Resumé

A concept is the name of any general element in one's experience, and,
consequently, in social science all conclusions are based on concepts. The main
object of this article is to indicate the existence of a logical gulf between the
concept as an object of analysis made by social science and the concept as a means
to accomplish the analysis.

I

Empirical definitions are an important contribution of neo-positivism to political science, and one of the most significant doctrinal implications of this development is a dissociation of political science from metaphysical idealism based upon formalism and legal conceptualism (Begriffsjurisprudenz).

It goes without saying that this little introduction to a few observations on concepts in social science in general, and political science in particular, cannot trace the development of its meaning and application in the course of time. The import of positivism on the thinking of social scientists has been a distinctive feature of recent schools of thought. The Vienna and its followers asserted that any statement which, in principle, could not be verified was, logically speaking, meaningless or nonsense. That is the positivist description of what later on was termed value premises. On the other hand, the factual aspect of a proposition refers to reality and can be verified or falsified by reference to the facts.

The Swedish philosopher Axel Hägerström, founder of the Uppsala developed positivism into a consistent value-nihilism. In his opinion any sentence that was not verifiable had to be meaningless. From this principle of verifiability follows the logical gulf between Is and Ought. The German philosopher Theodor Geiger (1960) went even further when

he adopted a practical positivism according to which the ethical demand was made that everyone must live in conformity with these principles. Geiger maintained the objective nature of reality understood as a part of the external physical world.

The European positivists rejected value judgments as a part of scientific argumentations and regarded these judgments as subjective attitudes, not to say theology. According to the positivist view, value judgments are meaningless because they cannot be 'proved' by any intersubjective lest. Apart from the purely theoretical questions in this connection this attitude led to an overemphasis of facts and an underestimation of values in social science.

The political philosophers mentioned above have been concerned with the substantial content of the sentence as a grammatical entity and the concept has been placed as an integrated part of the sentence only. We are talking about Is-sentences and Ought-sentences. Thorstein Eckhoff (1971, 19fO goes further in this direction and speaks about the concept of strategy as a goal-oriented behaviour choosing among the several behaviour-alternatives. His observations have a certain resemblance to Herbert A. Simon's concept of strategy as the series of decisions which determines behaviour over some stretch of time (Simon 1947, 67).

American political science turned against the speculative character of social science even before the reaction in Europe to the legal conceptualism. It was James Bryce who opened the way to the new perception of science in the study of politics. American pragmatism rebelled against all a priori statements and against the traditional practice of deductively deriving postulates from such principles. Political science, Bryce points out, 'is not a deductive science any more than it is a branch of speculative philosophy'. He rhetorically exclaimed that 'it is Facts that are needed: Facts, Facts, Facts. When facts have been supplied, eac'i of us can try to reason from them' (Bryce 1924, 12).

Even earlier American students began to acknowledge that in the past they had been inordinately immersed in speculation. Part of this reaction developed in the shape of rationalistic jurisprudence, and this was the basis of American realism in legal science.

Beginning with the latter half of the 19305, some ne opositivists abandoned their original positions, and a number of authors independently arrived at a conception of these problems which may be named Value Relativism. One of them was Gunnar Myrdal (1958), who entered the scene from the outset as both a positivist and a relativist, and Arnold Brecht (1959, 117f0, who formulated the so-called Scientific Value Relativism in which he maintained that the question whether something

is 'valuable' can be answered scientifically only in relation to some goal which it is useful to pursue or to the ideas held by some persons regarding what is valuable. Consequently, he found it impossible to establish scientifically what goals are valuable irrespective of the value they have in the pursuit of other goals or of someone's ideas about ultimate goals or purposes. Like the positivists the above authors admit that the value aspect of a proposition expresses only the emotional response to a state of facts, but consider that the adoption of this view does not preclude the possibility of examining values as observable facts.

Based upon these observations and the works of David Easton (1953, 220ff) and Herbert A. Simon (1947, 45ff), an attempt has been made by Poul Meyer (1973, 136ff; 1979, 296ff) to construct an abstract decisionmaking model based on the conception of the decision as a synthesis of premises brought about by a syllogistic argument. From this follows that the decision must be right by logical necessity, provided the premises are right, so that any other decision will lead to a contradiction in terms.

II

This brief review of the development of theory within this field has established that the sentence is the principal element in social theory. Turning to the corresponding aspect of the problem within analytic philosophy we will find that opinions differ much more on the subject. Some schools concentrate on the meaning of sentences, and identify the meaning of an expression with what it might refer to or be true about, at least in some cases of an observable physical object or situation.

In contrast, Niels Egmont Christensen (1961, 92) has advanced the thesis that 'the meaning of an expression pertains to the capacity of that expression of being rightly produced when and where, and only when and where, something specific of a non-linguistic kind is present, be it an object, property, relation, situation, or whatever it may be'.

The British philosopher H. L. A. Hart (1954) has asserted that the concept is irreducibly defeasible, and owing to the defeasible character of a statement it may well express the necessary and sometimes sufficient conditions for the application of the concept, but would not be able to express conditions which are always sufficient. By the term defeasible Hart understands that the concept can be neither defined nor adequately characterized.

The greater part of modern philosophers keep to the use of the concept as an analytical entity. P. L. Health (1967, 177ff) has offered a systematic summary of the writers in this field. The alleged opinion of the majority is expressed in a category named 'dispositional theories', according to

which concepts are peculiar to the individual. They point out the fact that the social development normally will lead to a sharing of concepts and, to the eventual acquisition of a standard repertory of concepts held in common by virtually all members of a given cultural or linguistic group. This opinion is related to the popular understanding of the problem which I shall describe in further detail below.

On the other hand, Health is opposed to the so-called 'entity-theories' according to which concepts are to be treated merely as logical constructions out of the actual or possible occasion of their employment. Health points to the fact that according to these theories concepts are phenomena which are calculated into existence. The entity theories form the basis of my attempt in the following to construct the concept as a technical term.

As indicated above, it is necessary to distinguish between concepts in everyday speech and concepts as technical terms. The logical fallacy of many philosophers and social scientists is, however, that they try to compare phenomena which logically cannot be compared. It stand!» to reason that natural science is able to define the colour red by way of the wavelengths of the spectrum, whereas the word has many varying meanings in everyday speech. On the other hand, it is t!ie task of social science to analyse the different meanings of the word red in everyday speech and in other fields of human activity and to accomplish the survey by way of concepts which are of exactly the same logical character as the concepts used in natural science.

Alfred Schutz (1963, 247) explains the behaviour of the actors on the
social scene by way of the following social action model which comprises
everyday concepts as well as technical concepts:

The social scientist observes certain facts and events within social reality and constructs typical patterns from what has been observed. Thereupon he co-ordinates to these patterns models of an ideal actor. To this fictitious actor he ascribes a set of typical notions, purposes and goals. These constructs are not arbitrary; they may be constructed in such a way that a human act performed within the real world would be understandable to the actor himself as well as to his fellow-men in terms of common-sense interpretation of everyday life.

Another conceptual confusion should be mentioned in passing, namely the confusion of the term method, understood as a conceptual analysis of relevant data, and technique understood as the process of collecting material for the analysis. It ought to be evident that an argument about the method of social science must be incapable of being supported by any example purporting to show that there are great difficulties in establishing theories in social science or that some social phenomena are relatively inaccessible

to observation or experiment. All kinds of science - social as
well as natural - have the same logical foundation and the ultimate goal
is to verify hypotheses and theories (Meyer 1962, 36f; Rudner 1966, 4f).

Consequently, the distinction between natural science and social science appears on two different logical levels. It is no concern of natural science to analyse the concepts of everyday speech, whereas it is an important task of the social sciences to carry out analyses of this kind. On the other hand, both natural science and social science use empirically constructed concepts as instruments for their scientific analyses of reality. The inferences that underlie everyday speech are of great complexity, and they depend, furthermore, on the application of unconscious assumptions.

It goes without saying that analysis of everyday speech is a small part of social science only, but it is an important part, and general sementics is thereby made part and parcel of our scientific work. Semantics is a very confusing discipline, and it stands to reason that this article cannot contribute to any attempt to simplify the question raised by the several authors.

The core of semantics is the concept of meaning, the complexity of which has been mentioned earlier in this article. The meaning assigned to signs (or sounds) is conventionally established. The significance assigned to the linguistic symbols is determined by the customs of the community, and, consequently, the linguistic customs which establish the symbolic function of language can only be discovered by a study of the way in which people express themselves. The political scientists must make it their concern to find out the 'meaning' of the oral and written utterances of the political actors.

It is a well known fact that many legal concepts will at some point when their application is in question prove indeterminate. Concepts of this kind have been characterized by the term open texture, which means that there are areas of conduct where much must be left to be developed by the courts. As pointed out by Friedrich Waismann (1945), this is a general feature of human language.

/

Consequently, the concepts of everyday speech have an open texture as their application cannot be stipulated and a possible definition can be applied at a given situation only. New situations may arise when the application of the concept is uncertain or impossible.

III

The main object of this article is to indicate the existence of a logical gulf
between the concepts as objects of the analysis of social science and the

concepts which are used by social science to accomplish the analysis. The concluding part of this article will be confined to some observations on the concept as a technical term used by social science for the analysis of texts.

By way of introduction it is necessary to notice that some part of human communication can be carried out without any use of concepts. Considerations of language lead to the observation that naming is the most elementary factor in description. L. Wittgenstein (The Concise Encyclopedia of Western Philosophy and Philosophers, London 1960, 409f) has pointed to the fact that conventional name-giving is the fundamental element to any description. He describes the distinct ways of using language as language games. This, however, presupposes knowledge of the language which indicates the use of concepts in the embryonic stage as name-giving. In this way we think of a word as being always the name of something which is described by ostensive definition or pointing out. It is, however, only a very small part of human communication which can be carried out without any use of concepts.

If each description of experience were assumed to be unique, no science would be possible because generalizations are necessary for describing a group of phenomena when every element of the description is present in every one of the phenomena in question. David Easton (1953, 4) rightly claims that Tacts must be ordered in some way so that we can see their connections. The higher the level of generality in ordering such facts and clarifying their relations, the broader will be the range of explanation and understanding'. Actually every experience contains elements of both abstractness and uniqueness. In this study, however, I am not interested in the actual result of empirical studies but in the kind of interrelationship between social phenomena which may be defined in the abstract.

Concepts are generalizations in that they describe some attributes to two or more objects. In social science all conclusions are; based on concepts, and out of the whole stream of experiences we must abstract limited aspects that are interesting and communicable through natural or symbolic language. All concepts are abstractions because; we cannot observe real objects corresponding to the terms 'man', "politician', 'prime minister' etc. 'That which all conceptions of an object must have in common, is the concept of the object' (Greer 1970, 43).

The preceding remarks have by their evident theoretical implications served to illustrate the connection between conceptualism and description. Nevertheless, it has been argued in some recent writings that political research ought to emphasize the fact-finding techniques and - basing

their results on public opinion surveys and statistics - these authors avoid as much as possible the use of concepts (Damgaard 1977, 12, 25; Christensen 1980, 9). Of these works it may be generally said that they bear directly on the questions raised by the school of Public Policy understood as the use of elements which are not easily reconstructed into a logical, orderly presentation such as political judgments and political values. Strictly speaking, it is, of course, possible to some degree only to avoid the use of concepts, but apart from the technical development during the twentieth century this outlook is a strange return to the 'hyperfactualism' which regarded science as the objective collection and classification of facts. I agree with Scott Greer (1970, 70) in his characteristics of this outlook: 'Unimaginative empiricism, in which one simply collects cases in terms of a sterile framework, is not only improfltable, it may be a barrier to discovery'.

Empirical concepts are more or less complex. Some concepts refer to a few factual phenomena, but most of the concepts which are used by jurisprudence and social science are of a very composite character. Composite concepts are not only syntheses of observable facts, but include other concepts of a lower degree of abstraction. The necessary condition of composite concepts in the technical sense of the word is that the individual concepts of which they are composed can be reduced to observable phenomena. The greater the degree of abstraction, the more difficult the determination of the empirical elements of the concepts.

It is rather easy to define the concept of 'member of parliament' by means of relevant rules and the necessary action within the framework of these rules, but it is very complicated to present an exhaustive definition of the concept of 'parliamentarism' as an empirical concept in the above sense.

It has been seen earlier that empirical concepts are man-made and that the several elements of a concept are chosen on the basis of expediency. I have indicated the difficulties bound up with the complicated character of some concepts. Another difficulty is the fact that in practice technical terms are under the influence of common language. This should be remembered when criticism is levelled against the use in social science of foreign words and the amplification of new words that sound terrible in the ears of 'the man in the street'. These artificial words are not always the outcome of academic affectation, but may be an attempt to avoid expressions which soon after their introduction into scientific terminology will absorb everyday implications. For that reason Svend Ranulf (1946, 54) recommends the use of technical designations of concepts rather than using traditional terms, which may be loaded up with inexact

platitudes from everyday speech. Concepts in social science must have
both fidelity to the social system and clarity and communicability.

The concomitant weakness of a large part of technical language is a frequent development of the meaning of the words during the space of time. By way of example may be mentioned that the institution named madhouse was changed to lunatic asylum and later to mental hospital or nursing-home to avoid the derogatory connotation linked up with uhe older designations. This is a further argument for using technical terms.

These inconveniences may in some cases be avoided by using mathematical symbols. It is, of course, impossible to define mathematics in this article. Common to all definitions is, however, the demand of pure reasoning without recourse to observation and experiment. The mathematical symbols produce a complete abstract generality, i.e. a language for all occasions. Consequently, mathematical knowledge is gained independently of empirical verification. When we use mathematical symbols in social science, however, it is indispensable to be aware of the fact that the symbols in these cases represent well defined social phenomena. They are no longer purely logical axioms, but indications of empirical unities. The application of mathematical symbols in social science must not be confused with statements of chance and tendency statements.

The definition of a concept presupposes static social surroundings and the expediency of the formal definition must be considered on the basis of the society in question. Consequently, if we consider the context in which we normally use this terminology, it will be seen that the qualitative change of the entities tends to influence longitudinal research. The elements of the concept and the societal factors are interchanging. This is a clear warning against any uncritical use of longitudinal series of numbers representing qualitative entities.

There are other dangers of distorting the concepts no matter how well-defined they are, for example the so-called reification, i.e. treating an abstraction as concrete. According to the above terminology the state is a well defined empirical concept even though it is a difficult question of expediency to choose the factual elements of the concept. In practice, however, the state is frequently treated as a fictitious person which predicates the real existence of an entity quite apart from the human beings who are the visible actors. Alf Ross (1953, 218) has rightly called this usage a metaphysical pseudo-problem as it is impossible to treat a collective unit as an individual person. There is nothing to say in support of this terminology, and there are many dangers in it (Popper 1958, 2 Iff).

IV

Speaking about norms, it seems necessary to touch on the relation between jurisprudence and social science. In a few words and rather superficially jurisprudence may be described as a study of legal norms, i.e. the norms which are the basis of the decisions of courts of law. Jurisprudence, consequently, is a separate part of social science when this discipline is understood as the description of all kinds of social facts and values, for example social norms in general.

In the main I share these views contrary to Jørgen Dahlberg-Larsen (1977), who advocates a change in the discipline from pure dogmatics to a mixed study of legal and factual relationships. Furthermore, it should be emphasized that the concepts - whether legal or not - are logically constructed at a lower level than the definition of scientific disciplines, and at this level all concepts are of an equal standing. Consequently, the question of conceptualization within jurisprudence is identical with the corresponding question within other branches of social science.

It is impossible in a concise work of this kind to go into a detailed study of the conceptional relations between jurisprudence and other social sciences. Let us content ourselves with a few observations on the law of property. It has been mentioned earlier that neo-positivists reacted against the traditional conceptualism in a way that led to the opinion that 'rights' were terms without any semantical reference at all. As we have seen in the preceding this is logically nonsense, because a concept of right as well as any other empirical concept has exactly the substantial content that we have given it ourselves, and no more. Knud Ilium (1945, 177) points to the fact that it is impossible to draw out of a concept more than we have injected into the concept by definition. From this it follows that the conception of property is a synthesis by way of abstraction of a number of factual elements. Intellectually it is the same process by which the courts decide from case to case to whom a certain right belongs.

If we consider the contexts in which we normally treat the question of empirical concepts, this article may seem to be concentrated on the concept as a part of factual sentences. Therefore, let it be clearly stated that these concepts appear in ought-sentences as well as in is-sentences. When we are saying that Danish politics ought to fight unemployment, the concept of unemployment could be empirically defined according to the above description of empirically defined concepts.

An empirical concept as defined in this article has been named operational, understood as a concept which assigns meaning to a construct by specifying the activities necessary to measure the construct. It has been asserted that the weakness of operationalism in social science stems from

the difficulty of measuring the phenomena. By way of example may be cited Scott Greer (1970, 43), who has expressed himself very categorically with the following remark about political scientists: 'Their concepts did not seem to imply measurable qualities while what could be measured did not seem to imply their concepts'.

The first thing that needs to be said about this criticism is that concepts which do not imply measurable qualities do not belong to the category of empirical concepts either. Furthermore, since the concepts are constructs, it is a matter of convenience only whether any measurable element will be used in a process of conceptualizaton. Finally, many empirical concepts are defined by elements which cannot be the object of measurement in the traditional sense of the word.

As mentioned before, we cannot observe the concepts of 'member of parliament', 'parliamentarism' etc., but they are empirical concepts because we are able to observe the constituting elements, the text of the law, the behaviour of the elected etc. Classification consists of grouping phenomena on the basis of their shared characteristics.

Some philologically oriented sociologists throw doubt on the propriety with which operational definitions are thought of as definitions at all and conclude that the conveyance of meaning is not a necessary and sufficient condition for something to achieve the status of definition (Rudner 1966, 20f). Statements of this kind are, however, based on the traditional confusion of concepts in everyday speech and technical concepts as defined in this article.

REFERENCES

Brecht, A. 1959. Political Theory. New Jersey: Princeton University Press.

Bryce, J. 1924. Modern Democracies. I. New York: Macmillan.

Christensen, J. G. 1980. Centraladministrationen: organisation og poli.'isk placering. Copenhagen:
Samfundsvidenskabeligt forlag.

Christensen, N. E. 1961. On the Nature of Meanings. Copenhagen: Munksgaard.

Dahlberg-Larsen, J. 1977. Retsvidenskaben som samfundsvidenskab. Copenhagen: Juristforbundets
Forlag.

Damgaard, E. 1977. Folketinget under forandring, Copenhagen: Samfundsvidenskabeligt
Forlag.

Eckhoff, E. 1971. Rettferdighet. Oslo: Universitetsforlaget.

Geiger, T. 1960. "Die Gesellschaft zwischen Pathos und Niichternkeit. Acta Jutlandlca.
Copenhagen: Munksgaard.

Greer, S. 1970. The Logic of Social Inquiry. Chicago: John Wiley.

Hart, H. L. A. I 954. 'Definition and Theory Jurisprudence. Law Quarterly Review (London).

Health, P. L. 1967. 'Concept' The Encyclopedia of Philosophy, Vol. 11. New York: Macmillan.

Ilium, K. 1945. Lov og Ret. Copenhagen: Busck.

Meyer, P. 1962. Politisk Videnskab. Copenhagen: Berlingske.

Meyer, P. 4th ed., 1979. Offentlig Forvaltning. Copenhagen: Gad.

Myrdal, G. 1959. Value in Social Theory. London: International Library of Sociology.

Popper, K. 1958. The Open Society and Its Enemies. New Jersey: Princeton University
Press.

Ranulf, S. 1946. Socialvidenskabelig Metodelære. Copenhagen.

Ross, A. 1953. Ret og Retfærdighed. Copenhagen: Busck.

Rudner, R. S. 1966. Philosophy of Social Science. New Jersey: Englewood.

Schutz, A. 1963. "Concepts and Theory Formation', in Philosophy ofthe Social Sciences,
Ed. Maurice Natanson. New York: Random House.

Simon, H. Administrative Behavior. lst edition 1947. 3rd ed. 1976. New York: Macmillan.

Waismann, F. 1945. Verifiability, Logic and Language, London.