theoretische pedagogiek (pp. 13-24). Meppel, The Netherlands: Boom.

Cooper, D. E. (1986). Education, values and mind: Essays for R.S. Peters. London: Routledge & Kegan Paul.

Davie, M. (1986). The Titanic: The full story of a tragedy. London: Bodley Head.

Delft, D. van (1997, July 21). Alle talen van de wereld. Vergelijkende taalwetenschap worstelt met betekenissen. Interview met Prof.dr F.H.J.Kortlandt. NRC-Handelsblad.

Geerts, G. & Heestermans, H. (1992). Van Dale. Groot woordenboek der Nederlnadse taal. Utrecht/Antwerpen, The Netherlands: Van Dale lexicografie.

Geertz, C. (1973). Thick description: Towards an interpretative theory of culture. In The interpretation of cultures: Selected essays (pp. 3-30). New York: Basic Books.

Goldschmidt, T. (1998, May 16). De biologie van het pesten. Onbevooroordeeld etholoog node gemist op het schoolplein. NRC-Handelsblad.

Harré, R. (1986). An outline of a social constructionist viewpoint. In R. Harré‚ (Ed.), The social construction of emotions (pp. 2-14). Oxford: Basil Blackwell.

Hendriks, M. (1997, June 4). De tijd op school heeft grote invloed op kinderen. De Volkskrant.

Hirst, P. H. & Peters, R. S. (1970) The logic of education. London: Routledge and Kegan Paul.

Junger-Tas, J. & Kesteren, J. (1998). Bullying and delinquency in a Dutch school population. Leiden: Vakgroep Criminologie en Rechtsgeleerdheid.

Kopland, R. (1969). Alles op de fiets. Amsterdam, The Netherlands: Van Oorschot.

Kopland, R. (1996, Februaru 24). Jonge sla in februari. De Volkskrant.

Krol. G. (1984, November 24). Nimmer teruggekeerd. De Volkskrant.

Levering, B. (1987). Conceptuele analyse en normativiteit. Pedagogische verhandelingen, 10 (3), 34-48.

Levering, B. (1993). The Language of Disappointment: On the language analysis of feeling words. Phenomenology and Pedagogy. 10, 53-74.

Levering, B. (1995). Education has something to do with democracy indeed. Abstracts of the 14th International Human Science Research Conference. Pretoria: University of Pretoria.

Pieper, J. (1963). Vom Sinn der Tapferkeit. München: Kösel Verlag.

Schuyt, C. (1995). De problematiek van de gammawetenschappen. In T.Koopmans (Ed.), De toekomst van het Nederlands als wetenschapstaal (pp. 21-25). Amsterdam: KNAW/Noord Hollandsche.

Searle, J. (1969). Speech acts. An essay in the philosophy of language. Cambridge: CUP.

Stäblin, R (1997). Vorwort. Zwischen Furcht und Zuversicht. In R. Stäblin, (Ed.), Mut. Wiederentdeckung einer persönlichen Kategorie. Frankfurt a.M.: Fischer.

van Manen, M., & Levering B. (1996). Childhood’s Secrets. Intimacy, privacy and the self reconsidered. New York: Teachers College Press.

Williams, B. (1985). Ethics and the limits of philosophy. London: Fontana Press Collins.

Wittgenstein, L. (1974). Über Gewissheit/On Certainty. Oxford: Basil Blackwell.

Wittgenstein, L. (1997). Philosophische Untersuchungen/Philosophical Investigations (2nd ed.). Oxford: Basil Blackwell.

White, C. & Boucke, L. (1994). The Undutchables. Leven in Holland. Amsterdam, The Netherlands: Nijgh en Van Ditmar.